Simultaneous Interpreting from a Signed Language into a Spoken Language: Quality, Cognitive Overload, and Strategies 2020048387, 9780367416997, 9780367815769, 9780367757885

Citation preview

Simultaneous Interpreting from a Signed Language into a Spoken Language

This book examines conference-level simultaneous interpreting from a signed language into a spoken language, drawing on Auslan (Australian Sign Language)to-English simultaneous interpretation data to explore the skills, knowledge, strategies, and cognitive abilities needed for effective interpretations in this language direction. As simultaneous interpreting from a spoken language into a signed language is the widely accepted norm within the field of signed language interpreting, to date little has been written on simultaneous interpreting in the other language direction. In an attempt to bridge this gap, Wang conducts microanalysis of an experimental corpus of Auslan-to-English simultaneous interpretations in a mock conference setting to investigate different dimensions of quality assessment, interpreting strategies, cognitive load, and the interpreting process itself. The focus on conference-level simultaneous interpreting not only allows for insights into the impact of signed language variation on the signed-to-spoken language simultaneous interpreting process but also sheds light on the unique demands of conference settings such as the requirement of using a formal register. Acting as a bridge between spoken language interpreting studies and signed language interpreting studies and highlighting implications for future research on simultaneous interpreting of other language combinations (spoken and signed), this book will be of interest to scholars in translation and interpreting studies as well as active practitioners in these fields. Jihong Wang is a lecturer in Chinese/English translation and interpreting at The University of Queensland, Australia. She completed a PhD thesis on the relationship between professional Auslan/English interpreters’ working memory capacity and simultaneous interpreting performance. She conducts research on signed language interpreting, spoken language interpreting, cognitive aspects of interpreting, simultaneous interpreting, remote interpreting, machine interpreting, and sight translation.

Routledge Advances in Translation and Interpreting Studies

52 Mapping Spaces of Translation in Twentieth-Century Latin American Print Culture María Constanza Guzmán Martínez 53 A Century of Chinese Literature in Translation (1919–2019) English Publication and Reception Edited by Leah Gerber and Lintao Qi 54 Advances in Discourse Analysis of Translation and Interpreting Linking Linguistic Approaches with Socio-cultural Interpretation Edited by Binhua Wang and Jeremy Munday 55 Institutional Translation and Interpreting Assessing Practices and Managing for Quality Edited by Fernando Prieto Ramos 56 M ultimodal Approaches to Chinese-English Translation and Interpreting Edited by Meifang Zhang and Dezheng Feng 57 Translational Spaces Towards a Chinese-Western Convergence Sun Yifeng 58 African Perspectives on Literary Translation Edited by Judith Inggs and Ella Wehrmeyer 59 Interpreters and War Crimes Kayoko Takeda 60 Simultaneous Interpreting from a Signed Language into a Spoken Language Quality, Cognitive Overload, and Strategies Jihong Wang For more information about this series, please visit https://www.routledge. com/Routledge-Advances-in-Translation-and-Interpreting-Studies/bookseries/RTS

Simultaneous Interpreting from a Signed Language into a Spoken Language Quality, Cognitive Overload, and Strategies Jihong Wang

First published 2021 by Routledge 52 Vanderbilt Avenue, New York, NY 10017 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2021 Taylor & Francis The right of Jihong Wang to be identified as author of this work has been asserted by her in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data Names: Wang, Jihong (Translation teacher) author. Title: Simultaneous interpreting from a signed language into a spoken language : quality, cognitive overload, and strategies / Jihong Wang. Description: New York : Routledge, 2021. | Series: Routledge advances in translation and interpreting studies | Includes bibliographical references and index. | Summary: "This book surveys conference-level simultaneous interpreting from a signed language into a spoken language, drawing on data from Auslan-English interpretations to explore the skills, and knowledge, and cognitive demands needed for effective interpreting in this language direction"-- Provided by publisher. Identifiers: LCCN 2020048387 | ISBN 9780367416997 (hardback) | ISBN 9780367815769 (ebook) Subjects: LCSH: Simultaneous interpreting. | Simultaneous interpreting--Psychological aspects. | Australian sign language. Classification: LCC P306.95 .W36 2021 | DDC 418/.02--dc23 LC record available at https://lccn.loc.gov/2020048387 ISBN: 978-0-367-41699-7 (hbk) ISBN: 978-0-367-75788-5 (pbk) ISBN: 978-0-367-81576-9 (ebk) Typeset in Bembo by SPi Global, India

To Colin Allen, 31 professional Auslan/English interpreter participants, Jemina Napier, Della Goswell, Andy Carmichael, Brenda Nicodemus, and Teresa Cumpston-Bird who inspired me to complete my research project and write this book.

Contents

Foreword Acknowledgements 1 Setting the Stage 1.1 An Unforgettable Experience  1 1.2 Interpreting Deaf Professionals’ Signed Monologues into Spoken Languages  4 1.3 Research Questions of This Study  6 1.4 The Significance of This Study  8 1.5 Book Structure  9 References  9 2 Overview 2.1 Directionality in Signed Language Interpreting  13 2.2 Simultaneous Interpreting from a Signed Language into a Spoken Language  16 2.3 Cognitive Load in Simultaneous Interpreting  21 2.3.1 Cognitive Load  21 2.3.2 Cognitive Load Models of Simultaneous Interpreting  23 2.3.3 Working Memory and Simultaneous Interpreting  24 2.4 Gile’s Effort Model of Simultaneous Interpreting and ‘Tightrope Hypothesis’  25 2.5 Adapting Gile’s Effort Model  29 2.6 Processing Time in Simultaneous Interpreting  30 2.6.1 Temporal Range of Processing Time  32 2.6.2 Variability in Processing Time  33 2.6.3 External and Internal Factors That Influence Processing Time  34 2.6.4 Relationship Between Processing Time and Simultaneous Interpreting Performance  37

xiii xiv 1

13

viii Contents 2.7 Numbers as a Problem Trigger in Simultaneous Interpreting  41 2.8 Strategies in Simultaneous Interpreting  47 2.8.1 Comprehension Tactics, Preventive Tactics, and Reformulation Tactics  47 2.8.2 Literal Interpretation and Free Interpretation  49 2.8.3 Strategic Omissions  49 2.8.4 Strategic Additions  50 2.9 Summary 52 References  53 3 Methodology 60 3.1 The Auslan Source Text  61 3.2 Creating an Experimental Corpus  62 3.2.1 The Interpreters  62 3.2.2 Auslan-to-English Interpretation Data Capture  64 3.3 Analytic Assessment of Simultaneous Interpreting Performance  65 3.3.1 External Raters  65 3.3.2 Developing an Assessment Rubric  65 3.3.2.1 Selecting Assessment Criteria  65 3.3.2.2 Determining Weightings  68 3.3.2.3 Defining the Rating Scale  68 3.3.2.4 Writing Level Descriptors  68 3.3.3 External Raters’ Analytic Assessment Process  68 3.4 Microanalysis of the Interpretation Corpus Using ELAN and Excel  69 3.5 Thematic Analysis of the Retrospective Interview Data  74 References  74 4 Quality, Self-perceived Challenges, and Self-reported Strategies 4.1 External Raters’ Scores for the Interpreters  77 4.2 Native Signers Versus Non-native Signers  79 4.3 Self-perceived Challenges  80 4.3.1 Signed Language Comprehension Problems  80 4.3.1.1 International Sign, Unfamiliar Signs, and Sign Variation  81 4.3.1.2 The Deaf Presenter’s Signing Style  83 4.3.1.3 Unfamiliarity with the Subject Matter  86 4.3.2 Cognitive Overload  88 4.3.3 Target Speech Production Difficulties  91 4.3.3.1 Trying to Produce Comprehensible English Sentences  91

77

Contents ix 4.3.3.2 Using a Formal Register in English  92 4.3.3.3 Maintaining Coherence in the Target Speech  94 4.3.4 Artificial Testing Conditions  95 4.4 Self-reported Strategies  96 4.5 Discussion 99 4.5.1 Native Signers and Non-native Signers  99 4.5.2 Improving Signed Language Comprehension  102 4.5.3 Enhancing Native Language Proficiency  106 4.5.4 Cognitive Overload and Interpreting Strategies  108 4.5.5 Familiarity and Designated or Preferred Interpreters  109 4.6 Summary 110 References  111 5 Processing Time Regarding Sentence Beginnings, Numbers, and Negation 114 5.1 Measuring Processing Time on 40 Auslan Sentences  115 5.2 Variability in Onset Processing Time  122 5.3 Accuracy of Interpreting Numbers and Processing Time Regarding Numbers  124 5.3.1 Error Taxonomy for Number Interpretations  124 5.3.2 Accuracy Rate of Interpretations of Individual Numbers  125 5.3.3 Accuracy Rate of Interpretations of Numbers in Clusters  128 5.3.4 Processing Time for Accurate Interpretations of a Number Versus Processing Time for Inaccurate Interpretations of the Number  128 5.4 Onset Processing Time Versus Number Processing Time  131 5.5 Consequences of Excessively Short Onset Processing Time  136 5.6 Corollaries of Exceptionally Long Onset Processing Time and Tail-to-tail Span  138 5.7 Interrelation Between Processing Time and Interpreting Strategies  140 5.8 Discussion 142 5.9 Summary 148 References  149 6 Cognitive Overload and Coping Strategies Regarding Numbers 152 6.1 Cognitive Overload and Sentence-level Accuracy Regarding Numbers  153 6.1.1 Number Accuracy Versus Sentence-Level Accuracy  154 6.1.2 Sentence Accuracy Versus Segment Accuracy  156

x Contents 6.2 Free and Literal Interpretations for Coping with Numbers in the Middle of Sentences  160 6.2.1 Long Processing Time, a Lack of Suppression, Self-correction, and Free Interpretation  160 6.2.2 Short Processing Time and Literal Interpretation  164 6.3 Coping Strategies for Numbers Near or at the End of Sentences  169 6.3.1 Topic-Comment Structure and Numbers  171 6.3.1.1 Use of Different Strategies and Printed PowerPoint Slide  171 6.3.1.2 Use of Strategic Additions and Linearity  173 6.3.2 Pseudo-Cleft Structure and Numbers  176 6.3.2.1 Coping Strategies, Self-correction, and Cognitive Overload  176 6.3.2.2 Literal Interpretation, Strategic Additions, and Unjustifiable Omissions  182 6.4 Summary 188 References  190 7 Cognitive Overload and Coping Strategies Regarding End Negation 7.1 Cognitive Overload and Sentence-Level Accuracy Regarding End Negation  192 7.2 Coping Strategies for End Negation  194 7.2.1 Long Onset Processing Time and Free Interpretation  195 7.2.2 Switch from Free Interpretation to Literal Interpretation  200 7.2.3 An Exception: Use of Short Processing Time and Pause to Cope with End Negation  203 7.2.4 Prediction  205 7.3 Discussion of Results  206 7.4 Comments on the Local Analysis  213 7.5 Summary 214 References  215 8 Summary and Recommendations 8.1 Answers to the First Research Question  218 8.2 Answers to the Second Research Question  220 8.3 Answers to the Third Research Question  222 8.4 Key Knowledge and Skills for Success  224 8.5 Strengths and Limitations of This Study  225

192

218

Contents xi 8.6 Implications of the Findings  227 8.6.1 Implications for Deaf Professionals  227 8.6.2 Implications for Signed Language Interpreters Working at Formal Settings  229 8.6.3 Implications for Spoken and Signed Language Interpreters in General  230 8.7 Suggestions for Further Research  231 8.8 Final Thoughts  232 References  232 Appendices Author Index Subject Index

234 251 253

Foreword

I had the pleasure of being Jihong Wang’s supervisor through her PhD candidature, and co-authoring some articles from her thesis, so I am delighted to see that she has returned to her data to produce this monograph. She has worked collaboratively with various people to re-engage with the data and dig deeper into the analysis to enhance our understanding of simultaneous interpreting from a signed language into a spoken language and issues around quality, cognitive overload, and strategies. This monograph makes an important contribution to literature in interpreting studies, and in sign language interpreting studies specifically. As the focus is on simultaneous interpreting from a signed into a spoken language, it extends our understanding of simultaneous interpreting processes and conference interpreting in general, and in particular the process of interpreting from a signed into a spoken language, which is an under-researched area. This book complements Jihong Wang’s other published work that has concentrated on interpreting from a spoken into a signed language (i.e., in the other language direction), so a monograph such as this makes a useful contribution to understanding her work, but also sign language interpreting more generally. This book is written in an engaging personal style that draws the reader in and is very accessible. For that reason, I think that it will be appealing to students studying sign language interpreting as well as interpreter practitioners, educators, and researchers. The structure is logical and helpful, moving from an overview of existing research/concepts to the methodology for the study, followed by dividing discussion of the data into four chapters giving a different focus for each chapter, and finishing with conclusions. Although his book focuses on Auslan (Australian Sign Language)-English interpreting, the findings and discussion are broadly applicable to other interpreting language combinations (spoken and signed).The book is situated primarily in interpreting studies, and primarily sign language interpreting studies, but will also be of interest to scholars in a wide range of cognate research areas, including sign language acquisition, sign language linguistics, deaf studies, and applied linguistics. Jemina Napier Sign language interpreter practitioner, educator, and researcher 27 August 2020

Acknowledgements

This book focuses on simultaneous interpreting from a signed language into a spoken language at a conference setting, an area that I do not work in but have research expertise. Signed language interpreting is somewhat similar to spoken language (e.g., Mandarin/English) simultaneous interpreting, a skill that I was trained in, use in my interpreting career, and teach to interpreting students. This book reports on new findings from my in-depth analysis of Auslan (Australian Sign Language)-to-English simultaneous interpretation data in my PhD research. I hope that the findings can broaden our understanding of the complexity and creativity in both spoken and signed language simultaneous interpreting. In addition to the people I particularly acknowledge here, I would like to thank all my family members, friends, and colleagues for their encouragement, support, and interest in my work. First and foremost, I would like to thank Mr Colin Allen for giving the Auslan presentation, which served as the source text for the signed-to-spoken language simultaneous interpreting task in this research project. I am also deeply grateful to the 31 professional Auslan/English interpreters for participating in my PhD research, for agreeing to be filmed and interviewed, and for allowing their Auslan-to-English simultaneous interpreting performance to be scrutinised by external raters, research assistants, and myself. It is the analysis of their Auslan-to-English simultaneous interpretation data and qualitative interview data that revealed so many interesting results in this book. I owe a debt of gratitude to Professor Jemina Napier, my principal supervisor during the PhD study (2009–2013), for her inspiration, ideas, guidance, research training, enthusiasm, positive attitude, and enormous support. My secondary supervisor, Dr Helen Slatyer, also made useful contributions to the development of assessment rubrics for evaluating the participants’ interpreting performance and provided tremendous support throughout my research. They have significantly influenced my academic journeys in numerous ways. I am also forever grateful to the three external raters – Professor Jemina Napier, Dr Della Goswell, and Andy Carmichael – for independently assessing the interpreters’ Auslan-to-English interpreting performance and agreeing to be recorded during the assessment process. Their scores and oral comments for the participants enabled me to have a good understanding of the participants’ interpreting performance and interpreting process.

Acknowledgements xv I also extend my deep appreciation for Professor Brenda Nicodemus, who invited me to Gallaudet University to work as a visiting scholar during August and December 2019, when I worked intensely on key chapters of this book. I would like to give special thanks to Brenda for mentoring me, sharing wisdom and brilliant ideas, providing detailed useful feedback on the whole book except for Chapters 2 and 8, inspiring me to persevere to achieve my goals, being my role model in so many ways, and for being a close friend. Many particular thanks go to Dr Lori Whynot, Neil Ray, Dr Danielle Hunt, Dr Emily Shaw, Dr Pamela Collins, Dr Paul Harrelson, Professor Paul Dudis, Dr Steven Collins, and Ms Stephanie Roche at Gallaudet University for their strong support, friendship, and interest in my research. I would also like to give special thanks to Ms Phyllis Rogers, an experienced American Sign Language (ASL)/English interpreter, for sharing her insights on simultaneous interpreting from a signed language into a spoken language at formal settings. In addition, I am grateful to Professor Xiaoyan Xiao for igniting my interest in exploring spoken and signed language interpreting when I was a postgraduate student. I also very much appreciate Professor Daniel Gile’s useful comments on my Chapter 6 regarding analysis of failure sequence in simultaneous interpreting. Moreover, I would like to thank Teresa Cumpston-Bird, an experienced Auslan/English interpreter, for being my research assistant, enhancing my understanding of the Auslan source text, assisting me in evaluating the sentencelevel accuracy of the interpreters’Auslan-to-English simultaneous interpretations, and encouraging me throughout my research project. Sincere thanks also go to Dr Jing Fang for her unswerving friendship and encouragement, especially in difficult times. I am extremely grateful to the anonymous reviewers for providing constructive feedback on my book proposal and manuscript. A big thanks to Elysee Preposi and Helena Parkinson, two editors of Routledge, for editing and producing my book. Finally, my deepest gratitude goes to my parents Shuguo Wang and Taoxiu Li, my husband Baohong (Frankie) Zhang, son Leo Zhang, daughter Emily Zhang, and my parents in law Shilian Zhang and Yuezhao Huang for their love, understanding, patience, and support.

1 Setting the Stage

1.1 An Unforgettable Experience On 2 April 2017, Dr Robert Adam,1 a renowned Deaf2 scholar as well as qualified Deaf interpreter and translator, gave a keynote speech in British Sign Language (BSL) titled ‘Mind the Gap:What Is Missing for Deaf Interpreters and Translators?’ at the Second International Symposium on Signed Language Interpretation and Translation Research, at Gallaudet University, Washington, DC. While Robert acquired Australian Sign Language (Auslan) from his Deaf parents in Melbourne, he lives in London and uses BSL, which is markedly similar to both Auslan and New Zealand Sign Language (NZSL) in terms of vocabulary and grammar (Johnston & Schembri, 2007). The audience for his keynote address was comprised of hundreds of signed language interpreting practitioners, educators, researchers, and students from the United States and many other countries; numerous Deaf professionals from around the world; and two Mandarin/English interpreters (Professor Xiaoyan Xiao [another keynote speaker at the conference] and I) who were interested in research on signed language interpreting. A fascinating feat of linguistic engineering of various types of simultaneous interpreting work took place as Robert gave his keynote speech in BSL. Before further elaboration, some key terms are explained here. The phrase simultaneous interpreting refers to a mode of interpreting where an interpreter renders understood messages into the target language at almost the same time as the speaker or signer delivers the messages in the source language (Diriker, 2015; Napier, 2002). When engaged in simultaneous interpreting, an interpreter typically operates with processing time of some seconds in order to receive and understand sufficient source language information before expressing its meaning in the target language (Diriker, 2015; Napier, 2002; Pöchhacker, 2015; Wang, 2020).While the time interval between receiving a source message and conveying its meaning in the target language has been referred to as time lag, lag time, earvoice span (EVS) in spoken language interpreting or eye-voice span in signed language interpreting, and décalage in interpreting studies literature (e.g., Cokely, 1992a; Gile, 2009; Timarová, 2015; Wang, 2020), Haualand and Nilsson (2019, p. 43) maintained that these terms indicate that simultaneous interpreting takes

2 Setting the Stage time but do not explain the underlying reasons.Therefore, Haualand and Nilsson suggested the use of the term processing time because it makes people more aware of the precious time needed by interpreters to analyse source messages and convey their meaning in the target language. I use the term processing time throughout my book because it makes readers more conscious of the cognitive process of simultaneous interpreting, the interpreter’s processing speed, and cognitive load. As simultaneous interpreting is a rather demanding cognitive task, simultaneous interpreters typically work in pairs to support each other, and take turns to interpret for approximately 20 minutes (Gile, 2009; Liu, Schallert, & Carroll, 2004). At the aforementioned symposium, Mr Gino Gouby, an experienced Deaf interpreter sitting in the first row of the audience, simultaneously interpreted Robert’s BSL presentation into American Sign Language (ASL), working from his non-native language into his native language. Another qualified Deaf interpreter standing next to Robert on the stage shadowed Gino’s ASL. A pair of highly skilled hearing ASL/English interpreters sitting in the second row of the audience simultaneously transferred the ASL messages into spoken English, which was instantly transcribed into a written English text (captions) shown on a large screen. With some proficiency in Auslan at that time, I still had to rely on the hearing interpreters’ ASL to English simultaneous interpretation in order to fully access Robert’s BSL monologue.Through these interpreters, I could appreciate Robert’s humour and understand his research expertise regarding Deaf interpreters and translators. I smiled when he said that one of the two interpreters working from ASL into English happened to be a fellow Australian interpreting from her nonnative language into her native language, and therefore her Australian English accent would perfectly match his own. Robert utilised PowerPoint slides to share interesting stories and explain dense information such as five Articles in the United Nations Convention on the Rights of Persons with Disabilities (UNCRPD) that specifically mention sign language as part of Deaf people’s human rights. Moreover, he emphasised a key point in his speech: compared to hearing signed language interpreters, Deaf interpreters and translators still face serious gaps in professionalisation, formal training, certification, professional status, recognition, and employment opportunities. Robert inspired the audience to reflect on what needs to be done to bridge these gaps for Deaf interpreters and translators. Honestly, I could not judge the accuracy of the ASL-to-English interpretations without comparing the BSL source text with the ASL interpretations and the English target text. However, at the time, I did find the ASL/English interpreters’ spoken English renditions remarkably informative, idiomatic, and fluent. I wondered how these interpreters managed to provide accurate and effective interpretations when working simultaneously from a signed language into a spoken language at a formal setting. Later, an insider told me that those signed language interpreters who interpreted the keynote speeches at the symposium had been handpicked out of hundreds of local practitioners, because they were considered

Setting the Stage 3 as the strongest and most experienced practitioners working at academic conferences and other formal settings. Robert, the Deaf keynote speaker, also demystified the smooth interpreting process by revealing to me his best practice of working with these Deaf and hearing signed language interpreters who interpreted his BSL presentation: (i) he had sent his PowerPoint slides to all the interpreters prior to the conference so that they could prepare for this interpreting assignment; (ii) before his presentation, he had met these interpreters for about 30 minutes to answer their questions and rehearse his entire presentation; (iii) during his signed presentation, he intentionally maintained eye contact with the interpreters to monitor if they could understand him and follow him; and (iv) he paused at certain points for the interpreters who were still interpreting to catch up with him. This authentic interpreted event suggests that a successful simultaneous interpretation from a signed language into a spoken language on a formal occasion requires a minimum of three conditions: (i) conference organisers need to select signed language interpreters who are the most competent and suitable for the particular (Deaf or hearing) signer, the content, and the setting; (ii) the signer and the interpreters need to work together to prepare for the interpreted event, as well as (iii) the signer and the interpreters need to cooperate with one another during the interpreted signed presentation. However, due to the inherent difficulties of simultaneous interpreting (e.g., time constraints, concurrent processing), having all three conditions may not guarantee that even a highly competent interpreter can render signed information into spoken language effectively and smoothly throughout the presenter’s signed monologue. For example, despite receiving adequate preparation materials and having extensive experience in various fields, a spoken or signed language interpreter may still encounter scenarios in which he or she does not understand a technical term or a phrase in the source text or struggle with finding an appropriate target language word to convey an understood concept (Gile, 2009, p. 191). In such situations, the interpreter may rely on a team interpreter’s support to deal with the particular obstacle, then move on to the next segment of the source text. This book focuses on investigating quality assessment, cognitive processes, and effective strategies in simultaneous interpreting from a signed language (Auslan) into a spoken language (English) at formal settings. Solid empirical evidence affirms that both spoken and signed language simultaneous interpreting are highly complex and challenging cognitive activities (Christoffels, 2004; Liu et al., 2004; Macnamara, Moore, Kegl, & Conway, 2011; Timarová et al., 2014; Wang, 2016). Yet, it still remains a mystery as to how the best spoken and signed language interpreters can calmly and simultaneously transfer source language messages into target language renditions that are faithful in meaning, idiomatic in target language use, and fluent in delivery. How do interpreters’ brain (‘black box’) function when they are engaged in simultaneous interpreting? What strategies do practitioners employ to cope with challenges in simultaneous interpreting? What guiding principles regarding attention management are universal to all simultaneous interpreters, and what strategies are language-pair specific?

4 Setting the Stage How can interpreting students, less experienced interpreters, interpreting educators, and researchers learn from trained, highly skilled simultaneous interpreters? These broad and marvellously complex questions require researchers from a wide range of disciplines (e.g., interpreting studies, cognitive science, bilingualism, applied linguistics) to collaborate, conduct interdisciplinary studies, and make ongoing endeavours in search of answers. As one may suspect, these intriguing questions go beyond the scope of this book; however, it is hoped that this book will shed some new light on these broad issues by reporting on findings from detailed linguistic and cognitive analyses of an experimental corpus of professional signed language interpreters’ simultaneous interpretations from a signed language (Auslan) into a spoken language (English) at a mock conference setting.

1.2 Interpreting Deaf Professionals’ Signed Monologues into Spoken Languages Although interpreting from a signed language into a spoken language has been variously referred to as voice-over, voicing, reverse interpreting, reading back, working into English, doing a voice-over (Nilsson, 2016, p. 20), voice-over interpreting, and signto-voice interpreting (Pöchhacker, 2016, p. 18), recently Haualand and Nilsson (2019, p. 43) argued that the outdated term voicing focuses on the modality of voice and tends to ‘hide the work the interpreter is doing’ and thus should be changed to the term interpreting from a signed language into a spoken language. I assert the use of this new term throughout my book because it not only precisely describes what the interpreter does when interpreting a Deaf or hearing signer’s signed messages into a spoken language for a hearing audience, but it also paves the way for in-depth discussions about bimodality and cognitive processing in this language direction. In addition, regarding directionality (an emerging topic in signed language interpreting studies), the term interpreting from a signed language into a spoken language reminds signed language interpreting practitioners, researchers, educators, and students of the other language direction, namely interpreting from a spoken language into a signed language. Furthermore, the consistent use of the term probably inspires spoken language interpreters to recognise that signed language interpreting is to some extent similar to spoken language interpreting. The broader term signed language interpreting (SLI) is used to refer to meaning transfer that occurs between a spoken language and a signed language or between two different signed languages (Napier & Leeson, 2015). Practitioners who ­provide signed language interpreting services are commonly referred to as sign – or signed – language interpreters. Sign – or signed – languages vary from country to country; they are naturally occurring, visual-spatial languages distinct from, but related to, the spoken or written languages of the countries where they are used (Johnston & Schembri, 2007; Napier & Leeson, 2015; Sandler & Lillo-Martin, 2006). Members of the Deaf community use signed languages as their first, dominant, or preferred language of communication. As a result, when Deaf

Setting the Stage 5 people and hearing non-signers attempt to communicate with each other, they may choose to rely on signed language interpreters who serve as bridges and mediate language barriers by transferring signed messages into a spoken language and vice versa (De Meulder & Haualand, 2019). This research monograph examines the quality, cognitive overload, and strategies in professional signed language interpreters’ simultaneous interpreting from Auslan into spoken English in a mock conference situation. Although numerous studies have investigated spoken language simultaneous (conference) interpreting (see Gile, 2009; Liu et al., 2004; Pöchhacker, 2016) and simultaneous interpreting from spoken languages into signed languages (see Cokely, 1992b; McDermid, 2018; Napier, 2002; Taylor, 1993), very few studies have examined simultaneous interpreting from signed languages into spoken languages (see Best, Napier, Carmichael, & Pouliot, 2016; Kalata-Zawłocka & van den Bogaerde, 2016; Napier, 2007; Taylor, 2002), let alone from Auslan into English. While some esteemed scholars who are also experienced signed language interpreters and educators (e.g., Napier & Leeson, 2015; Nilsson, 2016; Padden, 2000) have identified various challenges in working from signed languages into spoken languages (such as using a linear spoken language to reformulate a visual picture described in a signed language), few empirical studies have been undertaken to explore how interpreters cope with these challenges in order to produce accurate and coherent spoken language renditions. Interpreting for Deaf professionals and signed language conference interpreting emerge as two new research trends in signed language interpreting studies. Increasingly, Deaf individuals are pursuing higher education and achieving university degrees (e.g., bachelor’s degree, master’s degree, PhD), in part because more support services are now available to them (Napier, 2011). As a result, more Deaf people are attaining professional, managerial, and leadership roles in society at various levels (e.g., academics, engineers, lawyers, and healthcare providers) (Dickinson, 2017; Harrelson, 2019; Hauser, Finch, & Hauser, 2008; Napier & Leeson, 2015). Deaf people continue their services as leaders in national and international Deaf advocacy organisations, such as the national associations of the Deaf in many countries, and the World Federation of the Deaf. An increasing number of recent studies have explored signed language interpreting involving Deaf professionals, discussing topics such as (i) how Deaf PhD students and Deaf academics work with their interpreters in such situations as PhD thesis defences, signed or spoken lectures, workshops, and social functions (Burke & Nicodemus, 2013; De Meulder, Napier, & Stone, 2018), (ii) cooperation strategies employed by Deaf professionals and their preferred interpreters at seminar presentations to ensure both effective interpretations and smooth ­ ­communication (Napier, 2007; Napier, Carmichael, & Wiltshire, 2008), (iii) Deaf leaders’ criteria for selecting signed language interpreters and their strategies for ­working with them (Haug et al., 2017), (iv) hearing consumers’ (mis)perceptions of Deaf museum lecturers based on listening to the spoken English interpretations

6 Setting the Stage of their signed monologues (Feyne, 2015, 2016), (v) ­designated interpreters for Deaf professionals (Hauser et al., 2008), (vi) ­interpreting for Deaf professionals at both professional workplaces and social networking situations (Dickinson, 2017; Goswell, Carmichael, & Gollan, 2008), as well as (vii) Deaf professionals’ perceptions on effective signed language interpreting services in the workplace (Harrelson, 2019). Interpreting for Deaf professionals often requires signed language interpreters to render a signed monologue into a formal speech for hearing non-signers. However, many signed language interpreters, particularly novices, perceive interpreting from a signed language into a spoken language as being more challenging than working from a spoken language into a signed language (Napier, Rohan, & Slatyer, 2005; Nicodemus & Emmorey, 2013). Indeed, the quality of signed-to-spoken language simultaneous interpretation can have a considerable impact on hearing consumers’ perceptions of Deaf professionals’ expertise, competence, intelligence, and communication skills (Feyne, 2015, 2016). Research is needed to identify specific difficulties for interpreters when transferring signed monologues into spoken languages and find strategies to overcome them. The significance of exploring the quality of interpreters’ spoken language renditions of Deaf professionals’ signed language monologues is evidenced in Nilsson’s (2016, p. 44) words: As deaf people in many countries increasingly gain access to higher education and obtain professional positions of various kinds, professionalising the voicing of deaf professionals continues to be of vital importance. Another new trend of signed language interpreting research is the examination of hearing and Deaf interpreters’ work at national and international conferences and other formal settings (e.g., university lectures, professional workplaces). These studies have investigated topics such as (i) preparation and communication strategies utilised by interpreters when working at international conferences (Sheneman & Collins, 2016), (ii) collaboration strategies adopted by the hearing and Deaf interpreting teams to provide relay interpretation (Stone & Russell, 2014), (iii) complexities of interpreting in multilingual international settings, and (iv) interpreting between International Sign and spoken English (Best et al., 2016; McKee & Napier, 2002; Rosenstock & Napier, 2016; Whynot, 2016). Nevertheless, few of these studies have specifically focused on the performance quality and cognitive process of conference-level simultaneous interpreting from a signed language into a spoken language.

1.3 Research Questions of This Study The research in this book investigated how well 20 professional Auslan/English interpreters transferred an Australian Deaf professional’s formal Auslan ­monologue into spoken English for hearing non-signers at a mock conference

Setting the Stage 7 setting. The book provides both a comprehensive overview of the relevant literature and my in-depth analyses of the experimental corpus of Auslan-toEnglish simultaneous interpretations. Aiming to bridge the gaps in the existing signed language interpreting studies literature, I conducted this study to answer the following key research questions: 1 What are the challenges in simultaneous interpreting from a signed language (Auslan) into a spoken language (English) at a formal setting? 2 What causes cognitive overload in simultaneous interpreting from a signed language into a spoken language? 3 What strategies do signed language interpreters employ to cope with the challenges and cognitive overload in simultaneous interpreting from a signed language into a spoken language? In this research, I explored Auslan-to-English simultaneous interpreting from linguistic, cognitive, and communicative perspectives. I compared the Deaf professional’s Auslan source text with the 20 professional interpreters’ English renditions sentence by sentence, in order to identify particular Auslan segments that caused many of the participants to make significant errors in their Auslanto-English simultaneous interpretations. I then investigated the practitioners’ cognitive overload in relation to the specific challenges by analysing their processing time through ELAN, which is a linguistic annotation software program developed for creating text annotations of audio and video files (Crasborn & Sloetjes, 2008). Cognitive (over)load is a complex concept in translation and interpreting studies, and cognitive science, without a unified definition. Cognitive overload (cognitive saturation) in simultaneous interpreting refers to moments when an interpreter’s (total or specific) available processing capacity falls short of the (total or specific) required processing capacity, thus leading to errors or problems in the interpreter’s simultaneous interpreting performance (Gile, 2009, pp. 169–171). According to Gile (2009, p. 159), ‘when the processing capacity available for a particular task is insufficient, performance deteriorates.’ Note that alternative names for processing capacity include attention, attentional resources, mental energy, and effort (Gile, 2009, p. 159).While many scholars have examined cognitive overload (cognitive saturation) in spoken language simultaneous interpreting (Seeber, 2013; Seeber & Kerzel, 2012), few researchers have investigated cognitive overload in signed language simultaneous interpreting. My scrutiny of the moments when signed language interpreters experienced cognitive overload in Auslan-to-English simultaneous interpreting provided insight into the relationship between cognitive processing and performance quality of signed language simultaneous interpreting. Further, I explored how some practitioners in this study successfully coped with those challenges in the Auslan-to-English simultaneous interpreting task, specifically focusing on strategies and tactics that they utilised to deal with the difficulties. These analyses culminated in a list of key knowledge and skills that signed language interpreters

8 Setting the Stage need in order to perform well when rendering a signed monologue into a spoken language at formal settings. The primary goal of this book is to report on findings regarding the quality of professional signed language interpreters’ Auslan-to-English simultaneous interpreting performance as assessed by three external evaluators. A second goal is to contribute to the understanding of both challenges and cognitive overload in simultaneous interpreting from a signed language into a spoken language. Finally, this monograph aims to create new knowledge about strategies, tactics, and skills that are needed for simultaneously interpreting Deaf professionals’ signed monologues into spoken languages. The overarching theme throughout this book addresses the need to establish what really happens when signed language interpreters render a Deaf professional’s signed monologue into a spoken language for hearing non-signers at a formal setting, and then harnesses that knowledge into signed and spoken language interpreting practice, training, and research to produce highly skilled and reflective interpreters for Deaf and hearing people. While this empirical research concentrates on signed language interpreting in a unique language pair (Auslan/English), in only one language direction (from Auslan into English), and involves only one type of source text (a signed monologue), this book discusses the research findings in relation to the relevant literature on signed and spoken language interpreting, and it explores how the results can be applied equally among signed language interpreters working in other language combinations, and possibly to spoken language interpreters.

1.4 The Significance of This Study As noted earlier, due to the growing number of Deaf professionals around the world, signed language interpreters are increasingly required to simultaneously render the Deaf professionals’ signed monologues into spoken languages for hearing non-signers in formal situations. However, scant research has investigated signed-to-spoken language simultaneous interpreting. Only a few textbooks on interpreting from signed languages into spoken languages have been written; and almost all of them focus on ASL-to-English interpreting (e.g., Kelly, 2004; Owen, 2012; Taylor, 2002) rather than other language pairs. In this way, this substantive research into the quality, cognitive overload, and strategies of simultaneous interpreting from a signed language (Auslan) into a spoken language (English) will contribute to signed language interpreting training. This study is also useful for understanding signed language comprehension problems, which have been identified as a major challenge in signed-to-spoken language simultaneous interpreting. Moreover, this research creates new knowledge about the relationship among cognitive processing, strategy use, and performance quality of simultaneous interpreting from a signed into a spoken language. This research also serves as an example of analysing a signed language interpretation corpus. Finally, this study adds to the larger body of research on spoken and signed language simultaneous (conference) interpreting.

Setting the Stage 9

1.5 Book Structure This book comprises eight chapters. This opening chapter primarily provides the background, research questions, and significance of the study. Chapter 2 gives an overview of previous research on various aspects of simultaneous interpreting such as numbers, processing time, and strategies. Chapter 3 describes the methodology of the present study such as the stimulus material, participants, quality assessment of Auslan-to-English simultaneous interpreting performance, and the system for analysing the data. Chapter 4 presents findings regarding the quality of the participants’ Auslanto-English simultaneous interpreting performance, their perceived challenges in working in this language direction, as well as their perceived strategies. Chapter 5 reports on quantitative results of the interpreters’ processing time regarding Auslan sentences containing numbers and negation. It also uses representative examples from the interpretation corpus to illustrate the consequences of excessively short and long processing time. Chapters 6 and 7 present findings regarding cognitive overload due to problem triggers such as numbers, end negation, syntactical differences between Auslan and English, as well as dense information in the source text. The chapters also incorporate typical examples from the interpretation data to illustrate the interpreters’ coping strategies. Chapter 8 ties it all together by summarising the key findings, discussing their implications for interpreting practice and interpreter education, acknowledging the limitations, providing a conclusion, and offering suggestions for further research.

Notes 1 For further information about Dr Adam’s recent research, please refer to his book (Adam, Stone, Collins, & Metzger, 2014) and visit https://www.ucl.ac.uk/dcal/people/core-team/robertadam. 2 Upper-case ‘Deaf ’ is used to describe people who use signed language as their first or preferred language and identify themselves as members of a linguistic and cultural minority group. In other words, the capitalised ‘Deaf ’ indicates being a member of the Deaf community and having a Deaf identity. Lower-case ‘deaf ’ is used to refer to individuals who have an audiologically defined hearing impairment but may not perceive themselves as members of the Deaf community. This distinction is in line with the widely recognised convention used by scholars in signed language interpreting studies and Deaf studies (see Adam et al., 2014; Napier, 2002; Nicodemus, 2009).

References Adam, R., Stone, C., Collins, S. D., & Metzger, M. (Eds.). (2014). Deaf Interpreters at Work: International Insights. Washington, DC: Gallaudet University Press. Best, B., Napier, J., Carmichael, A., & Pouliot, O. (2016). From a Koine to Gestalt: Critical points and interpreter strategies in interpretation from International Sign into spoken English. In R. Rosenstock & J. Napier (Eds.), International Sign: Linguistic, Usage and Status Issues (pp. 136–166). Washington, DC: Gallaudet University Press.

10 Setting the Stage Burke, T. B., & Nicodemus, B. (2013). Coming out of the hard of hearing closet: Reflections on a shared journey in academia. Disability Studies Quarterly, 33(2). Christoffels, I. K. (2004). Cognitive Studies in Simultaneous Interpreting. (Unpublished doctoral dissertation). University of Amsterdam, Amsterdam, Netherlands. Cokely, D. (1992a). The effects of lag time on interpreter errors. In D. Cokely (Ed.), Sign Language Interpreters and Interpreting (pp. 39–69). Burtonsville, MD: Linstok Press. Cokely, D. (1992b). Interpretation: A Sociolinguistic Model. Burtonsville, MD: Linstok Press. Crasborn, O., & Sloetjes, H. (2008). Enhanced ELAN functionality for sign language corpora. In Proceedings of LREC 2008, Sixth International Conference on Language Resources and Evaluation (pp. 39–43). (Marrakesh, Morocco, 28–30 May 2008). De Meulder, M., & Haualand, H. (2019). Sign language interpreting services: A quick fix for inclusion? Translation and Interpreting Studies. doi: 10.1075/tis.18008.dem. De Meulder, M., Napier, J., & Stone, C. (2018). Designated or preferred? A deaf academic and two signed language interpreters working together for a PhD defense: A case study of best practice. International Journal of Interpreter Education, 10(2), 5–26. Dickinson, J. (2017). Signed Language Interpreting in the Workplace. Washington, DC: Gallaudet University Press. Diriker, E. (2015). Simultaneous interpreting. In F. Pöchhacker (Ed.), Routledge Encyclopedia of Interpreting Studies (pp. 382–385). New York: Routledge. Feyne, S. (2015). Typology of interpreter-mediated discourse that affects perceptions of the identity of Deaf professionals. In B. Nicodemus & K. Cagle (Eds.), Signed Language Interpretation and Translation Research: Selected Papers from the First International Symposium (pp. 49–70). Washington, DC: Gallaudet University Press. Feyne, S. (2016). Impact of ideology and interpreter-mediated discourse on perceptions of identity of Deaf professionals. In A. Kalata-Zawłocka & B. van den Bogaerde (Eds.), To Say or Not to Say – Challenges of Interpreting from Sign Language to Spoken Language. Proceedings of the 23rd efsli Conference in Warsaw, Poland, 11th–13th September 2015 (pp. 49–57). European Forum of Sign Language Interpreters (efsli). Gile, D. (2009). Basic Concepts and Models for Interpreter and Translator Training (Rev. ed.). Amsterdam/Philadelphia: John Benjamins. Goswell, D., Carmichael, A., & Gollan, S. (2008). Lights, camera … interpretation! In P. C. Hauser, K. L. Finch, & A. B. Hauser (Eds.), Deaf Professionals and Designated Interpreters: A New Paradigm (pp. 196–209). Washington, DC: Gallaudet University Press. Harrelson, P. (2019). Deaf employees’ perspectives on effective American Sign Language-English interpreting in the workplace:An investigation using the critical incident technique. (Unpublished doctoral dissertation). Gallaudet University, Washington, DC, United States. Haualand, H., & Nilsson, A.-L. (2019). Working with active interpreters: A commentary about interpreting terminology and concepts. International Journal of Interpreter Education, 11(2), 40–45. Haug,T., Bontempo, K., Napier, J., Nicodemus, B., van den Bogaerde, B., &Vermeerbergen, M. (2017). Deaf leaders’ strategies for working with signed language interpreters: An examination across seven countries. Across Languages and Cultures, 18(1), 107–131. doi: 10.1556/084.2017.18.1.5. Hauser, P. C., Finch, K. L., & Hauser, A. B. (Eds.). (2008). Deaf Professionals and Designated Interpreters: A New Paradigm. Washington, DC: Gallaudet University Press. Johnston, T., & Schembri, A. (2007). Australian Sign Language (Auslan): An Introduction to Sign Language Linguistics. Cambridge, UK: Cambridge University Press.

Setting the Stage 11 Kalata-Zawłocka, A., & van den Bogaerde, B. (Eds.). (2016). To Say or Not to Say – Challenges of Interpreting from Sign Language to Spoken Language. Proceedings of the 23rd efsli Conference in Warsaw, Poland, 11th–13th September 2015. European Forum of Sign Language Interpreters (efsli). Kelly, J. (2004). ASL-to-English Interpretation: Say It Like They Mean It. Alexandria, VA: Registry of Interpreters for the Deaf (RID) Press. Liu, M., Schallert, D. L., & Carroll, P. J. (2004). Working memory and expertise in simultaneous interpreting. Interpreting, 6(1), 19–42. doi: 10.1075/intp.6.1.04liu. Macnamara, B. N., Moore, A. B., Kegl, J. A., & Conway, A. R. A. (2011). Domain-general cognitive abilities and simultaneous interpreting skill. Interpreting, 13(1), 121–142. doi: 10.1075/intp.13.1.08mac. McDermid, C. (2018). Learning to Interpret: Working from English into American Sign Language. New York: Rochester Institute of Technology (RIT) Press. McKee, R., & Napier, J. (2002). Interpreting into International Sign Pidgin: An analysis. Journal of Sign Language and Linguistics, 5(1), 27–54. Napier, J. (2002). Sign Language Interpreting: Linguistic Coping Strategies. Coleford, Gloucestershire: Douglas McLean. Napier, J. (2007). Cooperation in interpreter-mediated monologic talk. Discourse & Communication, 1(4), 407–432. doi: 10.1177/1750481307082206. Napier, J. (2011). Signed language interpreting. In K. Malmkjær & K. Windle (Eds.), The Oxford Handbook of Translation Studies (pp. 357–376). Oxford/New York: Oxford University Press. Napier, J., Carmichael, A., & Wiltshire, A. (2008). Look-pause-nod: A linguistic case study of a deaf professional and interpreters working together. In P. C. Hauser, K. L. Finch, & A. B. Hauser (Eds.), Deaf Professionals and Designated Interpreters: A New Paradigm (pp. 22–42). Washington, DC: Gallaudet University Press. Napier, J., & Leeson, L. (2015). Signed language interpreting. In F. Pöchhacker (Ed.), Routledge Encyclopedia of Interpreting Studies (pp. 376–381). New York: Routledge. Napier, J., Rohan, M., & Slatyer, H. (2005). Perceptions of bilingual competence and preferred language direction in Auslan/English interpreters. Journal of Applied Linguistics, 2(2), 185–218. doi: 10.1558/japl.2005.2.2.185. Nicodemus, B. (2009). Prosodic Markers and Utterance Boundaries in American Sign Language Interpretation. Washington, DC: Gallaudet University Press. Nicodemus, B., & Emmorey, K. (2013). Direction asymmetries in spoken and signed language interpreting. Bilingualism: Language and Cognition, 16(3), 624–636. doi: 10.1017/S1366728912000521. Nilsson, A.-L. (2016). Interpreting from signed language into spoken language: The skills and knowledge needed to succeed. In A. Kalata-Zawłocka & B. van den Bogaerde (Eds.), To Say or Not to Say – Challenges of Interpreting from Sign Language to Spoken Language. Proceedings of the 23rd efsli Conference in Warsaw, Poland, 11th–13th September 2015 (pp. 15–48). European Forum of Sign Language Interpreters (efsli). Owen, A. (2012). Say What You See. Sign Language Interpreting: Voice-Over. London: The Wakeman Trust. Padden, C.A. (2000). Simultaneous interpreting across modalities. Interpreting, 5(2), 169–185. doi: 10.1075/intp.5.2.07pad. Pöchhacker, F. (2015). Modes. In F. Pöchhacker (Ed.), Routledge Encyclopedia of Interpreting Studies (pp. 268–269). New York: Routledge.

12 Setting the Stage Pöchhacker, F. (2016). Introducing Interpreting Studies (2nd ed.). London/New York: Routledge. Rosenstock, R., & Napier, J. (Eds.). (2016). International Sign: Linguistic, Usage and Status Issues. Washington, DC: Gallaudet University Press. Sandler, W., & Lillo-Martin, D. (2006). Sign Language and Linguistic Universals. New York: Cambridge University Press. Seeber, K. G. (2013). Cognitive load in simultaneous interpreting: Measures and methods. Target, 25(1), 18–32. doi: 10.1075/target.25.1.03see. Seeber, K. G., & Kerzel, D. (2012). Cognitive load in simultaneous interpreting: Model meets data. International Journal of Bilingualism, 16(2), 228–242. doi: 10.1177/1367006911402982. Sheneman, N., & Collins, P. F. (2016). The complexities of interpreting international conferences: A case study. In R. Rosenstock & J. Napier (Eds.), International Sign: Linguistic, Usage and Status Issues (pp. 167–191). Washington, DC: Gallaudet University Press. Stone, C., & Russell, D. (2014). Conference interpreting and interpreting teams. In R. Adam, C. Stone, S. D. Collins, & M. Metzger (Eds.), Deaf Interpreters at Work: International Insights (pp. 140–156). Washington DC: Gallaudet University Press. Taylor, M. M. (1993). Interpretation Skills: English to American Sign Language. Edmonton, Alberta: Interpreting Consolidated. Taylor, M. M. (2002). Interpretation Skills: American Sign Language to English. Edmonton, Alberta: Interpreting Consolidated. Timarová, Š. (2015). Time lag. In F. Pöchhacker (Ed.), Routledge Encyclopedia of Interpreting Studies (pp. 418–420). New York: Routledge. Timarová, Š., Čeňková, I., Meylaerts, R., Hertog, E., Szmalec, A., & Duyck, W. (2014). Simultaneous interpreting and working memory executive control. Interpreting, 16(2), 139–168. doi: 10.1075/intp.16.2.01tim. Wang, J. (2016). The relationship between working memory capacity and simultaneous interpreting performance: A mixed methods study on professional Auslan/English interpreters. Interpreting, 18(1), 1–33. doi: 10.1075/intp.18.1.01wan. Wang, J. (2020). Striking a cognitive balance: Processing time in Auslan-to-English simultaneous interpreting. In D. Hunt & E. Shaw (Eds.), The Second International Symposium on Signed Language Interpretation and Translation Research: Selected Papers (pp. 108–131). Washington, DC: Gallaudet University Press. Whynot, L. (2016). Understanding International Sign: A Sociolinguistic Study. Washington, DC: Gallaudet University Press.

2 Overview

This chapter sets the scene by providing a comprehensive overview of the relevant research on spoken and signed language interpreting that has informed the present study. The review focuses on the following topics: (i) directionality in signed language interpreting, (ii) simultaneous interpreting from a signed into a spoken language, (iii) cognitive load in simultaneous interpreting, (iv) Gile’s (2009) Effort Model of spoken language simultaneous interpreting, (v) adapting Gile’s Effort Model to signed-to-spoken language simultaneous interpreting, (vi) processing time in simultaneous interpreting, (vii) numbers as a problem trigger in simultaneous interpreting, and (viii) strategies in simultaneous interpreting.

2.1 Directionality in Signed Language Interpreting This section briefly describes some salient features of signed language interpreting, then gives an overview of previous studies on directionality in signed language interpreting. The most obvious difference between signed language interpreting and spoken language interpreting is modality (Kellett Bidoli, 2002; Napier, McKee, & Goswell, 2010; Nicodemus & Emmorey, 2013). Spoken language (unimodal) interpreters work between different spoken languages that are both perceived and produced in the auditory-verbal modality. In stark contrast, hearing signed language (bimodal) interpreters work between a spoken language and a signed language that are perceived and produced in different modalities.When working from a signed language into a spoken language, the interpreter needs to comprehend a visual picture comprised of manual signs, facial expressions, mouth movements, and body movement, decide whether it is necessary or natural to render all the visual-spatial details into a spoken language for the target audience, and express the meaning of the visual picture in the spoken language by uttering one word after another (Napier, 2011; Nilsson, 2016). Signed language interpreters work predominantly in simultaneous mode, not only because there is no acoustic overlap between a signed (visual-spatial) language and a spoken (­auditory-verbal) language, but also because simultaneous interpreting is more time-efficient than consecutive interpreting (Diriker, 2015; Napier, 2002; Napier & Leeson, 2015; Pöchhacker, 2015).

14 Overview Given that the majority of signed language interpreters have learned a signed language as a non-native language and most signed language interpreting work is from a spoken language into a signed language, signed language interpreters work primarily from their native language (A, or L1) into their non-native ­language (B, or L2).1 Only a small percentage (approximately 10%) of signed language interpreters have acquired a signed language from signing Deaf parents and have both a spoken language and a signed language as their native languages (Napier & Leeson, 2015; Napier, Rohan, & Slatyer, 2005; Nicodemus & Emmorey, 2013). Some empirical studies have explored directionality in signed language interpreting. Nicodemus and Emmorey’s (2013) large-scale survey study revealed that 49.6% of spoken language interpreters (179 out of 361) preferred working from their non-dominant language (B) into their native language (A), but 63.9% of ASL/English interpreters (228 out of 357) favoured working from A (English) into B (ASL). Compared with expert ASL/English interpreters who had ten or more years of full-time interpreting experience, novice ASL/English interpreters who had three years or less of full-time interpreting experience reported a stronger preference for A-to-B interpreting. Interestingly, 63.6% (28 out of 44) of ASL/English interpreters who were balanced bilinguals reported no directionality preference. However, signed language interpreters’ directionality preference may contradict their actual interpreting performance. Nicodemus and Emmorey (2015) asked 30 ASL/English interpreters, all native users of English and non-native users of ASL who were divided into 15 novices who had five years or less of full-time professional interpreting experience and 15 experts who had ten years or more of full-time professional interpreting experience, to complete simultaneous interpreting tasks from English into ASL and vice versa. The majority of the novices preferred working from English (A) into ASL (B), whereas the majority of the experts reported no directionality preference. Two native speakers of English and two highly fluent ASL signers (none were professional interpreters) assessed participants’ simultaneous interpreting performances in terms of accuracy of interpretation and articulation quality (including flow, production speed, and use of prosodic features). Results revealed that the novices were significantly more accurate, exhibited better presentation speed, and used more natural prosody when working from ASL (B) into English (A) than vice versa, and that their flow was similar in both directions. In contrast, the expert interpreters’ English-to-ASL (A-to-B) simultaneous interpreting performance was as good as their ASL-to-English (B-to-A) simultaneous interpreting performance in terms of accuracy, flow, presentation speed, and prosodic quality. These findings indicate that novice signed language interpreters achieve better simultaneous interpreting performance when working from their non-native language into their native language than in the opposite direction, and that ten or more years of full-time professional interpreting work mitigates directionality effect on simultaneous interpreting performance to some extent.

Overview 15 Based on their findings in the above two studies, Nicodemus and Emmorey (2015) maintained that the dissonance between novice signed language interpreters’ directionality preference and actual simultaneous interpreting performance may be attributable to the following three factors. First, when working from a spoken language into a signed language, novices may rely heavily on using fingerspelling to cope with situations in which they do not know the equivalent sign for a spoken word. Second, when interpreting from a spoken language into a signed language, novices may frequently use transcoding, namely using a predominantly literal interpretation approach. However, if interpreters use transcoding when working from a signed language into a spoken language, it is highly likely that their target language renditions are incomprehensible, ungrammatical, and unacceptable to the hearing audience. Third, when working from a spoken language into a signed language, novices cannot effectively monitor their signed language output, but they can hear themselves when working in the other language direction, detect errors, and make self-corrections. Nicodemus and Emmorey also emphasised that these explanations were closely related to modality differences between a spoken language and a signed language, a challenge for signed language interpreters. In Wang and Napier’s (2015) experimental study on directionality in signed language interpreting, the majority of 14 professional Auslan/English interpreters who were balanced bilinguals (native signers) reported no directionality preference, and slightly over half of 17 professional Auslan/English interpreters who were unbalanced bilinguals (non-native signers) also expressed no interpreting directionality preference. Results revealed that the native signers were similar to the non-native signers in terms of both overall English-to-Auslan simultaneous interpreting performance and overall Auslan-to-English simultaneous interpreting performance. However, regarding sub-scores of English-to-Auslan simultaneous interpreting performance, the 14 native signers were significantly better than the 17 non-native signers in terms of both target text features (e.g., vocabulary, grammar, register, affect) and delivery features (e.g., clarity of articulation, fluency, prosody), indicating that the native signers were significantly better than the non-native signers in terms of Auslan productive skills. In addition, as expected, the native signers’ English-to-Auslan simultaneous interpreting performance was similar to their Auslan-to-English simultaneous interpreting performance. Although the non-native signers’ Auslan-to-English (B-to-A) simultaneous interpreting performance was also comparable to their English-to-Auslan (A-to-B) simultaneous interpreting performance in terms of overall score, the former was significantly better than the latter in terms of target text features.This finding suggests that the non-native signer interpreters needed to continue to improve their signed language (non-native language) productive skills, especially the ability to use grammar, vocabulary, register, and affect in their signed language appropriately. When replicating the aforementioned analyses on 11 native signers and 11 nonnative signers who all had ten or more years of interpreting experience,Wang and Napier found that all the aforementioned result patterns remained the same, except that the 11 native signers were similar to the 11 non-native signers in terms of both target text features and delivery features of English-to-Auslan simultaneous

16 Overview interpretation. Interestingly, in relation to both accuracy and articulation/delivery features, Wang and Napier’s (2015) directionality findings regarding the 11 expert non-native signer interpreters were in line with Nicodemus and Emmorey’s (2015) directionality findings regarding the 15 expert non-native signer interpreters, despite some methodological differences in ­quality assessment. In addition, van Dijk, Boers, Christoffels, and Hermans (2011) examined directionality effect on 25 experienced interpreters’ simultaneous interpretation between Sign Language of the Netherlands (SLN) and Dutch. Ten of the interpreters were native signers of SLN, and the other 15 were non-native users of SLN. Results revealed that the native signers were similar to the non-native signers in terms of both SLN-to-Dutch simultaneous interpreting performance and Dutch-to-SLN simultaneous interpreting performance. Moreover, for both the native signers and the non-native signers, Dutch-to-SLN simultaneous interpreting performance was significantly better than SLN-to-Dutch simultaneous interpreting performance. van Dijk et al. explained that this finding may be due to two factors: first, the interpreters may have experienced severe difficulties in comprehending the SLN source text; second, the interpreters had more experience in working from Dutch into SLN than vice versa. In summary, previous experimental studies on directionality in signed language interpreting have yielded inconsistent results, which may be due to differences in participant profile, sample size, and interpreting performance assessment method, among other factors. More research on directionality in spoken and signed language interpreting is needed.

2.2 Simultaneous Interpreting from a Signed Language into a Spoken Language Interpreting from a signed into a spoken language is an under-researched area. In her seminal work, Taylor (2002) analysed a large corpus of ASL-to-English interpretations produced by certified ASL/English interpreters, uncertified practitioners, and interpreting students in Canada and the United States. She asked an expert panel made up of Deaf and hearing ASL/English interpreter educators to assess the ASL-to-English interpretations line by line and provide comments on the interpreting performance. As a result, she identified the following six major categories of necessary ASL-to-English interpreting skills: 1 Comprehension of ASL lexicon: (i) fingerspelling, (ii) numbers, (iii) ASL lexicon, (iv) classifiers (SASS), (v) non-manual markers, (vi) negation, (vii) emphasis, (viii) modulation, (ix) temporal aspects, (x) distribution of an action, and (xi) borrowed vocabulary. 2 Comprehension of ASL discourse: (i) facial markers, (ii) ASL referencing, (iii) spatial relationships, (iv) utterance boundaries, and (v) hand dominance. 3 Production of English lexicon: (i) interpretation of name signs, (ii) acronyms and abbreviations, (iii) interpretation of ASL referencing, (iv) vocabulary range, and (v) grammatical accuracy.

Overview 17 4 Production of English discourse: (i) ASL intrusions, (ii) English discourse strategies, (iii) register, (iv) cohesion markers, (v) flow, (vi) emphasis, (vii) repairs, and (viii) managing the interpretation process. 5 Delivery or public speaking: (i) breathing, (ii) voice projection, (iii) pausing, (iv) pace, (v) pitch, and (vi) articulation. 6 Composure and appearance: (i) minimising unnecessary sounds, (ii) expression, (iii) eye contact, (iv) posture, (v) hands and body, and (vi) appearance. In addition to this comprehensive list of essential skills, Taylor proposed two concepts, namely knowledge-lean skills (basic and fundamental language skills) and knowledge-rich skills (context-sensitive interpreting skills that allow an interpreter to convey the nuances in meaning and tone that a signer is expressing). In relation to ASL-to-English interpreting, knowledge-lean skills include: (i) ASL comprehension at the lexical level, (ii) ASL comprehension at the discourse level, and (iii) production of English lexicon. Knowledge-rich skills comprise: (i) production of English discourse, (ii) delivery of English speeches, and (iii) composure and appearance. Taylor’s taxonomy serves as a useful framework for analysing interpreting performance from a signed language into a spoken language. Feyne (2014, 2015, 2016) compared Deaf professionals’ perceptions of Deaf museum docents’ ASL lectures with hearing non-signing museum staff members’ perceptions of four trained and certified ASL/English interpreters’ spoken English renditions of the ASL lectures. The Deaf and hearing assessors used the following assessment criteria to evaluate the communicative competence and professionalism of the Deaf lecturers: (i) knowledge of topic, (ii) language use, (iii) communicative style, and (iv) museum calibre. The Deaf assessors perceived all the Deaf lecturers as confident and knowledgeable. In stark contrast, the hearing assessors often deemed the Deaf lecturers overly casual, not knowledgeable, not confident, unprofessional, and not satisfying the expectations of institutional talk in museum lectures. Another interesting finding is that the hearing assessors attributed almost all aspects of the spoken English renditions – such as improper register, fillers, vocal inflections, and prosodic markers – to the Deaf lecturers. This finding indicates that some of the hearing assessors may have mistakenly assumed that interpreters’ renditions are verbatim representations of source language utterances in terms of both content and manner. Furthermore, Feyne (2015, 2016) identified four aspects of signed-to-spoken language interpretations that most strongly affected the hearing assessors’ perceptions of the professionalism and calibre of the Deaf lecturers: 1 Target language utterances that violated the expected institutional norms such as precision in the use of spoken language for professional discourse, appropriate terminology for the particular setting, and formal register. 2 Literal interpretations2 of constructed dialogues in signed language that sounded casual for the target audience in the institutional setting.

18 Overview 3 Interpreters speaking in a manner that indicated a lack of confidence, specialised knowledge about the subject, and expertise. 4 Interpretation process artefacts such as interpreters’ errors, repetitions, false starts, repairs, hesitations, and disfluency that resulted in an impression that the Deaf lecturer was ill-prepared or unconfident. Feyne found that the hearing assessors perceived it very important for the interpreters to ensure that their spoken language renditions matched the expected communicative norms in the formal setting. Feyne’s findings indicate that both the content and delivery of interpreters’ signed-to-spoken language ­interpretations influence hearing non-signers’ perceptions of Deaf presenters’ professionalism, communication skills, and knowledgeability. In addition, Napier (2007), Napier, Carmichael, and Wiltshire (2008), as well as De Meulder, Napier, and Stone (2018) conducted case studies of real-life simultaneous interpreting from a signed language into a spoken language at a seminar and a PhD thesis defence. The Deaf professionals in both situations selected their preferred interpreters, either because they had worked together previously or because the interpreters were familiar with the subject matter and discourse environment. Results from both studies revealed that the interpreters worked closely with the Deaf presenter as a team before, during, and after the interpreting assignment. The interpreters proactively prepared for the assignment by reading preparation materials, summarising key points, familiarising with terminology, and writing down names that would be mentioned frequently. Before giving the signed presentation, the Deaf presenter briefed the interpreters about the content and context of the presentation and gave them relevant texts such as verbatim quotes and a script of the signed monologue. Apart from jointly discussing signs for specific technical terms, the Deaf presenter and the interpreters also established various visual cues to be used by them during the interpreting assignment to ensure good mutual communication. During the signed presentation, the Deaf presenter not only maintained eye contact with both interpreters, but also paused every now and then to give sufficient time for the active interpreter to decode and encode information and unpack particular concepts such as culturally bound concepts. The active interpreter nodded to the Deaf presenter to indicate that he or she had completed an interpretation; and the Deaf presenter nodded back before proceeding to the subsequent episode of signed discourse. The support interpreter laid out documents and held notes for the active interpreter, used a hand gesture to signal the Deaf presenter to pause and informed the Deaf presenter that the active interpreter was catching up, supported the active interpreter by offering prompts and compliments, and used the thumbs-up sign to reassure the Deaf presenter that the interpretation was accurate. Napier et al. (2008) coined the term ‘lookpause-nod’ to represent these visual cues or cooperation strategies that were used by the Deaf presenter and the interpreters. The Deaf presenter and the interpreters also held a debriefing session after the interpreting assignment.

Overview 19 Based on her findings, Napier (2007) argued that signed language interpreters and Deaf signers need to abide by the following six linguistic, communicative, and attitudinal maxims in order to cooperate with each other effectively in interpreter-mediated interactions, co-construct meaning, and achieve communicative goals: (i) trust, (ii) preparation, (iii) negotiation, (iv) eye contact, (v) turn-taking, and (vi) visual cues. De Meulder et al. (2018) concluded that, although the Deaf presenter and the two highly skilled signed language interpreters had not worked together before, both the Deaf presenter’s preference for selecting the particular interpreters and the interpreters’ familiarity with the Deaf presenter’s topic and domain were critical for the success of the trio working together. There are heated debates on whether it is better for Deaf professionals to work with designated interpreters or preferred interpreters (see Burke & Nicodemus, 2013; De Meulder et al., 2018; Dickinson, 2017; Harrelson, 2019; Hauser, Finch, & Hauser, 2008). Given the remarkable sign variation in the Deaf community and Deaf professionals’ distinct domains of professional expertise, the Deaf professionals may prefer to work with a small pool of professional signed language interpreters on a regular basis who possess the requisite knowledge and interpreting skills, and the Deaf professionals may proactively assume the responsibility for familiarising the interpreters with the technical terms in their specialised fields (Burke & Nicodemus, 2013; De Meulder et al., 2018; Dickinson, 2017; Harrelson, 2019). Further research is needed to expand our knowledge about how to interpret effectively for Deaf professionals. Additionally, Haug et al. (2017) interviewed two Deaf leaders in each of seven countries on their perceptions of signed language interpreters. The Deaf leaders reported that they used various criteria for selecting interpreters, including: 1 the interpreter’s expertise in both signed-to-spoken language interpreting and spoken-to-signed language interpreting, 2 the interpreter having good skills in signed language comprehension and production, 3 a sense of trust in the interpreter, 4 the interpreter’s knowledge set (e.g., familiarity with the setting, topics, technical terms) and willingness to provide contextual information, 5 the interpreter’s familiarity with the Deaf leader, 6 the interpreter transferring messages as closely as possible to the source text, 7 the interpreter having received formal interpreting training, 8 the interpreter being a native signer, and 9 the interpreter being actively involved in the Deaf community. The Deaf leaders also revealed that they used the following methods to assess the quality of interpreters’ signed-to-spoken language interpreting performance: 1 reviewing transcripts (e.g., the report system at the European Parliament) of the communication, 2 looking at a speech-to-text monitor to see what the interpreter said in the target language,

20 Overview 3 observing the facial expression of their hearing colleagues, 4 asking other interpreters to comment on the interpreter’s skills, 5 maintaining eye contact with the audience when giving a signed presentation, 6 asking the hearing audience if they understood the Deaf leader, 7 asking other people about the quality of signed-to-spoken language interpretation, 8 asking for confirmation if the interpreter actually said what the Deaf leader signed, 9 identifying mistakes based on the response of other participants in an interpreter-mediated interaction, 10 observing the body language of the interpreter, and 11 lip reading the interpreter. Moreover, the Deaf leaders reported that they employed the following strategies to work effectively with signed language interpreters: 1 adapting their signing style (e.g., sign more slowly, try to sign very clearly, repeat their signs, fingerspelling or sentences before going on in signing) to make sure that they were understood by the interpreter, 2 making regular eye contact with the interpreter to check if he or she could understand their signed discourse, 3 repeating their signs or rephrasing their signed messages if the interpreter could not understand their signing, 4 providing feedback to the interpreter, pointing out issues, and discussing how to resolve them, 5 giving feedback if the interpreter was willing to accept criticism and advice, 6 supporting the interpreter at the beginning but giving up due to too many problems occurring, 7 calling the interpreting agency for a replacement, 8 giving direct feedback and then not requesting the same interpreter, 9 providing the interpreter with technical terms in their signed language, and 10 directing the interpreter to the Deaf leader’s own website (containing videos of signing) so that the interpreter became familiar with their signing style. Haug et al.’s (2017) results indicate that some Deaf professionals take the initiative to collaborate with signed language interpreters in order to ensure that the interpreter-mediated communication proceeds smoothly and successfully. The findings also suggest that more work needs to be done in order to foster trust between Deaf professionals and signed language interpreters, improve interpreters’ signed language proficiency and free interpretation skills, and encourage honest conversations on how to cooperate effectively and provide and receive feedback appropriately. In summary, some empirical studies have found that professional or experienced signed language interpreters performed equally well in both language

Overview 21 directions, but others have revealed that signed language interpreters, particularly novices, performed significantly better when working from a spoken language into a signed language than vice versa. Previous research has also revealed that Deaf professionals and their preferred or designated interpreters employed cooperation strategies before, during, and after interpreted signed presentations in order to work seamlessly as a team.

2.3 Cognitive Load in Simultaneous Interpreting This section first introduces the term cognitive load, then reviews previous research on cognitive load in simultaneous interpreting, and finally summarises the key findings of studies on working memory and simultaneous interpreting. 2.3.1 Cognitive Load There is a broad consensus among interpreting practitioners and researchers that simultaneous interpreting is a highly complex language processing activity (Christoffels, 2004; Liu, Schallert, & Carroll, 2004; Nicodemus & Emmorey, 2013; Seeber, 2011;Timarová et al., 2014;Wang, 2016). Salient features of simultaneous interpreting include time pressure, concentration of attention, allocation of attention to concurrent sub-tasks, frequent switching of attention between the sub-tasks, ephemeral presentation of source language information, and almost immediate production of target language renditions. When engaged in simultaneous interpreting, an interpreter concurrently performs various sub-tasks such as listening to or watching the source language input, comprehending source language utterances,3 retrieving relevant knowledge from one’s long-term memory, temporarily storing information in working memory for later use, transferring meaning from the source language to the target language, reformulating the understood messages in the target language through speaking or signing, monitoring the target language output, and coordinating these parallel sub-tasks (Gile, 2009; Liu et al., 2004; Wang, 2013a). Cognitive research on simultaneous interpreting can enrich our understanding of both simultaneous interpreting process and language processing in general. The concept of cognitive load is a multidimensional construct used in psychology without a universal definition. It is based on the hypothesis that the human brain is not only limited in the number of cognitive operations it can carry out at the same time but also constrained in the amount of information it can keep available for cognitive processing at a particular juncture (Miller, 1956; Seeber, 2015). The term cognitive load has been used interchangeably with cognitive workload (Gevins & Smith, 2003), mental workload (Charles & Nixon, 2019; Galy, Cariou, & Mélan, 2012; Xie & Salvendy, 2000), mental effort (the cognitive capacity that an individual allocates to a task) (Galy et al., 2012; Seeber, 2013), and cognitive effort (Seeber, 2013). Mental workload influences an individual’s performance on complex tasks (Charles & Nixon, 2019; Xie & Salvendy, 2000). Cognitive load has been defined

22 Overview as ‘the cognitive demand of a task’ (Galy et al., 2012, p. 269), ‘the amount of mental work or effort necessary for a person or group to complete a task over a given period of time’ (Xie & Salvendy, 2000, p. 76), and ‘the load that performing a particular task imposes on the cognitive system of a particular learner’ (Paas & van Merriënboer, 1994, p. 122). Some researchers (e.g., Charles & Nixon, 2019; Xie & Salvendy, 2000) distinguish task workload (the objective processing load that a particular task imposes on a person) from mental workload (the perceived mental effort that the person allocates to the task). To put it differently, the same task workload may mean different mental workloads for different people. Cognitive load can be influenced by an array of factors such as task difficulty, temporal pressure, environmental variables, as well as people’s experience. Cognitive load has been categorised in various ways.Two types of cognitive load that are especially relevant to the present study are instantaneous workload (the load at a given point in time that fluctuates throughout a task) and peak workload (the highest value of instantaneous workload reached during a task) (for a description of other types of mental workload, see Xie & Salvendy, 2000, p. 88). Given that cognitive load ‘cannot be detected directly’ (Xie & Salvendy, 2000, p. 76), four methods can be used to measure it: analytical methods, subjective measures, performance measures, and psychophysiological measures (Galy et al., 2012; Paas,Tuovinen,Tabbers, & Van Gerven, 2003; Seeber, 2013). Each approach has its advantages and disadvantages that may limit its use for investigating cognitive load in simultaneous interpreting. Only the subjective and psychophysiological measures are discussed here because they are widely used and they are relevant to the present study. Subjective methods utilise self-ratings and self-reported data elicited from introspection, retrospection, and concurrent verbalisation (think-aloud protocol) to quantify participants’ perceived mental workload. Two subjective measures of mental workload are commonly used (Galy et al., 2012; Xie & Salvendy, 2000). One is the National Aeronautics and Space Administration Task Load Index (NASA-TLX) (Hart & Staveland, 1988) that evaluates six aspects of mental workload, including mental demand, physical demand, temporal demand, performance, frustration level, and effort. The other is the Subjective Workload Assessment Technique (SWAT) (Reid & Nygren, 1988) that assesses mental workload from three dimensions: time load, mental effort load, and psychological stress load. Both subjective measures are similar and mainly used in aeronautics. Weaknesses of the subjective methods include possible contamination of data caused by delayed recall (except in concurrent verbalisation), subjectivity in self-reports, and participants’ unconsciousness about automatic processing. In addition, the following six psychophysiological measures have been used to estimate cognitive load (Charles & Nixon, 2019): (i) electrocardiac activity (e.g., heart rate variability), (ii) respiration (e.g., respiratory rate), (iii) skin-based measures (e.g., electrodermal activity), (iv) blood pressure, (v) ocular measures (e.g., pupil dilation, blink rate), and (vi) brain measures (e.g., electroencephalogram [EEG], functional magnetic resonance imaging [fMRI]). The main disadvantage

Overview 23 of these psychophysiological measures is ‘the difficulty of identifying and determining what is actually measured’ (Seeber, 2013, p. 25). Seeber (2013) maintains that cognitive pupillometry using eye-tracking is one of the most promising methods to objectively measure local (online, instantaneous, or moment-by-moment) cognitive load during simultaneous interpreting, mainly because of its high temporal resolution, limited invasiveness, and affordability.The primary limitation of pupillometry is that it is suitable for short auditory stimuli (e.g., phrases and sentences) but much less suitable for long stimuli (e.g., discourse of several minutes). As a result, experiments that use pupillometry need to be carefully controlled and should use short periods of interest (Seeber, 2013). 2.3.2 Cognitive Load Models of Simultaneous Interpreting This section presents an overview of Seeber’s research on cognitive load in simultaneous interpreting. Seeber (2013, p. 19) defines cognitive load as ‘the amount of capacity the performance of a cognitive task occupies in an inherently capacity-limited system.’ Seeber (2013, p. 26) also points out that cognitive overload occurs ‘when the task exceeds the cognitive resources available to perform it.’ Based on Wickens’s (1984) Multiple Resource Model, which posits that performing two (or more) tasks at the same time demands more cognitive resources than performing either (or any) of the tasks separately, Seeber (2011) proposed theoretical analytical cognitive load models for simultaneous interpreting that illustrate the conjectured cognitive resource allocation when four different strategies (waiting, stalling, chunking, and anticipation) are used during simultaneous interpreting of syntactically asymmetrical languages such as German (a SOV structure) and English (a SVO structure). The four strategies are defined as follows. Waiting refers to a strategy by which an interpreter temporarily stops target language production in order to wait for more source language information (Seeber, 2011). Stalling is the strategy by which the interpreter produces a ‘neutral padding’ (padding phrase or filler which fills the gap but does not add any new meaning) while waiting for more source language information (Seeber, 2011). While both stalling and waiting enable the interpreter to buy time in order to receive more source language input, draw on the context to understand the meaning of source language messages, and plan target language production, waiting typically results in a period of silence in the interpreter’s target language speech, but stalling does not. Chunking (also called segmentation) refers to a process whereby the interpreter divides a source language sentence into shorter segments that can be processed without waiting for the whole sentence to end (Seeber, 2011). After using chunking, the interpreter needs to add cohesive devices in order to maintain cohesion and coherence in the target language speech. Anticipation refers to a strategy by which the interpreter predicts a sentence element before it is said by the source language speaker (Seeber, 2011).

24 Overview Seeber’s (2011) theoretical cognitive load models of simultaneous interpreting predict that, waiting, stalling, and chunking may lengthen the interpreter’s processing time and cause a spillover effect (i.e., a remarkable increase in cognitive load when the information the interpreter has been waiting for is finally received), which in turn may lead to interpretation breakdowns. His models also suggest that, when working between two languages that are syntactically asymmetrical, the amount of syntactical restructuring that interpreters engage in may increase the amount of cognitive load they experience during simultaneous interpreting. Moreover, Seeber and Kerzel (2012) conducted an experimental study to measure local cognitive load in German-to-English simultaneous interpreting. They used task-evoked pupillary responses to measure cognitive load experienced by ten professional interpreters who simultaneously rendered German sentences into spoken English. All participants had German as their C language (i.e., interpreters’ non-native passive language) and English as their native language, and they had a minimum of seven years of professional interpreting experience. A key finding is that the interpreters experienced significantly higher local cognitive load towards the end of sentences when processing asymmetrical syntax than when processing symmetrical syntax.This finding indicates that syntactical asymmetry between the source language and the target language significantly increases cognitive load in simultaneous interpreting. Another finding is that discourse context significantly reduced local cognitive load during simultaneous interpreting, especially towards the end of source language sentences which were syntactically asymmetrical to target language sentences.This finding highlights that it is important for interpreters to use context during simultaneous interpreting. Taken together, these empirical findings support Seeber’s (2011) analytical cognitive load models. However, it warrants attention that Seeber and Kerzel (2012) excluded from their data analysis source language sentences that participants had omitted due to excessive processing time; in my opinion, these sentences can shed light on why interpreters experienced cognitive overload. 2.3.3 Working Memory and Simultaneous Interpreting Another concept relevant to cognitive load is working memory. Baddeley (2003, p. 189) defines working memory as a limited capacity system that involves ‘the temporary storage and manipulation of information that is assumed to be necessary for a wide range of complex cognitive activities.’ Working memory capacity refers to the amount of information that a person can store temporarily while performing processing operations (Baddeley & Hitch, 2000). Some studies (e.g., Christoffels, De Groot, & Kroll, 2006) have found that professional interpreters have a significantly larger working memory capacity than interpreting students and/or noninterpreters, while other studies (e.g., Liu et al., 2004) have not found significant differences between professional interpreters, interpreting students, and noninterpreters in terms of working memory capacity. Some studies (e.g., Zhang, 2009) have found significant correlations between professional interpreters’

Overview 25 working memory capacity and simultaneous interpreting performance, while other studies (e.g., Wang, 2016) have not found such significant correlations. In view of the discrepancies in the existing results, Mellinger and Hanson (2019) conducted meta-analyses of a number of previous studies on working memory and simultaneous interpreting.Their analyses of the combined evidence revealed that professional simultaneous interpreters had greater working memory capacity than comparison groups. They also found an overall positive correlation between participants’ working memory capacity and simultaneous interpreting performance. Given the inconsistent findings and methodological differences in the previous studies on working memory and simultaneous interpreting, further research is needed in this area before clear conclusions can be drawn. The study in this book extends my doctoral research on the relationship between signed language interpreters’ working memory capacity and simultaneous interpreting performance (e.g., Wang, 2013a, 2016) in an attempt to investigate how the interpreters’ working memory functions during simultaneous interpreting. Rather than measuring cognitive load, I aim to explore potential links between cognitive processing and performance quality of Auslan-toEnglish simultaneous interpreting. I analyse the intricate relations among signed language interpreters’ processing time (measured in ELAN), interpreting strategies, online cognitive load, and performance quality (especially accuracy) in Auslan-to-English simultaneous interpreting. I focus on examining how different interpreters cope with challenges in the signed monologue. It is important to note that cognitive research on simultaneous interpreting also needs to take into account linguistic, motivational, psychological, situational, social, cultural, and many other factors that may influence information processing during simultaneous interpreting (Gerver, 1976). Simultaneous interpreting is not only a cognitive task but also a linguistic and communicative activity involving people, context, and communicative goal.

2.4 Gile’s Effort Model of Simultaneous Interpreting and ‘Tightrope Hypothesis’ Gile’s (1983, 1995, 2009) Effort Model of simultaneous interpreting is closely related to both cognitive load and working memory in simultaneous interpreting and thus will be used as the theoretical underpinning of the present study. Drawing on his extensive experience as a French/English/Japanese conference interpreter, Gile initially developed the Effort Model of spoken language simultaneous interpreting to explain to his students why simultaneous interpreting is so challenging and to find ways to address the challenges.Through close scrutiny of participants’ simultaneous interpreting performance, many empirical studies have revealed that both interpreting students and experienced interpreters make many errors (e.g., unjustifiable omissions, distortions, unidiomatic expressions in the target language, unnatural intonation) in their simultaneous interpretations despite their best efforts, not only when interpreting fast, dense, and technical source speeches, but also when interpreting slow, clear, and non-technical

26 Overview speeches (see Cokely, 1992; Gile, 1999, 2009; Liu et al., 2004; Napier, 2016; Wang, Napier, Goswell, & Carmichael, 2015). Gile’s Effort Model of ­simultaneous interpreting provides good explanations for recurrent errors in simultaneous interpretation. Gile’s (2009, pp. 160–171) Effort Model of simultaneous interpreting represents the process of spoken language simultaneous interpreting (SI) as encompassing four Efforts: SI = L + P + M + C, with each Effort being explained as follows: 1 the Listening and Analysis Effort (L, receiving and understanding messages in the source speech), 2 the Production Effort (P, reformulating the messages in the target language), 3 the Memory Effort (M, short-term memory operations up to a few seconds), and 4 the Coordination Effort (C, attention that is used to coordinate the other three Efforts). A good understanding of Gile’s Effort Model of simultaneous interpreting requires familiarity with some basic terms and assumptions. Efforts refer to nonautomatic operations in simultaneous interpreting that require an interpreter’s deliberate decisions and attention (Gile, 2009, p. 160). Processing capacity, a central concept in this model, is also referred to as attention, attentional resources, and mental energy (Gile, 2009, p. 159). The basic assumptions underlying Gile’s Effort Model of simultaneous interpreting include: (i) the Efforts are non-automatic operations that are slow and require processing capacity; (ii) human attention has a limited capacity; (iii) people can allocate at least part of their attention to specific tasks; and (iv) these Efforts compete for available attentional resources (Gile, 2009, pp. 159–160; 2018).Total processing capacity requirements (TR) comprise four components: processing capacity requirements for Listening and Analysis (LR), for short-term Memory (MR), for Production (PR), and for Coordination (CR), as shown by the formula TR = LR + MR + PR + CR. Moreover, Gile (2009, pp. 169–171) maintains that the following conditions must be met at any time during simultaneous interpreting for the interpreting process to proceed smoothly: 1 The total processing capacity requirements (TR) should not exceed an interpreter’s total available processing capacity (TA). 2 For each individual Effort (L, P, M, and C), the processing capacity requirement should not exceed the available processing capacity. Cognitive overload occurs when either condition is not met. That is, when the total processing capacity requirements exceed an interpreter’s total available processing capacity (i.e., supply cannot meet demand), cognitive overload occurs, causing deterioration in simultaneous interpreting performance, immediately or

Overview 27 subsequently. In addition, when the processing capacity requirement for an individual Effort (e.g., Listening and Analysis) exceeds the interpreter’s available processing capacity for that particular Effort, cognitive saturation also arises, leading to performance degradation. To put it differently, even if the interpreter has sufficient total available processing capacity, if s/he does not allocate the available attentional resources appropriately to each of the four individual Efforts, s/he may experience cognitive overload. Gile (1999, p. 157) also postulated a tightrope hypothesis, which states that ‘most of the time, interpreters work near saturation level.’ If this assumption is true, when the interpreter has a momentary attention lapse during simultaneous interpreting, he or she is likely to lose cognitive balance and make mistakes in target language renditions. In light of the tightrope hypothesis, many errors in simultaneous interpreting performance arise not because experienced interpreters do not have sufficient knowledge of their working languages, knowledge of the subject matter, or interpreting skills, but because attentional resources required for the interpreters to successfully perform a specific sub-task (e.g., comprehension of a source language sentence, temporary storage of a technical term, retrieval of information from long-term memory, or production of a target language equivalent) were unavailable at a particular moment when needed (Gile, 2009, p. 160; 2018). To put it differently, numerous errors in simultaneous interpretations are often due to interpreters’ mismanagement of their limited attention at particular junctures. For instance, when an interpreter directs excessive attention to using perfect syntax and vocabulary in the target language to convey understood ideas, s/he may not have sufficient processing capacity left to understand the incoming source language information. Gile’s Effort Model of simultaneous interpreting and his tightrope hypothesis indicate that interpreters’ attention allocation ability is critical to the quality of their simultaneous interpretation. Apart from the ability to allocate attention appropriately, interpreters’ fast processing of information can unload short-term memory and avoid cognitive saturation. Problems in simultaneous interpreting usually occur in the following situations: (i) when processing capacity requirements increase due to problem triggers in the source speech, (ii) when mismanagement of attentional resources happens (e.g., an interpreter allocates a considerable amount of attention to a particular Effort and there is insufficient attention left for other Efforts; his/her processing time is excessively long or short), and (iii) when an interpreter has a momentary attention deficit when receiving vulnerable segments (e.g., short words, numbers, different words that have the same pronunciation) in the source speech. Problem triggers in simultaneous interpreting (Gile, 2009) include: names, numbers, enumerations (lists of items), fast source speech, dense information in the source speech, prepared speeches read out by speakers, speakers’ strong regional or foreign accents, speakers’ incorrect grammar and lexical usage, poor logic of the source speech, inadequate sound quality, technical jargons, idiomatic expressions, and syntactical differences between the source language and the target language, among others.

28 Overview Problem triggers can cause processing capacity requirements to increase. For instance, when an interpreter needs to re-order informational sequence so as to deal with syntactical asymmetry between the source language and the target language, s/he has to memorise information for a longer time and allocate more attention to the Memory Effort; thus, s/he may not have sufficient processing capacity left to analyse the incoming information (Gile, 2009, p. 172). Another example is, when a speaker talks with a heavy accent, the interpreter may have to direct a remarkable amount of processing capacity to source speech comprehension and slow down target speech production, thus causing long processing time and short-term memory overload (Gile, 2009, p. 173). Due to its simplicity and strong explanatory power, Gile’s Effort Model of simultaneous interpreting has evolved beyond its initial pedagogical purpose and has been adopted by many researchers as a theoretical framework to explain cognitive processes in spoken and signed language simultaneous interpreting. Interestingly, empirical studies have generated some evidence that supports Gile’s Effort Model of simultaneous interpreting and his tightrope hypothesis. For example, many empirical studies have found that numbers in the source speech often cause interpretation errors and affect interpreters’ processing of the neighbouring segments in the source speech (e.g., Mackintosh, 1983; Mazza, 2000; Wang, 2016). Empirical research (e.g., Cattaneo, 2004) has also shown that idiomatic expressions in the source speech adversely influence interpreters’ processing of the neighbouring segments. In addition, in a seminal study on working memory and Mandarin/English simultaneous interpreting, Liu et al. (2004) inserted continuation sentences immediately after critical sentences in English speeches in order to test the attention allocation ability of beginning interpreting students, advanced interpreting students, and professional interpreters. Continuation sentences refer to short and simple English sentences whose initial two or three words were crucial for interpreters to correctly understand the meaning of these sentences. Critical sentences refer to English sentences that the researchers carefully manipulated in terms of difficulty level; and the accuracy of their Mandarin renditions indicated participants’ abilities to cope with varying difficulty levels and their abilities to select essential idea units over secondary idea units to interpret. Liu et al. found that the professional interpreters significantly outperformed both the advanced students and the beginning students on interpreting both the continuation sentences and the critical sentences, but the advanced students were similar to the beginning students in terms of performance on interpreting these sentences. Liu et al.’s results indicate that two skills – allocating limited attention to concurrent sub-tasks and differentiating essential idea units from secondary idea units – are crucial in simultaneous interpreting, and that professional interpreters are more proficient than students at applying these skills. However, Seeber (2011) maintains that the fluctuation of local cognitive load during simultaneous interpreting may not support Gile’s tightrope hypothesis. More empirical research is needed to provide further evidence to corroborate Gile’s Effort Model of simultaneous interpreting and to test his tightrope hypothesis.

Overview 29

2.5 Adapting Gile’s Effort Model This section first presents Gile’s adaptation of his Effort Model of spoken ­language simultaneous interpreting to signed language simultaneous interpreting, then describes my revision of Gile’s Effort Model to suit signed-to-spoken language simultaneous interpreting. Based on Pointurier-Pournin’s (2016) research on signed language interpreting, Gile (2018) proposed the following Effort Model for spoken-to-signed language simultaneous interpreting:

Sim  L  M  P  SMS  OID  C Sim: Simultaneous interpreting from a spoken language into a signed language L: Listening and Analysis of source speech P: Production of target speech, including self-monitoring M: Short-term Memory SMS: Self-Management in Space OID: Online Interaction with the Deaf C: Coordination

When working from a spoken language into a signed language, an interpreter typically works in simultaneous mode, stands on stage near the speaker, faces Deaf clients who sit in the audience, and maintains regular eye contact with the Deaf clients to receive external feedback on their own signed output. With regard to Self-Management in Space, the interpreter needs to make deliberate decisions on spatial positioning, distance to the speaker, and angles to optimise both comprehension of the source speech and production of the signed messages to the Deaf clients. Online Interaction with the Deaf signers refers to the interpreter’s attending to the Deaf signers’ feedback on their own signed output. In contrast, as to simultaneous interpreting from a signed language into a spoken language, Gile (2009, p. 160) suggested changing the Listening and Analysis Effort in the original model to Viewing and Analysis, which means watching a Deaf signer’s signed information and analysing its meaning. Based on my doctoral research on working memory and signed language interpreting (Wang, 2013a, 2013b, 2016) and the aforementioned literature, I  adapted Gile’s Effort Model of spoken language simultaneous interpreting to signed-to-spoken language simultaneous interpreting as follows.

SLI  Signed Language Comprehension  Target Speech Production  Working Memory  Coordination

This adapted Effort Model shows that the process of simultaneous interpreting from a signed language into a spoken language (SLI) consists of four core Efforts: (i) watching the Deaf signer’s signing and analysing the meaning (Signed Language Comprehension), (ii) expressing understood messages in the target language (Target Speech Production), (iii) temporarily storing and processing information that is relevant to a specific simultaneous interpreting task (Working Memory), and (iv) coordinating the other three Efforts (Coordination).

30 Overview This revised Effort Model serves as the theoretical framework for the present study on Auslan-to-English simultaneous interpreting. Based on my previous work (Wang, 2013a, 2013b, 2016), I deem the short-term Memory Effort (M) in Gile’s Effort Model equal to working memory. According to Gile (2009, p. 166), ‘Psychologists traditionally make a distinction between long-term memory (what we refer to as ‘memory’ in everyday life) and short-term memory, which is the ability to keep information and process it over a short period.’ Intriguingly, this definition of short-term memory aligns with Baddeley’s (2000, p. 418) definition of working memory, namely, ‘a limited capacity system allowing the temporary storage and manipulation of information necessary for such complex tasks as comprehension, learning, and reasoning.’ However, Gile (2018) maintains that the short-term Memory Effort in his Effort Model is not based on psychological construct of working memory, though closely related to it. Gile used the term short-term memory in his Effort Model because it is immediately recognisable by both students and professional interpreters as memory operations. Gile (2009, p. 167) also acknowledges that working memory ‘is obviously part of the Memory Effort and perhaps conceptually very close to it.’ Furthermore, Gile (2018) considered working memory as part of the Reception Effort (Listening/Viewing and Analysis Effort) and Production Effort, so he believed that it would make no sense to propose a distinct Working Memory Effort. Given the solid evidence showing that working memory plays an important role in complex cognitive activities such as speech comprehension, speech production, and simultaneous interpreting (see Mellinger & Hanson, 2019; Wang, 2016), I maintain that in my aforementioned adapted Effort Model, the Working Memory Effort interacts with the Signed Language Comprehension Effort,Target Speech Production Effort, and Coordination Effort during signed-to-spoken language simultaneous interpreting. In addition, interpreters retrieve relevant linguistic knowledge (vocabulary, technical terms, grammar rules, etc.) and extralinguistic knowledge (knowledge about the subject matter, the presenter, etc.) from longterm memory during simultaneous interpreting to complete source speech comprehension and target speech production. In other words, long-term memory also interacts with the Signed Language Comprehension Effort, Target Speech Production Effort, and Working Memory Effort in the revised Effort Model.

2.6 Processing Time in Simultaneous Interpreting Processing time refers to the time interval between receiving source language input and producing the corresponding target language output (Timarová, Dragsted, & Hansen, 2011). Previous research has investigated interpreters’ or translators’ processing time in English-to-ASL simultaneous interpreting (Cokely, 1992), spoken language simultaneous interpreting (e.g., Timarová et al., 2011), note-taking in consecutive interpreting through the use of pen recording (see Chen, 2017), sight translation, and written translation through the use of technologies such as eye-tracking and key-logging (e.g., Timarová, 2015;

Overview 31 Timarová et al., 2011). Chapters 5, 6, and 7 of this book, however, focus on professional interpreters’ processing time in simultaneous interpreting from a signed language (Auslan) into a spoken language (English), an under-researched topic in interpreting studies. There are two broad types of processing time in simultaneous interpreting: one is based on corresponding sentences (e.g., onset processing time, tail-to-tail span), the other is based on equivalent lexical items (e.g., processing time for numbers, processing time regarding nouns). Onset processing time (previously known as onset time lag) refers to the time interval between the beginning of a source language sentence and the beginning of its target language interpretation (see Cokely, 1992, p. 64; Wang, 2020, p. 108). Tail-to-tail span refers to the temporal delay between the end of a source language sentence and the end of its target language rendition (see Lee, 2003, pp. 42–43; Wang, 2020, p. 110). In this book, processing time for lexical items refers to the time interval between the beginning of a source language lexical item and the beginning of its target language equivalent (see also Defrancq, 2015, p. 30; Timarová et al., 2011), though the term is defined differently in some other studies (see Chen, 2017; Ono, Tohyama, & Matsubara, 2008). Processing time is required for simultaneous interpreting, in which the interpreter needs to wait for a meaningful chunk of information before expressing the message in the target language (Timarová et al., 2014). Simultaneous interpreting is a highly complex cognitive activity (Cokely, 1992; Collard & Defrancq, 2019; Gile, 2009; Liu et al., 2004; Timarová et al., 2014; Wang, 2016) in that the interpreter is engaged in concurrent sub-tasks such as comprehending the source language input, converting source language information to target language information, memorising information temporarily, delivering target language renditions, and monitoring the target language output. Research on processing time in simultaneous interpreting sheds interesting light on the cognitive activities that take place during the time interval between receiving a source speech unit and producing a corresponding target speech unit (Timarová, 2015; Wang, 2020), on how the temporal delay impacts on simultaneous interpreting performance (Cokely, 1986), and on how a myriad of internal and external factors influence the length of processing time (Collard & Defrancq, 2019). Processing time is a sensitive, dynamic, interesting, elusive, and complex variable in simultaneous interpreting research. Processing time provides insight into the temporal characteristics regarding the cognitive processes of simultaneous interpreting (Pöchhacker, 2016), the interpreter’s segmentation of the source text into chunks (Barik, 1975/2002; Pöchhacker, 2016; Podhajská, 2008), cognitive load (Timarová et al., 2011; Treisman, 1965), cognitive processing in simultaneous interpreting (Collard & Defrancq, 2019;Timarová, 2015;Timarová et al., 2011), processing speed (Timarová et al., 2011), and the minimum time that the interpreter needs for processing source language information in order to produce a meaningful rendition in the target language (Lee, 2006). Although some researchers (e.g., Podhajská, 2008;Timarová et al., 2014) maintain that interpreters’ ability to work with processing time partly determines the quality of their simultaneous interpreting performance, there has been little empirical evidence

32 Overview showing how this actually operationalises. Hence, more research is needed to examine the characteristics of processing time in simultaneous interpreting. With more knowledge about interpreters’ processing time in simultaneous interpreting, interpreters, researchers, and interpreter educators can achieve a better understanding of the underlying complex cognitive processes in simultaneous interpreting, often known as invisible cognitive activities in the interpreters’ ‘black box.’ There is a long tradition of measuring and analysing processing time in spoken language simultaneous interpreting; by contrast, there has been little empirical research (e.g., Cokely, 1986, 1992; McKee & Napier, 2002) on processing time in signed language simultaneous interpreting. Some experienced interpreter educators who are also conference interpreters (e.g., Gile, 2015; Setton & Dawrant, 2016) have also discussed the elusive and complex nature of processing time and its interrelation with numerous factors such as language pair, syntactic differences between the source language and the target language, and interpreting strategies. What follows is an overview of previous research into processing time in simultaneous interpreting. Previous research on processing time in spoken and signed language simultaneous interpreting has primarily concentrated on the temporal range of processing time, variability in processing time, external and internal factors that influence the length of processing time, and the relationship between processing time and the quality of simultaneous interpreting performance. 2.6.1 Temporal Range of Processing Time Many studies have measured interpreters’ processing time in simultaneous interpreting in order to establish this metric’s average length and temporal range. There is a broad consensus in prior studies that interpreters’ average processing time is approximately between 1 and 5 seconds (see Anderson, 1994; Barik, 1973; Christoffels & De Groot, 2004; Cokely, 1986; Collard & Defrancq, 2019; Defrancq, 2015; Díaz-Galaz, Padilla, & Bajo, 2015; Gerver, 1969/2002; Kim, 2005; Lee, 2002, 2006, 2011; Oléron & Nanpon, 1965/2002; Paneth, 1957/2002; Podhajská, 2008; Timarová et al., 2011; Treisman, 1965). For example, in regard to professional interpreters, the average processing time was 2.8 seconds for English-to-ASL simultaneous interpreting (Cokely, 1992, p. 69), 2.98 seconds for French-to-English (B-to-A or L2-to-L1) simultaneous interpreting (Anderson, 1994, p. 114), 2.68 seconds for French-to-Dutch simultaneous interpreting (Defrancq, 2015, p. 31), 3.6 seconds for English-to-Czech (L2-to-L1) simultaneous interpreting (Timarová et al., 2011, p. 138), 2.446 seconds and 4.532 seconds for English-to-Japanese (L2-to-L1) and Japanese-to-English (A-to-B or L1-to-L2) simultaneous interpreting respectively (Ono et al., 2008, p. 3386), and 3 seconds for English-to-Korean (L2-to-L1) simultaneous interpreting (Lee, 2002, p. 599). Interestingly, the finding of average processing time being in the range of 1–5 seconds appears to be true for simultaneous interpreting involving various

Overview 33 language pairs (syntactically dissimilar or similar languages), different interpreting modalities (spoken and signed language simultaneous interpreting), different interpreters with varying levels of simultaneous interpreting experience, different types of source texts, and diverse conditions under which simultaneous interpreting tasks were performed (e.g., with or without advance preparation) (Timarová et al., 2011). In terms of temporal range, individual interpreters’ processing time values reported in the available literature seem to range from −6.27 seconds to 19.17 seconds (see Collard & Defrancq, 2019; Defrancq, 2015, p. 31; Lamberger-Felber, 2001; McKee & Napier, 2002, p. 42; Podhajská, 2008, pp. 103–104), with the negative value indicating that anticipation is adopted by an interpreter as an interpreting strategy. The lower limit of processing time (in seconds) indicates the minimum time required by an interpreter to obtain sufficient source language information in order to start producing a target language interpretation, whereas the upper limit of processing time (in seconds) is indicative of the limit of human working memory, namely the maximum time that the interpreter can lag behind the source language speaker/signer with or without producing interpretation errors. Processing time represented in the number of words in simultaneous interpreting may range from a few lexical items to several sentences, depending on factors such as language combination, syntactical similarities or dissimilarities between the source language and the target language, cognitive processing, cognitive load, interpreters’ working memory capacity, fatigue, and interpreting strategies (Goldman-Eisler, 1972/2002; Kohn & Kalina, 1996; Wang, 2020), among other factors. 2.6.2 Variability in Processing Time Processing time not only varies from one interpreter to another but also fluctuates throughout a simultaneous interpreting task. There has been ample empirical evidence that processing time varies considerably between different interpreters (Defrancq, 2015; Lamberger-Felber, 2001; Timarová et al., 2011; Timarová et al., 2014). Such inter-subject variations in processing time may be due to individual differences in such areas as level of comprehension of the source text, familiarity with the subject matter and technical terms, the predictability of the source speech, memory capacity for different types of information, work habits, proficiency in the source and target languages, desire to achieve high accuracy, and individual cognitive constraints (Oléron & Nanpon, 1965/2002; Setton & Dawrant, 2016; Timarová, 2015). Different interpreters may employ different lengths of processing time, different interpreting processes, different interpreting strategies, and different norms depending on their own strengths (e.g., working memory capacity), weaknesses, and preferences (Collard & Defrancq, 2019). Moreover, each interpreter’s processing time changes considerably at different measurement points of the same source speech (Timarová et al., 2011; Wang, 2020). In other words, an interpreter does not maintain a steady processing time

34 Overview throughout a simultaneous interpreting task. It may well be that internal f­ actors – such as an interpreter’s familiarity with source language information (closely linked to the interpreter’s pre-task preparation, interpreting experience, and other factors), use of interpreting strategies at different points in time, and target language availability – contribute to the fluctuation of the interpreter’s processing time throughout a simultaneous interpreting task. In summary, processing time in simultaneous interpreting is individualistic (similar to spoken language interpreters’ note-taking skills for consecutive interpreting), dynamic (changing at different points during a simultaneous interpreting task), and complex (linking to a multitude of speaker-related, interpreter-related, and context-related factors). These findings suggest that processing time at a specific juncture in simultaneous interpreting is probably determined by a combination of external and internal factors (Podhajská, 2008;Timarová et al., 2011). 2.6.3 External and Internal Factors That Influence Processing Time A broad range of external factors (e.g., task type, source speech delivery rate, the length of source language sentences, text type, complexity level of the source speech, language pair, linguistic properties, redundancy in the source text, use of the speaker’s manuscript) have been found to influence interpreters’ processing time in simultaneous interpreting. A consistent finding across previous studies is that processing time in simultaneous interpreting is significantly longer than processing time in shadowing (Anderson, 1994; Christoffels & De Groot, 2004; Gerver, 1969/2002;Timarová et al., 2011;Treisman, 1965), indicating that simultaneous interpreting involves more complex cognitive processing, deeper information processing, more processing efforts, and larger processing units than shadowing. As source speech presentation rate increased, interpreters lagged further and further behind the speaker (Barik, 1973; Gerver, 1969/2002), suggesting that interpreters need more time to comprehend, process, and assess the importance of source language information when the speaker speeds up. As to professional Korean/English conference interpreters, long source language sentences resulted in long ear-voice span in English-to-Korean (B-to-A, or L2-to-L1) simultaneous interpreting while short source language sentences incurred short ear-voice span (Lee, 2002). Additionally, regarding English-toDutch (L2-to-L1) simultaneous interpreting, interpreting students’ processing time for interpreting a prepared (written beforehand) speech was significantly shorter than that for a spontaneous speech (Podhajská, 2008). Lamberger-Felber (2001) found that for 9 of 12 German/English conference interpreters with a minimum of ten years of professional experience, their processing time was longer during English-to-German (L2-to-L1) simultaneous interpreting with text than without text. Lamberger-Felber’s finding indicates that interpreters’ use of the source text in the booth reduces their working memory load and therefore allows for longer processing time during simultaneous interpreting. In addition, difficult segments in a source language speech significantly prolonged interpreting students’ processing time in English-to-Spanish (L2-to-L1) simultaneous

Overview 35 interpreting, but these difficult segments did not significantly affect professional interpreters’ processing time in the same simultaneous interpreting task (DíazGalaz et al., 2015). Furthermore, processing time is shorter when the source language and the target language are similar in syntactic structure (e.g., French and English) than when they are markedly different in terms of syntactic structure (e.g., German and English) (Goldman-Eisler, 1972/2002; Kim, 2005). Some empirical studies have found that the length of processing time is partly determined by the linguistic properties of lexical items on which processing time was measured. In a corpus-based study on processing time and simultaneous interpreting, Ono et al. (2008) observed that, in Japanese-to-English (A-to-B or L1-to-L2) simultaneous interpreting, professional interpreters’ processing time measured on nouns was longer than their processing time measured on verbs, whereas the result pattern was opposite in English-to-Japanese (L2-to-L1) simultaneous interpreting. Given that Ono et al. did not conduct statistical tests such as t-tests to compare participants’ processing time in different conditions, it remains unclear whether these differences reached statistical significance. Moreover, in a study exploring methodological issues in measuring processing time in spoken language simultaneous interpreting, Timarová et al. (2011) found that one professional interpreter’s processing time measured on numbers was significantly shorter than that measured on verbs or sentence beginnings, with processing time measured on verbs being comparable to that measured on sentence beginnings, indicating that source language items that are different in terms of parts of speech may require different types of cognitive processing and different lengths of processing time in simultaneous interpreting. Similarly, in their aforementioned study on processing time in English-to-Spanish (L2-to-L1) simultaneous interpreting, Díaz-Galaz et al. (2015) found that seven experienced interpreters’ processing time for source language segments containing nonredundant information such as numbers and acronyms was significantly shorter than their processing time for source language segments containing specialised terminology or complex syntax. Furthermore, in simultaneous interpreting from Japanese into English (L1-to-L2) and vice versa, professional interpreters’ processing time measured on numerals was shorter than that measured on nouns (Ono et al., 2008), indicating that numerals were interpreted more quickly than nouns. These results support Podhajská’s (2008) as well as Setton and Dawrant’s (2016) argument and recommendation that interpreters should use fairly short processing time (i.e., follow the speaker quite closely) when processing source language units that require verbatim memory (e.g., numbers, names, dates) so as to avoid cognitive overload and interpret those precise messages accurately. Seeber (2011) also maintains that factors such as language combination, directionality, and features of source speeches may influence interpreters’ processing time in simultaneous interpreting. There is also some empirical evidence that internal factors (e.g., interpreters’ pre-task preparation, interpreting experience, interpreting strategies) influence processing time in simultaneous interpreting. Advance preparation for simultaneous interpreting tasks significantly reduced both professional interpreters and

36 Overview interpreting students’ processing time in simultaneous interpreting and contributed to a higher level of accuracy in simultaneous interpretation (Díaz-Galaz et al., 2015). Díaz-Galaz et al. (2015) found that professional interpreters were similar to interpreting students in terms of processing time in simultaneous interpreting. However, Timarová et al. (2014) found a significant, negative, and moderate correlation between professional interpreters’ interpreting experience (in days) and their median processing time in simultaneous interpreting, indicating that more experienced interpreters tend to use shorter processing time than less experienced interpreters, perhaps because the experienced interpreters can process source language information faster than the less experienced interpreters in order to ensure a high level of accuracy of interpretation. Timarová et al. (2014) also found that professional interpreters who were better able to shift attention between different tasks tended to use shorter processing time in simultaneous interpreting, indicating that the cognitive ability to switch attention efficiently is closely linked to fast processing in simultaneous interpreting. Recently, Collard and Defrancq (2019) conducted a corpus-based study on 16 predictors of ear-voice span in professional interpreters’ real-life simultaneous interpreting between French and English and between Dutch and English. The researchers found that the following eight factors did have a significant impact on the interpreters’ processing time in simultaneous interpreting: (i) interpreter’s language, (ii) interpreter’s delivery rate, (iii) interpreter’s number of lengthenings, (iv) interpreter’s filled pauses, (v) interpreter’s false starts, (vi) original speaker’s language, (vii) original speaker’s delivery, and (viii) the original speaker/interpreter speaking time ratio. For example, the interpreters’ ear-voice span extended when they produced lengthenings, filled pauses, and false starts, indicating that the interpreters’ own delivery features (e.g., conciseness, fluency) influenced their processing time.The interpreters’ ear-voice span also increased when the speaker spoke faster, corroborating previous findings (Barik, 1973; Gerver, 1969/2002). The interpreters’ processing time also became longer when the original speaker/ interpreter speaking time ratio increased and caused greater cognitive load. However, Collard and Defrancq (2019) found that the following eight factors had no significant impact on the interpreters’ ear-voice span in simultaneous interpreting: interpreter’s and original speaker’s length of segments between pauses, interpreter’s and original speaker’s speech/pause ratio, interpreter’s and original speaker’s average duration of silent pauses, interpreter’s gender, and interpreter’s delivery type. That is, neither silent pauses nor sentence lengths in the source speeches had any significant influence on the interpreters’ processing time during simultaneous interpreting. Moreover, they found that male interpreters were similar to female interpreters in terms of ear-voice span. Collard and Defrancq’s study focused on measuring processing time in real-life simultaneous interpreting and conducting many complex statistical analyses. Further research is needed to conduct qualitative analysis of processing time in simultaneous interpreting in order to unveil how processing time becomes shorter or longer in response to interpreters’ use of interpreting strategies and delivery features.

Overview 37 2.6.4 Relationship Between Processing Time and Simultaneous Interpreting Performance Interpreters use specific processing time in combination with interpreting strategies in order to perform well in simultaneous interpreting involving particular language pairs (Kim, 2005; Lee, 2002). For example, in English-to-Korean (L2-to-L1) simultaneous interpreting, some professional interpreters seemed to shorten their processing time in order to process the incoming source language information quickly in an attempt to avoid cognitive overload, while other professional interpreters with long processing time spoke fast in order to catch up with the speaker (Lee, 2002). Interpreters’ management or adjustment of processing time is considered as a crucial and complex skill and strategy for simultaneous interpreting (Best, Napier, Carmichael, & Pouliot, 2016; Gile, 2009; McKee & Napier, 2002; Podhajská, 2008; Setton & Dawrant, 2016; Wang et al., 2015). Leneham (2011) found that Auslan/English interpreters used their processing time strategically to ensure that the integrity of the source text was conveyed, and that their processing time would shorten or lengthen according to the density of the source text and their search for equivalence. Apart from the aforementioned findings from empirical studies, some scholars (Podhajská, 2008; Timarová, 2015) maintain that an interpreter’s processing time is determined partly by other internal factors such as the individual approach of the interpreter, his/her segmentation of the source text, self-correction, personality, knowledge of the context, skills, physical and psychological aspects, and physiological limitations of the interpreter’s cognitive systems. These hypotheses merit further exploration. In summary, interpreters’ processing time in simultaneous interpreting is likely to be jointly determined by a complex web of external factors (e.g., processing difficulty, syntactical similarity or asymmetry between the source language and the target language, source speech presentation rate, source text type) and internal factors (e.g., interpreters’ tactics or strategies, familiarity with the topic and terminology, working memory capacity) that interact with each other (Gile, 2015; Setton & Dawrant, 2016; Timarová et al., 2011). There is a broad consensus in the existing literature that an interpreter needs to use appropriate processing time at any point during simultaneous interpreting in order to produce an effective interpretation and avoid cognitive saturation (Barik, 1973; Cokely, 1986; Lee, 2002; Oléron & Nanpon, 1965/2002; Setton & Dawrant, 2016). Extremely short or long processing time is likely to be problematic and affect simultaneous interpreting performance and should be avoided (Cokely, 1986; Setton & Dawrant, 2016; Timarová et al., 2011; Wang, 2020). However, the difficult questions are, what is the ideal processing time in a specific situation at a particular moment, how to strike the delicate balance between interpreting as soon as possible to avoid cognitive overload and waiting sufficiently long to understand enough source language information so as to produce accurate and meaningful interpretations, and how to adjust one’s processing time to produce effective renditions? Kade and Cartellieri (1971)

38 Overview proposed that the theoretically ‘ideal’ moment for an interpreter to start speaking during simultaneous interpreting is when s/he has resolved the syntactical and semantic ambiguities of a source speech unit and has decided on an unambiguous target language equivalent for it. Jones (2002, p. 116), an experienced spoken language conference interpreter, pointed out that an interpreter should only start speaking when s/he knows exactly what to say in the target language. According to Nilsson (2016, p. 33), many signed language interpreter educators recommend to their students that in signed-to-spoken language interpreting, they should wait until they ‘know what the signer is saying.’ These ‘golden rules’ on processing time in simultaneous interpreting for interpreting students and professional interpreters are useful but need to be quantified. Some empirical studies have cast light on the length of the optimum processing time in simultaneous interpreting. For example, Goldman-Eisler (1972/2002) observed that interpreters’ processing unit typically consisted of a minimum of a complete predicative expression (noun phrase + verb phrase), with the verb phrase being a crucial piece of information for the interpreters to start interpreting. Given that the location of the verb phrase varies across different languages, it is reasonable to assume that processing time may vary from one language pair to another (i.e., the language-pair specificity effect on processing time). Oléron and Nanpon (1965/2002) observed that processing time corresponding to satisfactory simultaneous interpreting performance across multiple language combinations (German-to-French, French-to-Spanish, English-to-French, and French-to-English, all involving European languages) was typically 2–3 seconds. Further research is needed to examine whether this processing time range also yields accurate interpretations in simultaneous interpreting involving Asian languages and signed languages (e.g., Mandarinto-English or Auslan-to-English simultaneous interpreting). In addition, Defrancq (2015) used a corpus-based approach to investigate the potential effects of short processing time on the quality of real-life French-toDutch simultaneous interpretations. He found that professional interpreters’ short (2 seconds) and very short (1 second) processing time in French-to-Dutch simultaneous interpretations was significantly more frequent in contexts where cognates occurred than elsewhere in the target speeches. Cognates refer to target language lexical items which are phonetically similar to their corresponding source language lexical items. Defrancq therefore concluded that interpreters’ use of short processing time during simultaneous interpreting is likely to increase their use of cognates.This finding indicates that (overly) short processing time in simultaneous interpreting probably causes both source language intrusion and predominantly literal interpretations. In Lee’s (2002) aforementioned study on processing time in English-toKorean (L2-to-L1) simultaneous interpreting, he found that the quality of simultaneous interpretations with processing time above 4 seconds was significantly poorer than that with processing time less than 2 seconds. Moreover, Lee found that the quality of simultaneous interpretations following those sentences with processing time above 4 seconds was lower than the quality of simultaneous

Overview 39 interpretations following those sentences with processing time less than 2 s­ econds. In other words, excessively long processing time – in this case, more than 4 ­seconds – severely affected not only the interpreters’ processing of the current sentence but also their processing of the subsequent sentence. These findings indicate that interpreters’ processing time of longer than 4 seconds during English-to-Korean simultaneous interpreting reduces the accuracy of interpretation, but processing time of shorter than 2 seconds was more effective. Interestingly, while processing time of less than 2 seconds is deemed short in French-to-Dutch simultaneous interpreting (Defrancq, 2015), it appears to be optimum in English-to-Korean simultaneous interpreting (Lee, 2002), indicating that the length of optimal processing time may vary across language pairs. Although Lee (2002) clarified that the accuracy of simultaneous interpretation was evaluated in terms of word correspondence, it remains unclear as to who the assessors were, what assessment instruments were employed, and how assessment outcomes were represented in his study. Similarly, Lamberger-Felber (2001) found that experienced interpreters’ processing time which was longer than the average value resulted in long omissions (i.e., omissions of more than 15 words in the source text). Taken together, interpreters’ overly long processing time is likely to cause interpretation errors such as unjustifiable omissions. Moreover, Barik (1973) found a positive correlation between interpreters’ processing time and their total number of omissions. Furthermore, Timarová et al. (2014) found significant, negative correlations between professional interpreters’ ear-voice span and their success in interpreting complex source language sentences, numbers, and double negation in English-to-Czech or Dutch simultaneous interpreting tasks. These results indicate that shorter processing time is closely related to both higher accuracy of simultaneous interpretation involving some language combinations and higher accuracy of rendering frozen items such as numbers. A plausible explanation is that professional interpreters who use short processing time and still achieve high accuracy of simultaneous interpretation can process information extremely fast (due to their interpreting experience, familiarity with the topic and terminology, mastery of working languages, etc.) and can minimise the negative effects of short processing time (see also Defrancq, 2015; Timarová et al., 2011). To date, only a few empirical studies have examined processing time in signed language interpreting. In Cokely’s (1986) pioneering study on four certified ASL/English interpreters’ real-life English-to-ASL simultaneous interpreting performance, he found that two interpreters who had average onset processing time of 2 seconds made considerably more miscues – specifically, unjustifiable omissions, additions, substitutions, intrusions (source language interference), and anomalies (target language information that did not make sense) – than the other two interpreters who had average onset processing time of 4 seconds. These findings indicate an inverse relationship between the length of processing time and the number of interpretation errors. In other words, as for English-toASL simultaneous interpreting, onset processing time of 4 seconds is more favourable than onset processing time of 2 seconds. Cokely explained that those

40 Overview interpreters with longer processing time (4 seconds) had more time to both analyse source language information and seek appropriate vocabulary and syntactic structure in the target language to convey the understood meaning accurately. Similarly, in McKee and Napier’s (2002) study of interpreters’ strategies in real-life simultaneous interpreting from English into International Sign, they found that the interpreters typically worked with long processing time (usually 10–16 seconds and at times even longer) in this particular language direction in order to analyse the meaning of source speech units, convey the meaning by using conceptually equivalent target language renditions, and make use of contextual information. The interpreters in McKee and Napier’s study not only stretched the upper limits of their storage and processing capacity (working memory capacity) but also appeared to be highly conscious of how much processing time they needed in order to produce meaningful and effective interpretations. These interpreters lagged further behind the source language speaker when simultaneously interpreting from English into International Sign than when working into their national sign language (processing time being 3–6 seconds), indicating that they strategically adjusted their processing time to cope with syntactical differences between a particular source language and a particular target language. Taken together, these findings highlight the importance of using appropriate processing time at any specific point during simultaneous interpreting to strike a delicate cognitive balance. These results also provide insight into the intricate relationship among processing time, language combination, interpreting strategies, conscious awareness, and accuracy of simultaneous interpretation. Given the complex and dynamic nature of processing time in simultaneous interpreting, further research in this area is warranted. In summary, although interpreters’ average processing time in simultaneous interpreting across previous studies is typically between 1 and 5 seconds, the length of optimum processing time appears to vary from one language pair to another. Prior research has consistently shown that processing time not only varies from one interpreter to another, but also varies considerably from one measurement point to another within a simultaneous interpreting task. There is broad consensus in the interpreting studies literature that processing time is jointly determined by a plethora of external factors and internal factors that interact with each other in complicated ways. While some studies (e.g., Lee, 2002) found that processing time of longer than 4 seconds led to more interpretation errors than processing time of shorter than 2 seconds, other studies (e.g., Cokely, 1986) found that processing time at or less than 2 seconds resulted in source language intrusion and other interpretation miscues. The discrepancies in research findings may be due to language pair as well as methodological differences (e.g., in terms of points of measurement, measurement method, unit of measurement, use of mean versus median for statistical analysis, and tests for statistical analysis; see Timarová et al., 2011), among others. It is generally accepted that exceptionally long processing time may lead to unjustifiable omissions and reduced accuracy, whereas excessively short

Overview 41 processing time is likely to affect both source speech comprehension and target speech production and thus result in source language intrusion and predominantly literal interpretations. Therefore, interpreters need to strike a delicate cognitive balance between the amount of source language input needed for understanding the speaker/signer sufficiently and the limits of working memory capacity (see Collard & Defrancq, 2019; Wang, 2020). The overview presented here unveils that processing time in signed language interpreting, especially in simultaneous interpreting from a signed language into a spoken language, is a significant gap that warrants further research.

2.7 Numbers as a Problem Trigger in Simultaneous Interpreting This section discusses previous studies on numbers in simultaneous interpreting. Many practitioners perceive numbers as a challenge in simultaneous interpreting, irrespective of language combination (Cheung, 2009; Frittella, 2019; Gile, 2009; Korpal & Stachowiak-Szymczak, 2020; Mazza, 2001; Mead, 2015; Pinochi, 2009). Other problem triggers in simultaneous interpreting include but are not limited to: proper names, speakers’ fast delivery, speakers’ accents, and dense information in the source text (Gile, 2009). Since numbers represent factual information, it is important for interpreters to understand and memorise the figures and their associated concepts, then render them accurately in the target language. However, very few studies have examined reasons why numbers are so challenging for interpreters or have proposed effective strategies to cope with numbers in simultaneous interpreting. Findings of the existing studies are summarised in this section to inform the present research. Numbers can be particularly difficult for interpreters to understand, remember, process, convert, and reformulate during simultaneous interpreting because of their intrinsic features such as ephemeral nature, low redundancy, low predictability from context, exact content, univocal meaning with only one direct equivalent in the target language, lack of semantic content, as well as high informative content (Braun & Clarici, 1996; Frittella, 2019; Gile, 2009; Korpal & Stachowiak-Szymczak, 2020; Mazza, 2001; Mead, 2015; Pinochi, 2009; Taylor, 2002). Numbers are pronounced quickly; interpreters’ memory traces of numbers last for only a few seconds, then decay rapidly.The exact value of a number can rarely be inferred from its neighbouring elements in a sentence (i.e., low predictability) (Korpal & StachowiakSzymczak, 2020); therefore, if an interpreter misses or forgets the number, it is almost impossible for him/her to recover or reconstruct it from the context. As numbers represent precise numerical information and lack semantic content, interpreters need to devote sufficient attention to listening to or watching the figures carefully and memorising them temporarily. Given that each component of numbers signifies only one specific meaning, this characteristic deprives interpreters from applying interpreting strategies such as anticipation, paraphrasing, and reformulation to cope with them (Desmet,Vandierendonck, & Defrancq, 2018; Pinochi, 2009).

42 Overview While the semantic content of a sentence may be reconstructed or inferred from the context (i.e., intelligent hearing), numbers require interpreters to allocate adequate attention to the numerical items alone in order to understand their meaning and store them in working memory (i.e., literal hearing) (Braun & Clarici, 1996; Pinochi, 2009). As a result, when processing a source language segment embedded with numbers, interpreters need to constantly switch between intelligent hearing and literal hearing. Due to the aforementioned language-independent reasons, numbers, especially large ones, are highly likely to increase cognitive load (e.g., add pressure on working memory) or leave insufficient attentional resources for other concurrent sub-tasks during simultaneous interpreting, thus often causing errors in the rendition of the numbers themselves and/or disrupting the interpretation of their surrounding elements (Gile, 2009; Mazza, 2001; Pinochi, 2009). Prior experimental studies on numbers in simultaneous interpreting typically focused on error analysis (e.g., error types, error rate) of student interpreters’ renditions of numbers in speeches. For example, Pinochi (2009) compared eight students’ German-to-Italian (B-to-A) simultaneous interpretation with another eight students’ English-to-Italian (B-to-A) simultaneous interpretation in terms of number errors. While English and Italian share a linear numerical system, German requires the application of the inversion rule, according to which ‘25’ is pronounced ‘five and twenty’ rather than ‘twenty-five.’ Pinochi incorporated five categories of numbers in the source speeches: (i) numbers with four or more digits read at once (e.g., 920,000), (ii) numbers with four or more digits read in two blocks (e.g., 928,346), (iii) numbers with fewer than four digits, (iv) decimals, and (v) dates. Pinochi used the following number error typology (which was based on Braun and Clarici’s (1996) number error taxonomy): (i) omissions (the number was left out completely or replaced by a general expression such as ‘many’ or ‘few’), (ii) approximations (the numeral was rounded up or down, e.g., ‘8.1%’ was rendered as ‘about 8%’), (iii) lexical errors (the order of magnitude of the stimulus was maintained but one or more digits were distorted, e.g., ‘2004’ was skewed as ‘2005’), (iv) syntactical errors (although the number contained the correct figures in their correct order, it had a wrong order of magnitude, e.g., ‘300’ was changed to ‘300,000’), (v) errors of phonemic perception (a phonemically wrong perception of the stimulus, e.g., ‘17’ was distorted as ‘70’), (vi) errors of transposition of the digits (wrong order of the digits, e.g., ‘47’ was replaced by ‘74’), and (vii) other mistakes. Pinochi found that the interpreting students’ inaccuracy rate for numbers in both the German-to-Italian and English-to-Italian simultaneous interpreting tasks was considerably high, at 40.6% and 41.2%, respectively. These results confirm the broad consensus among interpreter practitioners that numbers are difficulties in simultaneous interpreting, irrespective of language pair. Interestingly, large numbers (with four or more digits) as well as decimals were found to be more susceptible to errors than small numbers (with fewer than four digits) and dates, possibly due to the word length effect on working memory. Another

Overview 43 intriguing finding is that omissions stood out as the most frequent type of number mistake, followed by approximations and lexical errors. In relation to language pair, transposition errors regarding numbers occurred significantly more frequently in German-to-Italian simultaneous interpreting than in English-toItalian simultaneous interpreting, indicating that differences between the source language and the target language in terms of numerical system affect the accuracy of number interpretation. Braun and Clarici (1996) investigated the effects of directionality and notetaking on 12 interpreting students’ accuracy of rendering numbers during simultaneous interpreting between German and Italian. Most participants had Italian as their native language (A) and German as their non-native language (B). Results showed that the students had an error rate of 69.49% in interpreting numbers within speeches. Similar to Pinochi (2009), Braun and Clarici found that common types of number errors in both directions of simultaneous interpreting included omissions, approximations, lexical mistakes, and syntactical errors. In relation to directionality, the students interpreted more numbers inaccurately when working from German into Italian (B-to-A interpreting for most participants) than vice versa, likely due to difficulties in comprehending a nonnative language (German). In addition, note-taking of numbers improved the accuracy of rendering them during German-to-Italian simultaneous interpreting, which was B-to-A interpreting for most participants. This result indicates that writing down numbers alleviates both the Listening and Analysis Effort and the short-term Memory Effort in simultaneous interpreting from a non-native language into a native language (see Gile’s Effort Model in Section 2.4). In addition, Mazza (2001) examined whether numbers affect the interpreting of neighbouring items and whether note-taking facilitates the interpreting of numbers. Fifteen interpreting students completed English-to-Italian (B-to-A) simultaneous interpreting tasks under two experimental conditions: with or without note-taking. Results showed that the students’ interpretation of speech segments without numbers was more accurate than their interpretation of speech segments with numbers, indicating that numbers were challenging to the students. A key finding is that the error rate of numbers in simultaneous interpreting with note-taking was 45.1%, whereas that for without note-taking was 49.9%.This finding indicates that note-taking improved the accuracy of number interpretation, thus corroborating Braun and Clarici’s aforementioned finding. Large numbers (with four or more digits), decimals, and ranges were found to be more difficult to interpret than small numbers (with fewer than four digits) and dates, thus lending support to Pinochi’s (2009) results. Results also revealed that omissions were the most common type of number errors, followed by approximations, thus supporting the findings obtained by Pinochi (2009) as well as Braun and Clarici (1996).What is particularly interesting is that the interpreting students in Mazza’s (2001) study frequently omitted the referents of numbers or speech elements surrounding the numbers, probably due to their working memory deficit at those junctures. Another intriguing finding is that participants’ performance on repeated numbers was not necessarily better than that of

44 Overview numbers that occurred only once, indicating that interpreters probably release numbers from their working memory as soon as they have transferred them into the target language. Furthermore, Desmet et al. (2018) conducted an experimental study to assess the impact of a mock-up technological support on interpreting students’ processing of numbers during B-to-A simultaneous interpreting. Seven students interpreted from French into Dutch in simultaneous mode and another three students worked from German into Dutch in simultaneous mode, all working from their B language into their A language. Each source speech contained numbers. The researchers distinguished the numbers into four categories: (i) simple whole numbers (e.g., 87 or 60,000), (ii) complex whole numbers (e.g., 387 or 65,400), (iii) decimals, and (iv) years. Desmet et al. prepared PowerPoint slides that showed numbers that had just been articulated by the speaker, simulating an ideal technological support that would quickly transform spoken numbers into written numbers and display the numbers on the screen in real time. Desmet et al.’s results revealed that the interpreting students’ error rate for complex numbers, decimals, simple numbers, and dates during simultaneous interpreting was 68%, 49%, 30%, and 28%, respectively. These results support previous findings (Mazza, 2001; Pinochi, 2009) that complex/large numbers and decimals are more challenging than simple/small numbers and dates. Unsurprisingly, visual access to the numbers shown on PowerPoint slides significantly enhanced the accuracy rate of interpreting each of the four types of numbers, especially complex numbers and decimals. In addition, omissions were the most frequent errors, followed by approximations and lexical mistakes; these findings regarding number error distribution corroborate the findings obtained by Braun and Clarici (1996), Mazza (2001), and Pinochi (2009). Taken together, these findings suggest that allowing interpreters to see numbers projected on PowerPoint slides during simultaneous interpreting not only reduces cognitive load but also improves the accuracy of number interpretation. While all the aforementioned experiments focused on student interpreters’ success rate of rendering numbers during simultaneous interpreting, few studies have explored professional interpreters’ accuracy of interpreting numbers. Korpal and Stachowiak-Szymczak (2020) investigated the influence of the speaker’s delivery rate on 30 professional interpreters and 24 interpreting students’ renditions of numbers during English-to-Polish (B-to-A) simultaneous interpreting with visual access to PowerPoint slides showing numbers, definitions, and important facts. As expected, for all participants, the accuracy score of numbers in the slow delivery rate condition was significantly higher than that in the fast delivery rate condition. In addition, the professional interpreters significantly outperformed the students in interpreting numbers not only in the slow speed condition (98% vs. 92%) but also in the fast speed condition (92% vs. 81%), indicating that the professional interpreters were better at coping with numbers and more automatic at multimodal (acoustic and visual modalities) processing. It is noteworthy that the accuracy rate of number interpretations was at ceiling level for the professional interpreters under both delivery rate

Overview 45 conditions and for the students under the slow delivery rate condition, probably due to participants’ access to the projected presentation slides and the fact that there were only ten numbers in each source speech.These results indicate that it is necessary to train interpreting students on how to deal with numbers during simultaneous interpreting, and that projected PowerPoint slides reduce the difficulty of rendering numbers. In contrast to the aforementioned experimental studies on the simultaneous interpretation of numbers, Collard and Defrancq (2021) conducted a corpusbased study to compare female professional interpreters with male ones in terms of renditions of numbers in 180 source speeches at the European Parliament. The researchers divided numbers into six categories: regular numbers, dates, and decimals, each category with three or more components or with two or fewer components. Collard and Defrancq classified interpreters’ number renditions into five categories: complete renditions (accurate interpretations), omissions, approximations, related substitutions, and unrelated substitutions. They also measured the interpreters’ ear-voice span regarding numbers, that is, the temporal interval between a number uttered by the speaker and a number interpretation articulated by the interpreter. Collard and Defrancq found no significant differences between the female interpreters and the male interpreters in terms of the accuracy rate of rendering numbers. Unlike the majority of previous studies of interpreting students’ number renditions that showed rather high error rate (e.g., Desmet et al., 2018; Mazza, 2001; Pinochi, 2009), the professional interpreters’ accuracy rate of interpreting numbers in the real-life conference setting in Collard and Defrancq’s study was as high as 79%. Similar to prior studies, omissions (14%) were found to be the most common type of number interpretation errors, followed by unrelated and related substitutions as well as approximations. A key finding from Collard and Defrancq’s (2019) data-driven study is that the professional interpreters’ processing time for numbers was the only significant predictor of the accuracy of number interpretations. In other words, the shorter the processing time for numbers, the more accurate the interpretations of the numbers. Collard and Defrancq (2019) found that the professional interpreters’ mean processing time (average of processing time measured on all corresponding lexical items, not just numbers, in the source text and the target text) was not a significant predictor of the accuracy of their number interpretations. However, Timarová et al. (2014) found a significant, negative correlation between professional interpreters’ median onset processing time (temporal interval between the beginning of a source language sentence and the beginning of its target language sentence) and accuracy of interpretation of numbers (i.e., interpreters who used shorter onset processing time rendered more numbers accurately). It is unclear if the discrepancies among the existing findings are due to the methodological differences in measuring processing time (Collard and Defrancq’s use of corresponding lexical items as measurement points versus Timarová et al.’s use of sentence beginnings as measurement points), the use of different types of

46 Overview descriptive statistics (Collard and Defrancq’s use of mean processing time versus Timarová et al.’s use of median onset processing time), or other factors. Another important finding of Collard and Defrancq’s (2019) study is that the professional interpreters’ ear-voice span for inaccurate renditions of numbers was significantly longer than their ear-voice span for correct interpretations of the numbers.This finding indicates that interpreters can render more numbers accurately during simultaneous interpreting by shortening their processing time to deal specifically with the numerical items. In summary, the aforementioned findings from prior empirical studies on numbers in simultaneous interpreting supported Gile’s (2009) claim that numbers are problem triggers in simultaneous interpreting of various language combinations. Most of the experimental studies on numbers in simultaneous interpreting have focused on interpreting students and found a rather high error rate (approximately 40%) in their renditions of numbers, highlighting a need to train students on how to cope with numbers during simultaneous interpreting. Only a few studies have analysed professional interpreters’ simultaneous interpretations of speeches containing numbers and found a considerably lower error rate regarding number interpretations in professional interpreters (approximately 20%) than in interpreting students. Omissions and approximations have been found to be the most common types of interpretation errors regarding numbers. Asking the booth mate (team interpreter) to write down numbers is a typical strategy used by interpreters to deal with numbers ­during real-life simultaneous interpreting. In addition, several studies (e.g., Braun & Clarici, 1996; Desmet et al., 2018; Lamberger-Felber, 2001; Mazza, 2001) have revealed that note-taking by the team interpreter or the working interpreter, seeing numbers projected on PowerPoint slides, and using the source text in the booth considerably improved the accuracy rate of rendering numbers during simultaneous interpreting. A corpus-based study by Collard and Defrancq (2021) has shown that professional interpreters strategically shortened their processing time in order to render numbers accurately during simultaneous interpreting. Given that omissions, approximations, and unjustifiable substitutions are common categories of inaccurate renditions of numbers, using a simple taxonomy of number interpretations (e.g., accurate interpretations, unjustifiable omissions, approximations, and unjustifiable distortions) may be more useful than the detailed taxonomy applied in many of the previous studies. To my knowledge, no empirical study has explored signed language interpreters’ renditions of numbers during simultaneous interpreting. In an attempt to bridge this gap, Chapter 5 of this book reports on findings regarding professional Auslan/English interpreters’ accuracy rate of rendering numbers in Auslan-toEnglish simultaneous interpreting and presents findings regarding the link between the accuracy of interpreting numbers and the length of processing time. In addition, Chapter 6 presents results from detailed analyses of interpreters’ English renditions of Auslan segments containing numbers, particularly discussing cognitive overload and the interpreters’ coping strategies regarding the numbers.

Overview 47

2.8 Strategies in Simultaneous Interpreting Compared with error analysis of interpreting performance, research on interpreting strategies strikes a more positive note because it regards interpreting as a dynamic process that involves interpreters’ conscious decision-making and goaloriented (intentional) problem-solving. This section presents an overview of previous studies on interpreting strategies. Strategy has been defined differently by various scholars. Bartłomiejczyk (2006, p. 152) defines interpreting strategies as ‘methods that are potentially conducive to solving particular problems encountered by interpreters or generally facilitating the interpreter’s task and preventing potential problems.’ Bartłomiejczyk (2006) maintains that interpreters’ successful use of automated strategies reduces cognitive load during simultaneous interpreting. Gile (2009, pp. 262–263) distinguishes strategies from tactics, with the former referring to planned actions taken by an interpreter to achieve a longer-term goal (e.g., preparing for an interpreting assignment) and the latter referring to immediate actions taken by the interpreter to cope with online problems encountered during interpreting (e.g., omitting secondary and redundant information in order to focus on conveying primary information, i.e., strategic omission). In this book, I use ‘strategies’ and ‘tactics’ interchangeably. In view of the aforementioned definitions, interpreters adopt strategies during (consecutive and simultaneous) interpreting to prevent problems from occurring and/or to solve problems. 2.8.1 Comprehension Tactics, Preventive Tactics, and Reformulation Tactics Gile (2009, pp. 201–211) proposed three main categories of coping tactics for simultaneous interpreting: comprehension tactics, preventive tactics, and reformulation tactics. Interpreters employ comprehension tactics when they encounter or sense difficulties in understanding the source speech. There are four types of comprehension tactics: (i) delaying response for up to a few seconds (increasing processing time or waiting), (ii) restructuring the segment with the help of the context, (iii) using the team interpreter’s help, and (iv) consulting documents in the booth. Interpreters use preventive tactics to avoid cognitive overload and prevent problems from occurring in the first place (Gile, 2009). Gile proposed four types of preventive tactics: (i) taking notes of numbers and names, (ii) lengthening or shortening processing time, (iii) segmentation (cutting a long sentence into short segments rather than waiting for the long sentence to finish, also called chunking) and unloading of short-term memory, and (iv) changing the order of elements in an enumeration (i.e., a list of items). In addition, interpreters adopt reformulation tactics to deal with target language production or short-term memory problems (Gile, 2009). Gile posited 14 sub-categories of reformulation tactics: (i) delaying response for up to a few seconds, (ii) using the team interpreter’s help, (iii) consulting documents in the

48 Overview booth, (iv) replacing a segment with a superordinate term or a more general speech segment (generalising), (v) explaining or paraphrasing, (vi) reproducing the sound heard in the source language speech, (vii) ‘instant naturalisation’ (adapting a source language word to the phonological and/or morphological rules in the target language when the interpreter does not know its target language equivalent), (viii) transcoding (translating a source language term or speech segment into the target language word for word, especially for rendering numbers, names, and lists of items), (ix) form-based interpreting (primarily using source language words and syntax to guide oneself in reformulating the target speech), (x) informing listeners of a problem, (xi) referring delegates to another information source such as PowerPoint slides, (xii) omitting the content of a source speech segment, (xiii) ‘parallel’ reformulation (in case of poor working conditions, interpreters may invent a target speech segment that matches the remaining part of the speaker’s source speech), and (xiv) switching off the microphone. The last five reformulation tactics are deemed emergency tactics. All these comprehension, preventive, and reformulation tactics during simultaneous interpreting proposed by Gile (2009) for spoken language interpreters, except reproducing the sound heard in the source language speech and ‘instant naturalisation,’ may be useful to signed language interpreters. Since some of the coping tactics may require time and result in working memory overload, it is necessary for interpreters to choose tactics carefully. Gile (2009, pp. 211–214) recommended that interpreters consider the following rules (in the order of decreasing importance) when selecting tactics: (i) maximising information recovery (interpreters try to reformulate all source messages in the target language), (ii) minimising interference in information recovery (interpreters attempt to recover as much information as possible on each segment without affecting the recovery of other segments), (iii) maximising the communication impact of the speech, (iv) the law of least effort (interpreters may favour tactics that require the least time and processing effort, but sometimes this rule results in interpreters selecting tactics that require less attention even when attentional resources are available), and (v) self-protection (e.g., some interpreters may choose not to inform consumers of a problem in their target language renditions for the sake of protecting themselves). Furthermore, anticipation (also referred to as prediction) plays an important part in simultaneous interpreting. Gile (2009, pp. 173–175) suggested that, the more interpreters use linguistic and extralinguistic anticipation, the less processing capacity they need to allocate to source language comprehension, and therefore the more processing capacity remains for short-term memory, target language production, and coordination. Note that linguistic anticipation refers to the interpreter’s use of grammatical rules, collocations, standard phrases, and transitional probabilities to facilitate source speech comprehension (Gile, 2009, pp. 173–174). Extralinguistic anticipation refers to the interpreter’s use of her/his good knowledge about the conference situation, the context, the topic, the PowerPoint slides shown on a screen, and the speaker to anticipate the information that the speaker is going to say (Gile, 2009, p. 174). Therefore, both interpreters’ high level of bilingual

Overview 49 proficiency and their thorough preparation for a simultaneous interpreting assignment (e.g., reading documents and preparatory materials beforehand to familiarise themselves with the relevant names, technical terms, expressions, facts, and ideas that may be referred to by the conference speakers) increase their ability to predict accurately what the speakers will say (Gile, 2009, p. 175). When an interpreter predicts something wrongly, s/he then needs to make a self-correction. 2.8.2 Literal Interpretation and Free Interpretation This section defines literal interpretation and free interpretation, which frequently occur in interpreting studies literature. In signed language interpreting studies, concepts such as transcoding, transcodage, transliteration, form-based translation, and literal interpretation are synonymous. Napier (2006, p. 45) defines literal interpretation as ‘retaining the form of the source message to some degree, providing a more formal equivalence in which either the original lexical items or syntactic structure are recognisable in the target language message.’ Napier (2006, p. 45) defines free interpretation as ‘achievement of linguistic, cultural, pragmatic, and dynamic equivalence, where the message is “freed” from the form of the source language and the focus is on meaning.’ In other words, free interpretation or dynamic translation not only focuses on reformulating source messages in linguistically and culturally acceptable ways in the target language, but also recognises that people who participate in the interpreter-mediated communication may bring their own life experiences to the interpreting situation (Napier & Leeson, 2015). To put it differently, literal interpretation primarily involves surface level or form-based processing, whereas free interpretation primarily requires deeper level or meaning-based processing. In an empirical study on accredited Auslan/English interpreters’ use of linguistic coping strategies in interpreting a dense university lecture into Auslan, Napier (2016) found that the interpreters were dominant or extremely dominant in employing a free or literal interpretation approach. She also found that interpreters who used a dominantly free interpretation approach switched to a more literal interpretation approach at key junctures in order to introduce subject-specific English terms to Deaf university students. Based on the results, Napier (2016a) proposed a translation style continuum, and posited that signed language interpreters should code-switch between a free interpretation style and a literal interpretation style in formal discourse environments which present features of language contact between a spoken language and a signed language, such as when interpreting university lectures from a spoken language into a signed language. 2.8.3 Strategic Omissions This section presents a review of previous studies on strategic omissions, highlighting that while some omissions are errors, others are strategic behaviours. Omissions and additions are often examined by comparing the source text with the target text in terms of message equivalence in order to assess the accuracy of interpreting performance. Error analysis (i.e., miscue taxonomy) has been widely used in spoken

50 Overview and signed language interpreting studies (e.g., Barik, 1975/2002; Cokely, 1992) to investigate the types and numbers of errors (i.e., inaccuracies, inadequacies, mistakes, miscues, problems) in interpreters’ target language renditions. Recent studies, however, have started to examine interpreters’ deliberate production of strategic omissions from linguistic, cultural, contextual, and pragmatic perspectives. In her aforementioned study on signed language interpreting, Napier (2004, 2016) proposed an omission taxonomy that comprised the following five types of omissions: (i) conscious strategic omissions, (ii) conscious intentional omissions, (iii) conscious unintentional omissions, (iv) conscious receptive omissions, and (v) unconscious omissions. Conscious strategic omissions refer to instances of omissions in which the interpreter draws on his/her linguistic and cultural knowledge to consciously omit redundant, untranslatable, and culturally irrelevant information in the source text in order to produce effective interpretations. In applying the omission taxonomy to analyse ten Auslan/English interpreters’ Auslan interpretations of a university lecture, Napier found that unconscious omissions and conscious strategic omissions were the most common types of omissions in the English-to-Auslan simultaneous interpretation data. She also found that those interpreters who were more familiar with the subject matter of the source speech made fewer omissions (all five types together) than those who were less familiar with the topic. Another interesting finding is that most interpreters who held postgraduate or undergraduate qualifications made fewer than the average number of omissions. However, Napier did not find a clear relationship between the interpreters’ frequency of conscious strategic omissions and either their familiarity with the lecture topic, educational background, or (literal or free) translation style. Based on these findings, Napier concluded that conscious strategic omissions appeared to have been used by all interpreters in her study as a linguistic coping strategy. She also concluded that some interpreters’ unfamiliarity with the lecture topic, subjectspecific terminology, and the general academic discourse environment resulted in the occurrence of omissions in their English-to-Auslan interpretations. The notion of conscious strategic omissions somewhat resembles Wadensjö’s (1998) concepts of reduced renditions (less explicitly expressed information in the target language rendition than the corresponding source language utterance), summarised renditions (a target language rendition that corresponds to two or more preceding original utterances), and zero renditions (source language messages that are not translated) in interpreter-mediated dialogic interactions. Napier’s (2004) concept of strategic omission is also similar to Major and Napier’s (2012) notion of strategic reduction (e.g., interpreter’s omission of redundant information that is already clear from the context). 2.8.4 Strategic Additions A number of studies (Braun, 2017; Kalina, 1998; Major & Napier, 2012;Wadensjö, 1998;Wang & Fang, 2019) have examined strategic additions made by interpreters in dialogue, consecutive, and simultaneous interpreting. In this book,

Overview 51 strategic additions refer to addition, explicitation, expansion, explanation, rephrasing, elaboration, and/or reiteration made deliberately by an interpreter in order to enhance the effectiveness of an interpretation, clarify intended meaning, convey the speaker’s/signer’s pragmatic meaning, maximise audience design (i.e., make the renditions better suited to the special needs of the target audience), narrow perceived gaps between source culture and target culture, and achieve an equivalent communicative effect. In her aforementioned study of interpreter-mediated dialogic interactions, Wadensjö (1998) discussed expanded renditions (renditions that contain more explicitly expressed information than the corresponding source language utterances) and non-renditions (renditions that the interpreter adds on his/her own initiative), both of which may be regarded as strategic additions if they are conducive to the interpreted communication. In addition, Merlini and Favaron (2005, pp. 287–293, as cited in Braun, 2017, p. 167) classified additions into emphatic additions (interpreter-initiated reiterations or repetitions), explanatory additions (explaining a message to clarify an ambiguous point), phatic additions, and request for clarification, among others. In addition, Jacobsen (2002, p. 167, as cited in Braun, 2017, p. 167) proposed an addition taxonomy that included emphasising additions (increasing the illocutionary force of an utterance), downtoning additions (reducing the force of an utterance), repetitions (fillers), explicating additions (adding culture-bound information, making non-verbal information explicit), elaborating additions (elaborating on a message that has already been interpreted), and new-information additions (adding a piece of information that is not mentioned in the original utterance). Some empirical studies have analysed examples of strategic additions in great detail. In a study of interpreter-mediated doctor-patient consultations, Major and Napier (2012) found that signed language interpreters made various types of expanded renditions (i.e., strategic additions), including making implicit source language information explicit in the target language, adding cohesive elements to link the current target language rendition back to a previous statement, and expanding source language information to ensure the clarity of meaning. In addition, Baraldi (2012, as cited in Braun, 2017, p. 167) analysed interpreters’ reformulation in healthcare interpreting, namely rephrasing and expanding source language information in order to ensure that the recipient understands the intended meaning. Moreover, Braun (2017) adopted an ‘enriched’ Conversation Analysis framework to investigate both the functions and motivations of additions and expansions made by interpreters during videoconference interpreting of police interviews. Braun found that many of the additions and expansions represented the interpreters’ strategic moves which aimed to ensure that their target language renditions meet the special needs of the receiver. Furthermore, in a comparative study of Mandarin/English telephone interpreting and on-site interpreting, Wang and Fang (2019) identified examples of one professional interpreter’s strategic additions such as making implicit information in a source language utterance explicit in the target language, adding

52 Overview turn-taking cues, and explaining source language concepts that were unfamiliar to the recipient. It is worthy of attention that the line between unjustifiable, acceptable, and strategic additions and omissions may be obscure and difficult to draw (Mikkelson, 2000), and thus researchers need to cautiously consider the context, the speaker’s/signer’s intent, the communicative goal, and the receiver’s needs when judging whether particular additions and omissions are strategic. Little research has explored strategic omissions and strategic additions employed by professional interpreters when working from a signed language into a spoken language in simultaneous mode at formal settings.

2.9 Summary This chapter has presented an overview of previous research that is relevant to simultaneous interpreting from a signed language into a spoken language. Specifically, the chapter has summarised research findings regarding directionality, interpreting from a signed language into a spoken language, cognitive load, Gile’s (2009) Effort Model of spoken language simultaneous interpreting, adapting the model to signed-to-spoken language simultaneous interpreting, processing time, numbers as a problem trigger, and strategies in simultaneous interpreting.

Notes 1 In this book, I use the terms native language, L1, and A language to refer to a person’s mother tongue, the language that s/he acquired at birth and which is usually deemed the strongest language. I use the terms non-native language, non-native active language, secondary language, L2, and B language to mean a language that the person acquired or learned later in life but which is not as strong as the L1. I use the terms non-native passive language and C language to mean a language that the person understands perfectly but has difficulties in expressing ideas fluently and clearly. AIIC (International Association of Conference Interpreters) classifies conference interpreters’ working languages as A, B, and C languages. AIIC stipulates that conference interpreters can work from their B and C languages into their A language in both consecutive and simultaneous modes, that they can work from their A and C languages into their B language preferably in consecutive mode only, and that they should not work from their A or B language into their C language. Spoken and signed language interpreters who are balanced bilinguals are exempt from AIIC’s directionality principles because they transfer meaning between two native languages. For further information, please visit https://aiic.org/site/world/about/ profession/abc. 2 Napier (2006, p. 45) defines literal interpretation as ‘retaining the form of the source message to some degree, providing a more formal equivalence in which either the original lexical items or syntactic structure are recognisable in the target language message.’ Feyne (2015) considers that literal interpretations, similar to intrusion miscues (Cokely, 1992), convey the overall meaning of the original utterances but retain some features of the source language form. 3 The term utterance refers to a phrasal linguistic unit or a sentence.

Overview 53

References Anderson, L. (1994). Simultaneous interpretation: Contextual and translation aspects. In S. Lambert & B. Moser-Mercer (Eds.), Bridging the Gap: Empirical Research in Simultaneous Interpretation (pp. 101–120). Amsterdam/Philadelphia: John Benjamins. Baddeley, A. D. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4(11), 417–423. doi: 10.1016/S1364-6613(00)01538-2. Baddeley, A. D. (2003). Working memory and language: An overview. Journal of Communication Disorders, 36(3), 189–208. doi: 10.1016/S0021-9924(03)00019-4. Baddeley, A. D., & Hitch, G. J. (2000). Development of working memory: Should the Pascual-Leone and the Baddeley and Hitch models be merged? Journal of Experimental Child Psychology, 77(2), 128–137. doi: 10.1006/jecp.2000.2592. Baraldi, C. (2012). Interpreting as dialogic mediation: The relevance of expansions. In L. Gavioli (Ed.), Coordinating Participation in Dialogue Interpreting (pp. 297–326). Amsterdam: John Benjamins. Barik, H. C. (1973). Simultaneous interpretation:Temporal and quantitative data. Language and Speech, 16(3), 237–270. doi: 10.1177/002383097301600307. Barik, H. C. (1975/2002). Simultaneous interpretation: Qualitative and linguistic data. In F. Pöchhacker & M. Shlesinger (Eds.), The Interpreting Studies Reader (pp. 78–91). London/New York: Routledge. Bartłomiejczyk, M. (2006). Strategies of simultaneous interpreting and directionality. Interpreting, 8(2), 149–174. doi: 10.1075/intp.8.2.03bar. Best, B., Napier, J., Carmichael, A., & Pouliot, O. (2016). From a Koine to Gestalt: Critical points and interpreter strategies in interpretation from International Sign into spoken English. In R. Rosenstock & J. Napier (Eds.), International Sign: Linguistic, Usage and Status Issues (pp. 136–166). Washington, DC: Gallaudet University Press. Braun, S. (2017). What a micro-analytical investigation of additions and expansions in remote interpreting can tell us about interpreter’s participation in a shared virtual space. Journal of Pragmatics. Special Issue ‘Participation in Interpreter-Mediated Interaction,’ 107, 165–177. doi: 10.1016/j.pragma.2016.09.011. Braun, S., & Clarici, A. (1996). Inaccuracy for numerals in simultaneous interpretation: Neurolinguistic and neuropsychological perspectives. The Interpreters’ Newsletter, 7, 85–102. Burke, T. B., & Nicodemus, B. (2013). Coming out of the hard of hearing closet: Reflections on a shared journey in academia. Disability Studies Quarterly, 33(2). Cattaneo, E. (2004). Idiomatic expressions in conference interpreting. (Graduation thesis). Universita degli studi di Bologna, Sede di Forli. Charles, R. L., & Nixon, J. (2019). Measuring mental workload using physiological measures: A systematic review. Applied Ergonomics, 74, 221–232. doi: 10.1016/j.apergo.2018.08.028. Chen, S. (2017). Note-taking in consecutive interpreting: New data from pen recording. Translation & Interpreting, 9(1), 4–23. doi: 10.12807/ti.109201.2017.a02. Cheung, A. K. F. (2009). Numbers in simultaneous interpreting: An experimental study. Forum, 7(2), 61–88. doi: 10.1075/forum.7.2.03che. Christoffels, I. K. (2004). Cognitive Studies in Simultaneous Interpreting. (Unpublished doctoral dissertation). University of Amsterdam, Amsterdam, Netherlands. Christoffels, I. K., & De Groot, A. M. B. (2004). Components of simultaneous interpreting: Comparing interpreting with shadowing and paraphrasing. Bilingualism: Language and Cognition, 7(3), 227–240. doi: 10.1017/S1366728904001609.

54 Overview Christoffels, I. K., De Groot, A. M. B., & Kroll, J. F. (2006). Memory and language skills in simultaneous interpreters: The role of expertise and language proficiency. Journal of Memory and Language, 54(3), 324–345. doi: 10.1016/j.jml.2005.12.004. Cokely, D. (1986). The effects of lag time on interpreter errors. Sign Language Studies, 53(Winter), 341–375. doi: 10.1353/sls.1986.0025. Cokely, D. (1992). Interpretation: A Sociolinguistic Model. Burtonsville, MD: Linstok Press. Collard, C., & Defrancq, B. (2019). Predictors of ear-voice span, a corpus-based study with special reference to sex.Perspectives,27(3),431–454.doi:10.1080/0907676X.2018.1553199. Collard, C., & Defrancq, B. (2021). Corpus study on the rendition of numbers in simultaneous interpreting, with special reference to sex. Interpreting. De Meulder, M., Napier, J., & Stone, C. (2018). Designated or preferred? A deaf academic and two signed language interpreters working together for a PhD defense: A case study of best practice. International Journal of Interpreter Education, 10(2), 5–26. Defrancq, B. (2015). Corpus-based research into the presumed effects of short EVS. Interpreting, 17(1), 26–45. doi: 10.1075/intp.17.1.02def. Desmet, B., Vandierendonck, M., & Defrancq, B. (2018). Simultaneous interpretation of numbers and the impact of technological support. In C. Fantinuoli (Ed.), Interpreting and Technology (pp. 13–27). Berlin: Language Science Press. Díaz-Galaz, S., Padilla, P., & Bajo, M. T. (2015). The role of advance preparation in simultaneous interpreting: A comparison of professional interpreters and interpreting students. Interpreting, 17(1), 1–25. doi: 10.1075/intp.17.1.01dia. Dickinson, J. (2017). Signed Language Interpreting in the Workplace. Washington, DC: Gallaudet University Press. Diriker, E. (2015). Simultaneous interpreting. In F. Pöchhacker (Ed.), Routledge Encyclopedia of Interpreting Studies (pp. 382–385). New York: Routledge. Feyne, S. (2014). Interpreting identity:The impact of sign language interpreters on the construction of the situated identity of Deaf professionals (unpublished MA thesis). City University of New York. Feyne, S. (2015). Typology of interpreter-mediated discourse that affects perceptions of the identity of Deaf professionals. In B. Nicodemus & K. Cagle (Eds.), Signed Language Interpretation and Translation Research: Selected Papers from the First International Symposium (pp. 49–70). Washington, DC: Gallaudet University Press. Feyne, S. (2016). Impact of ideology and interpreter-mediated discourse on perceptions of identity of Deaf professionals. In A. Kalata-Zawłocka & B. van den Bogaerde (Eds.), To Say or Not to Say – Challenges of Interpreting from Sign Language to Spoken Language. Proceedings of the 23rd efsli Conference in Warsaw, Poland, 11th–13th September 2015 (pp. 49–57). European Forum of Sign Language Interpreters (efsli). Frittella, F. M. (2019).‘70.6 Billion World Citizens’: Investigating the difficulty of interpreting numbers. Translation & Interpreting, 11(1), 79–99. doi: 10.12807/ti.111201.2019.a05. Galy, E., Cariou, M., & Mélan, C. (2012).What is the relationship between mental workload factors and cognitive load types? International Journal of Psychophysiology, 83, 269– 275. doi: 10.1016/j.ijpsycho.2011.09.023. Gerver, D. (1969/2002). The effects of source language presentation rate on the performance of simultaneous conference interpreters. In F. Pöchhacker & M. Shlesinger (Eds.), The Interpreting Studies Reader (pp. 52–66). London/New York: Routledge. Gerver, D. (1976). Empirical studies of simultaneous interpretation: A review and a model. In R. W. Brislin (Ed.), Translation: Applications & Research (pp. 165–207). New York: Gardner Press.

Overview 55 Gevins, A., & Smith, M. E. (2003). Neurophysiological measures of cognitive workload during human–computer interaction. Theoretical Issues in Ergonomics Science, 4(1–2), 113–131. doi: 10.1080/14639220210159717. Gile, D. (1983). Des difficultés de langue en interprétation simultanée. Traduire, 117(Octobre), 2–8. Gile, D. (1995). Basic Concepts and Models for Interpreter and Translator Training (1st ed.). Amsterdam/Philadelphia: John Benjamins. Gile, D. (1999).Testing the Effort Models’ tightrope hypothesis in simultaneous interpreting-A contribution. Hermes, Journal of Linguistics, 12(23), 153–172. doi: 10.7146/hjlcb. v12i23.25553. Gile, D. (2009). Basic Concepts and Models for Interpreter and Translator Training (Rev. ed.). Amsterdam/Philadelphia: John Benjamins. Gile, D. (2015). The contributions of cognitive psychology and psycholinguistics to conference interpreting: A critical analysis. In A. Ferreira & J. W. Schwieter (Eds.), Psycholinguistic and Cognitive Inquiries into Translation and Interpreting (pp. 41–66). Philadelphia: John Benjamins. Gile, D. (2018). The Effort Models and Gravitational Model: Clarifications and Update. http:// cirinandgile.com/powerpoint/The-Effort-Models-and-Gravitational-ModelClarifications-and-update.pdf. Accessed on September 17, 2019. Goldman-Eisler, F. (1972/2002). Segmentation of input in simultaneous translation. In F. Pöchhacker & M. Shlesinger (Eds.), The Interpreting Studies Reader (pp. 69–76). London/ New York: Routledge. Harrelson, P. (2019). Deaf Employees’ Perspectives on Effective American Sign Language-English Interpreting in the Workplace: An Investigation Using the Critical Incident Technique. (Unpublished doctoral dissertation). Gallaudet University, Washington, DC, United States. Hart, S. G., & Staveland, L. E. (1988). Development of NASA-TLX (Task Load Index): Results of empirical and theoretical research. In P. A. Hancock & N. Meshkati (Eds.), Human Mental Workload (pp. 139–183). Amsterdam: Elsevier Science. Haug,T., Bontempo, K., Napier, J., Nicodemus, B., van den Bogaerde, B., &Vermeerbergen, M. (2017). Deaf leaders’ strategies for working with signed language interpreters: An examination across seven countries. Across Languages and Cultures, 18(1), 107–131. doi: 10.1556/084.2017.18.1.5. Hauser, P. C., Finch, K. L., & Hauser, A. B. (Eds.). (2008). Deaf Professionals and Designated Interpreters: A New Paradigm. Washington, DC: Gallaudet University Press. Jacobsen, B. (2002). Pragmatic meaning in court interpreting: An empirical study in consecutively interpreted question-answer dialogues (Unpublished doctoral dissertation). The Aarhus School of Business, Aarhus, Denmark. Jones, R. (2002). Conference Interpreting Explained (2nd ed.). Manchester, UK: St. Jerome. Kade, O., & Cartellieri, C. (1971). Some methodological aspects of simultaneous interpreting. Babel: International Journal of Translation, 17(2), 12–16. doi: 10.1075/babel.​ 17.2.04car. Kalina, S. (1998). Strategische Prozesse beim Dolmetschen: Theoretische Grundlagen, Empirische Fallstudien, Didaktische Konsequenzen. Tübingen: Gunter Narr. Kellett Bidoli, C. J. (2002). Spoken-language and signed-language interpretation: Are they really so different? In G. Garzone & M.Viezzi (Eds.), Interpreting in the 21st Century: Challenges and Opportunities. Selected Papers from the 1st Forlì Conference on Interpreting Studies, 9–11 November 2000 (pp. 171–179). Amsterdam/Philadelphia: John Benjamins.

56 Overview Kim, H.-R. (2005). Linguistic characteristics and interpretation strategy based on EVS analysis of Korean-Chinese, Korean-Japanese interpretation. Meta, 50(4). http:// id.erudit.org/iderudit/019846ar. Kohn, J., & Kalina, S. (1996). The strategic dimension of interpreting. Meta, 41(1), 118– 138. doi: 10.7202/003333ar. Korpal, P., & Stachowiak-Szymczak, K. (2020). Combined problem triggers in simultaneous interpreting: Exploring the effect of delivery rate on processing and rendering numbers. Perspectives, 28(1), 126–143. doi: 10.1080/0907676X.2019.1628285. Lamberger-Felber, H. (2001). Text-oriented research into interpreting: Examples from a case study. Hermes, Journal of Linguistics, 14 (26), 39–64. doi: 10.7146/hjlcb.v14i26.25638. Lee, T. (2002). Ear voice span in English into Korean simultaneous interpretation. Meta, 47(4), 596–606. doi: 10.7202/008039ar. Lee,T. (2003).Tail-to-tail span: A new variable in conference interpreting research. Forum, 1(1), 41–62. doi: 10.1075/forum.1.1.03lee. Lee, T. (2006). A comparison of simultaneous interpretation and delayed simultaneous interpretation from English into Korean. Meta, 51(2), 202–214. doi: 10.7202/013251ar. Lee, T. (2011). English into Korean simultaneous interpretation of Academy Awards ceremony through open captions on TV. Meta, 56(1), 145–161. doi: 10.7202/1003514ar. Leneham, M. (2011). Modelling the process of simultaneous interpreting. Paper presented to the National Interpreter Trainers’ Network (ITN) Symposium, Sydney, 29–30 August 2011. Liu, M., Schallert, D. L., & Carroll, P. J. (2004). Working memory and expertise in simultaneous interpreting. Interpreting, 6(1), 19–42. doi: 10.1075/intp.6.1.04liu. Mackintosh, J. (1983). Relay Interpretation: An Exploratory Study. (Unpublished MA thesis). Birkbeck College, University of London, London. Major, G., & Napier, J. (2012). Interpreting and knowledge mediation in the healthcare setting: What do we really mean by ‘Accuracy’? Linguistica Antverpiensia, 11, 207–225. Mazza, C. (2000). Numbers in Simultaneous Interpretation. (Unpublished graduation thesis). University of Bologna, SSLMIT Forli. Mazza, C. (2001). Numbers in simultaneous interpretation. The Interpreters’ Newsletter, 11, 87–104. McKee, R., & Napier, J. (2002). Interpreting into International Sign Pidgin: An analysis. Journal of Sign Language and Linguistics, 5(1), 27–54. Mead, P. (2015). Numbers. In F. Pöchhacker (Ed.), Routledge Encyclopedia of Interpreting Studies (pp. 286–288). New York: Routledge. Mellinger, C. D., & Hanson, T. A. (2019). Meta-analyses of simultaneous interpreting and working memory. Interpreting, 21(2), 165–195. doi: 10.1075/intp.00026.mel. Merlini, R., & Favaron, R. (2005). Examining the ‘Voice of interpreting’ in speech pathology. Interpreting, 7(2), 263–302. doi: 10.1075/bct.9.08mer. Mikkelson, H. (2000). Introduction to Court Interpreting. Manchester, UK: St. Jerome. Miller, G. A. (1956). The magical number seven plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63, 81–97. Napier, J. (2002). Sign Language Interpreting: Linguistic Coping Strategies. Coleford, Gloucestershire: Douglas McLean. Napier, J. (2004). Interpreting omissions: A new perspective. Interpreting, 6(2), 117–142. doi: 10.1075/intp.6.2.02nap. Napier, J. (2006). Comparing language contact phenomena between Auslan-English interpreters and deaf Australians: A preliminary study. In C. Lucas (Ed.), Multilingualism and Sign Languages: From the Great Plains to Australia (pp. 39–77). Washington, DC: Gallaudet University Press.

Overview 57 Napier, J. (2007). Cooperation in interpreter-mediated monologic talk. Discourse & Communication, 1(4), 407–432. doi: 10.1177/1750481307082206. Napier, J. (2011). Signed language interpreting. In K. Malmkjær & K. Windle (Eds.), The Oxford Handbook of Translation Studies (pp. 357–376). Oxford/New York: Oxford University Press. Napier, J. (2016). Linguistic Coping Strategies in Sign Language Interpreting.  Washington, DC: Gallaudet University. Napier, J., Carmichael, A., & Wiltshire, A. (2008). Look-pause-nod: A linguistic case study of a deaf professional and interpreters working together. In P. C. Hauser, K. L. Finch, & A. B. Hauser (Eds.), Deaf Professionals and Designated Interpreters: A New Paradigm (pp. 22–42). Washington, DC: Gallaudet University Press. Napier, J., & Leeson, L. (2015). Signed language interpreting. In F. Pöchhacker (Ed.), Routledge Encyclopedia of Interpreting Studies (pp. 376–381). New York: Routledge. Napier, J., McKee, R., & Goswell, D. (2010). Sign Language Interpreting:Theory and Practice in Australia and New Zealand (2nd ed.). Sydney, Australia: The Federation Press. Napier, J., Rohan, M., & Slatyer, H. (2005). Perceptions of bilingual competence and preferred language direction in Auslan/English interpreters. Journal of Applied Linguistics, 2(2), 185–218. doi: 10.1558/japl.2005.2.2.185. Nicodemus, B., & Emmorey, K. (2013). Direction asymmetries in spoken and signed language interpreting. Bilingualism: Language and Cognition, 16(3), 624–636. doi: 10.1017/ S1366728912000521. Nicodemus, B., & Emmorey, K. (2015). Directionality in ASL-English interpreting: Accuracy and articulation quality in L1 and L2. Interpreting, 17(2), 145–166. doi: 10.1075/intp.17.2.01nic. Nilsson, A.-L. (2016). Interpreting from signed language into spoken language: The skills and knowledge needed to succeed. In A. Kalata-Zawłocka & B. van den Bogaerde (Eds.), To Say or Not to Say – Challenges of Interpreting from Sign Language to Spoken Language. Proceedings of the 23rd efsli Conference in Warsaw, Poland, 11th–13th September 2015 (pp. 15–48). European Forum of Sign Language Interpreters (efsli). Oléron, P., & Nanpon, H. (1965/2002). Research into simultaneous translation. In F. Pöchhacker & M. Shlesinger (Eds.), The Interpreting Studies Reader (pp. 42–50). London/ New York: Routledge. Ono,T.,Tohyama, H., & Matsubara, S. (2008). Construction and analysis of word-level timealigned simultaneous interpretation corpus. In N. Calzolari, K. Choukri, B. Maegaard, J. Mariani, J. Odijk, S. Piperidis, & D. Tapias (Eds.), Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC’08) (pp. 3383–3387). Paris: ELRA. Paas, F. G. W. C., Tuovinen, J. E., Tabbers, H. K., & Van Gerven, P. W. M. (2003). Cognitive load measurement as a means to advance cognitive load theory. Educational Psychologist, 38(1), 63–71. doi: 10.1207/S15326985EP3801_8. Paas, F. G. W. C., & van Merriënboer, J. J. G. (1994). Variability of worked examples and transfer of geometrical problem solving skills: A cognitive load approach. Journal of Educational Psychology, 86(1), 122–133. doi: 10.1037/0022-0663.86.1.122. Paneth, E. (1957/2002). An investigation into conference interpreting. In F. Pöchhacker & M. Shlesinger (Eds.), The Interpreting Studies Reader (pp. 30–40). London/New York: Routledge. Pinochi, D. (2009). Simultaneous interpretation of numbers: Comparing German and English to Italian: An experimental study. The Interpreters’ Newsletter, 14, 33–57. Pöchhacker, F. (2015). Modes. In F. Pöchhacker (Ed.), Routledge Encyclopedia of Interpreting Studies (pp. 268–269). New York: Routledge.

58 Overview Pöchhacker, F. (2016). Introducing Interpreting Studies (2nd ed.). London/New York: Routledge. Podhajská, K. (2008). Time lag in simultaneous interpretation from English into Czech and its dependence on text type. Folia Translatologica, 10, 87–110. Pointurier-Pournin, S. (2016). Cognitive challenges in sign-to-voice interpreting. In A. KalataZawłocka & B. van den Bogaerde (Eds.), To Say or Not to Say – Challenges of Interpreting from Sign Language to Spoken Language. Proceedings of the 23rd efsli Conference in Warsaw, Poland, 11th–13th September 2015 (pp. 71–85). European Forum of Sign Language Interpreters (efsli). Reid, G. B., & Nygren, T. E. (1988). Subjective workload assessment technique: A scaling procedure for measuring mental workload. In P. A. Hancock & N. Meshkati (Eds.), Human Mental Workload (pp. 185–217). Amsterdam: Elsevier Science. Seeber, K. G. (2011). Cognitive load in simultaneous interpreting: Existing theories – New models. Interpreting, 13(2), 176–204. doi: 10.1075/intp.13.2.02see. Seeber, K. G. (2013). Cognitive load in simultaneous interpreting: Measures and methods. Target, 25(1), 18–32. doi: 10.1075/target.25.1.03see. Seeber, K. G. (2015). Cognitive load. In F. Pöchhacker (Ed.), Routledge Encyclopedia of Interpreting Studies (pp. 60–61). New York: Routledge. Seeber, K. G., & Kerzel, D. (2012). Cognitive load in simultaneous interpreting: Model meets data. International Journal of Bilingualism, 16(2), 228–242. doi: 10.1177/1367006911402982. Setton, R., & Dawrant, A. (2016). Conference Interpreting: A Complete Course. Amsterdam: John Benjamins. Taylor, M. M. (2002). Interpretation Skills: American Sign Language to English. Edmonton, Alberta: Interpreting Consolidated. Timarová, Š. (2015). Time lag. In F. Pöchhacker (Ed.), Routledge Encyclopedia of Interpreting Studies (pp. 418–420). New York: Routledge. Timarová, Š., Čeňková, I., Meylaerts, R., Hertog, E., Szmalec, A., & Duyck, W. (2014). Simultaneous interpreting and working memory executive control. Interpreting, 16(2), 139–168. doi: 10.1075/intp.16.2.01tim. Timarová, Š., Dragsted, B., & Hansen, I. G. (2011). Time lag in translation and interpreting: A methodological exploration. In C. Alvstad, A. Hild, & E. Tiselius (Eds.), Methods and Strategies of Process Research: Integrative Approaches in Translation Studies (pp. 121–146). Amsterdam: John Benjamins. Treisman, A. M. (1965). The effects of redundancy and familiarity on translating and repeating back a foreign and a native language. British Journal of Psychology, 56(4), 369– 379. doi: 10.1111/j.2044-8295.1965.tb00979.x. van Dijk, R., Boers, E., Christoffels, I., & Hermans, D. (2011). Directionality effects in simultaneous language interpreting: The case of sign language interpreters in the Netherlands. American Annals of the Deaf, 156(1), 47–55. doi: 10.1353/aad.2011.0013. Wadensjö, C. (1998). Interpreting as Interaction (1st ed.). London/New York: Longman. Wang, J. (2013a). Bilingual working memory capacity of professional Auslan/English interpreters. Interpreting, 15(2), 139–167. doi: 10.1075/intp.15.2.01wan. Wang, J. (2013b). Working memory and signed language interpreting. (Unpublished doctoral dissertation). Macquarie University, Sydney, Australia. Wang, J. (2016). The relationship between working memory capacity and simultaneous interpreting performance: A mixed methods study on professional Auslan/English interpreters. Interpreting, 18(1), 1–33. doi: 10.1075/intp.18.1.01wan.

Overview 59 Wang, J. (2020). Striking a cognitive balance: Processing time in Auslan-to-English simultaneous interpreting. In D. Hunt & E. Shaw (Eds.), The Second International Symposium on Signed Language Interpretation and Translation Research: Selected Papers (pp. 108–131). Washington, DC: Gallaudet University Press. Wang, J., & Fang, J. (2019). Accuracy in telephone interpreting and on-site interpreting: A comparative study. Interpreting, 21(1), 36–61. doi:10.1075/intp.00019.wan. Wang, J., & Napier, J. (2015). Directionality in signed language interpreting. Meta, 60(3), 518–541. doi: 10.7202/1036141ar. Wang, J., Napier, J., Goswell, D., & Carmichael, A. (2015). The design and application of rubrics to assess signed language interpreting performance. The Interpreter and Translator Trainer, 9(1), 83–103. doi: 10.1080/1750399X.2015.1009261. Wickens, C. D. (1984). Processing resources in attention. In R. Parasuraman & D. R. Davies (Eds.), Varieties of Attention (pp. 63–102). New York: Academic Press. Xie, B., & Salvendy, G. (2000). Review and reappraisal of modelling and predicting mental workload in single- and multi-task environments. Work and Stress, 14(1), 74–99. doi: 10.1080/026783700417249. Zhang, W. (2009). Working memory in different directions of simultaneous interpreting (in Chinese). Foreign Language Teaching and Research, 41(5), 371–378.

3 Methodology

The primary focus of this research was to explore the intricate relationship among performance quality, cognitive overload, and strategies in signed-to-spoken language simultaneous interpreting at formal settings. Specifically, this study examined the interpreting product, cognitive process, and strategies of Auslanto-English simultaneous interpreting at a mock conference. This chapter will describe the research design and methodology, providing information about the development of stimulus material, the recruitment of professional Auslan/ English interpreters, the creation of an experimental corpus of Auslan-toEnglish simultaneous interpretation data, three external raters’ analytic assessment of participants’ interpreting performance, my microanalysis of the corpus, as well as thematic analysis of post-task interview data. In my completed PhD research project, 31 professional Auslan/English interpreters were instructed to simultaneously interpret an English speech into Auslan, simultaneously interpret an Auslan presentation into English, complete an English listening span task (for measuring their English working memory capacity), and then complete an Auslan working memory span task, with each task followed by a short retrospective interview (Wang, 2013, 2016; Wang & Napier, 2013, 2015a, 2015b;Wang, Napier, Goswell, & Carmichael, 2015). In this book, I concentrate on detailed analyses of 20 of the 31 professional interpreters’ Auslan-to-English simultaneous interpretations and their retrospective interview data. These 20 participants were selected because they scored above 65 out of 100 on the Auslan-to-English simultaneous interpreting task, as per the three external raters’ individual analytic assessment using a rubric. The external raters were all hearing native Auslan signers and professional Auslan/English interpreters with extensive experience in conference interpreting and community interpreting. I subsequently conducted microanalysis of the experimental corpus of Auslanto-English simultaneous interpretation data in terms of accuracy, processing time, cognitive load, strategies, coherence, and formal register, among facets, by using tools such as ELAN and Excel. As a professional Mandarin/English interpreter, I relied heavily on the three external raters’ oral and written comments made during their analytic assessment as well as research assistance provided by another experienced professional Auslan/English interpreter (who did not participate in the study)

Methodology

61

to thoroughly understand both the Auslan monologue and the Auslan-to-English simultaneous interpretation corpus. Furthermore, interpreters’ qualitative interview data was analysed for recurrent themes regarding their self-perceived challenges in the Auslan-to-English simultaneous interpreting task and coping strategies.

3.1 The Auslan Source Text The first step in this research project was to create an Auslan monologue that would serve as the source language stimulus material for the signed language interpreters. The Auslan source text was developed for a mock national conference themed ‘Interpreting and Human Rights.’ The audience of the signed presentation was described as including both English-speaking attendees who were familiar with human rights issues but did not understand Auslan or Deaf culture, and Deaf professionals who were bilingual in both Auslan and written English. In creating the Auslan material, attention was given to selecting a Deaf professional who was a native user of the language, demonstrated clear articulation when signing, had professional expertise in relation to the conference theme, and frequently worked with signed language interpreters in formal settings such as conferences and professional workplaces. A Deaf native Auslan signer in Sydney, the manager of a Deaf community service organisation, provided the Auslan source text by giving a signed presentation entitled ‘Deaf People and Human Rights.’ To make the source text as realistic as possible, the Deaf presenter had been briefed about the conference theme and the intended audience of his presentation, as described previously. At the time of filming, he sat on a chair in front of a black backdrop at a quiet room. The Deaf professional started with a three-minute introduction of himself and provided some background information about his presentation, paused for a few seconds, and then delivered a 17-minute formal presentation in Auslan, using PowerPoint slides as prompts. He reported findings of a questionnaire-based survey study on the human rights situations of Deaf people around the world. Specifically, he presented results about the varying levels of recognition of signed languages, different methods of Deaf education, the provision of professional signed language interpreting services, and Deaf people’s accessibility to various government services in countries that had responded to the survey. A hearing heritage signer of Auslan served as an audience member for the filming of the Auslan presentation and also advanced the PowerPoint slides for the Deaf presenter during his talk. The researcher filmed the 20-minute Auslan monologue using a Flip UltraHD video camera, creating an MP4 video file. Due to environmental constraints (e.g., layout of the room, location of facilities, lighting, angle of filming), the camera recorded the Deaf presenter but did not capture his PowerPoint slides. Coincidentally, the Deaf professional had presented the same content by using the same PowerPoint slides at an international conference on human rights some months before creating this Auslan source text for the current study. His PowerPoint slides end with a reference, namely a website link to his co-authored research report on the research project (Haualand & Allen, 2009). It is unclear how

62 Methodology many interpreters in the present study used this link to access the report and read more about his research project in preparation for the interpreting assignment. The Auslan presentation was glossed in English, and an English model translation was produced to facilitate later data analysis. Sign notation conventions can be found in Appendix A. The hearing heritage signer who served as the Deaf presenter’s audience transcribed the Auslan monologue by annotating English glosses of the signs in ELAN (see Appendix B). The English gloss of a sign is the most commonly associated and/or closest translation of the sign (Johnston & Schembri, 2007). Another hearing native signer of Auslan who was an experienced professional Auslan/English interpreter (who did not participate in the study) provided the English model translation (see Appendix B). He first simultaneously interpreted the Auslan presentation into spoken English, and then produced the English model translation based on his English interpretation.The process of creating this English model translation differs from the process and outcome of producing a standard English translation with many edits available. The goal of this English model translation was to develop a marking transcript for the external raters that realistically reflected the message transfer capability of an interpreter simultaneously rendering the Auslan video into spoken English in real time.

3.2 Creating an Experimental Corpus The second step of the research project was to film professional Auslan/English interpreters simultaneously interpreting the Auslan monologue into spoken English. 3.2.1 The Interpreters Given its focus on conference-level signed language simultaneous interpreting, this study required that participants be NAATI-accredited professional-level Auslan/English interpreters.1 As noted earlier, although a total of 31 NAATIaccredited professional-level Auslan/English interpreters participated in my doctoral research project (Wang, 2013), the present analysis focuses on only 20 participants who scored over 65 (out of a maximum score of 100) on the Auslanto-English simultaneous interpreting task. This is because the sample size of 20 professional interpreters was large enough to show diversity in the English renditions but was still manageable for conducting in-depth analyses of the corpus. Given that only 160 Auslan/English interpreters held NAATI accreditation at the professional level in 2012 (Robert Foote, Accreditation Manager, NAATI, personal communication, 3 October 2012), the sample size of 20 professional Auslan/English interpreters in the current study is considered as being a representative sample (13%) in relation to that actual population size. The study attempted to include a cross-section of signed language interpreters who were representative of practitioners working in the field. Participants comprised 10 native signers (heritage signers) and 10 non-native signers of Auslan. The native signers were regarded as native users of Auslan and English, having acquired both languages from birth – the first from their signing Deaf parents at home, the second through communication with the hearing people2

Methodology

63

in their environment. The native signers stated that they had been using Auslan for an average of 39 years (standard deviation = 14 years). According to selfreports, the non-native signers acquired English from birth from their hearing parents; they often had no family connections with the Australian Deaf community, but started to learn Auslan between the ages of 16 and 24 by attending Auslan classes and/or by interacting with Deaf signers at work or through social networks. The non-native signers reported that they had been using Auslan for an average of 23 years (standard deviation = 12 years). The non-native signers were considered to be unbalanced bilinguals, with English as their native, dominant language and Auslan as their non-native language. Participants’ demographic and professional data is summarised in Table 3.1. Sixteen (80%) of the interpreters were female, and four were male. Fourteen interpreters (five native signers and nine non-native signers) held a bachelor’s,

Table 3.1  Twenty professional Auslan/English interpreters’ demographic and professional data Demographic/professional profile

20 participants 10 native signers

10 non-native signers

Female vs. male Mean age, in years (standard deviation) Mean years of signing Interpreters with a BA, MA, or PhD degree (%) Formally trained vs. untrained interpreters Mean years of holding NAATI Professional Interpreter accreditation (standard deviation) Mean years of paid interpreting experience (standard deviation) Number of interpreters with a minimum of 10 years of paid interpreting experience Mean years of conference interpreting experience (standard deviation) Mean number of conferences participants interpreted each year Full-time vs. part-time interpreters Mean hours of paid interpreting per month (standard deviation) Mean percentage of working from English into Auslan (standard deviation) Mean percentage of working from Auslan into English (standard deviation) Interpreters with no directionality preference Interpreters preferring Auslan-to-English interpreting Interpreters preferring English-to-Auslan interpreting Interpreters who were uncertain about directionality preference

16 vs. 4 40 (12) 31 (15) 14 (70%)

7 vs. 3 39 (14) 39 (14) 5 (50%)

9 vs. 1 41 (10) 23 (12) 9 (90%)

14 vs. 6 12 (8)

6 vs. 4 13 (9)

8 vs. 2 11 (8)

16 (8)

16 (8)

16 (8)

15 (75%)

8 (80%)

7 (70%)

12 (8)

12 (9)

11 (8)

5 (5)

6 (6)

3 (2)

11 vs. 9 70 (58)

7 vs. 3 89 (59)

4 vs. 6 51 (51)

62% (0.17)

66% (0.12) 58% (0.20)

38% (0.17)

34% (0.12) 42% (0.20)

13 3

7 (70%) 0 (0%)

6 (60%) 3 (30%)

3

2 (20%)

1 (10%)

1

1 (10%)

0 (0%)

64 Methodology master’s, or doctoral degree. Fourteen participants received formal interpreting training. Participants’ paid interpreting experience ranged from 3 to 25 years, with the average being 16 years. Although most participants reported that they had interpreted at conference settings, there were considerable individual differences in terms of both years of conference interpreting experience (the mean value being 12 years, and the standard deviation being 8) and the number of conferences they interpreted each year (the mean value being 5 conferences, and the standard deviation being 5). This remarkable variability partly results from the limited availability of conference interpreting work for signed (and spoken) language interpreters in Australia, where the majority of interpreting work is community interpreting. Moreover, participants stated that on average, 62% of their interpreting work was interpreting from English into Auslan, and that the other 38% was working from Auslan into English. In terms of directionality preference, 13 interpreters (seven native signers and six non-native signers) stated that they equally liked both language directions. The professional Auslan/English interpreters were residing in Sydney (N = 6), Melbourne (N = 11), Perth (N = 2), and Adelaide (N = 1) at the time of data collection. Due to marked linguistic variation in Deaf signers’ use of a signed language (Johnston & Schembri, 2007; Taylor, 2002), these geographical differences meant that many interpreters might have experienced difficulties in understanding the Sydney-based Deaf professional’s signs and signing style. Each interpreter was assigned a pseudonym (see Appendix C) for confidentiality purposes. 3.2.2 Auslan-to-English Interpretation Data Capture To replicate an authentic conference interpreting assignment and encourage pre-assignment preparation, the following materials were sent to participants prior to data collection: task instructions (e.g., conference theme, description of the intended audience, the Deaf presenter’s name and background, presentation title), glossary (fingerspelled words3 in the Auslan source text), and PowerPoint slides for the Auslan presentation. All participants provided a spoken English simultaneous interpretation of the Auslan presentation individually under approximately the same condition between March and July 2011. Upon arrival at the Deaf community service organisation in their respective cities, the interpreters were re-informed about the research project and asked to complete both a consent form and a demographic questionnaire. They were given an opportunity to ask questions about the research and read the preparation materials that had been distributed to them earlier. Each participant was filmed simultaneously interpreting the Auslan video shown on a computer into spoken English. Each participant watched the first three minutes of the Auslan video as a warm-up4 and then simultaneously interpreted the remaining 17 minutes of the formal presentation. The participants were advised that they could use a hard copy of the PowerPoint slides during the interpreting task if so desired. During the task, the researcher sat next to a Flip UltraHD video camera as the participant’s audience, coordinating the

Methodology

65

filming. The camera synchronously captured both the Deaf presenter on the computer screen and the participant’s spoken English interpretation. After the task, the researcher engaged each participant in a brief semi-­ structured interview on issues such as how s/he felt about the interpreting performance, challenges s/he perceived in the interpreting task, and interpreting strategies that s/he had employed during the task. The interview was audiorecorded for later analysis. For each interpreter, the total participation time was approximately an hour.This experimental corpus consists of 20 video clips (each video is approximately 17 minutes long) and, altogether, approximately 5.7 hours of Auslan-to-English simultaneous interpretation data.

3.3 Analytic Assessment of Simultaneous Interpreting Performance 3.3.1 External Raters Three external raters (named Morgan, Ashley, and Jamie) were selected to evaluate the corpus of Auslan-to-English simultaneous interpreting performance. Each rater met the following selection criteria: (i) native English speaker, (ii) hearing native signer, and (iii) NAATI-accredited professional Auslan/English interpreter with extensive interpreting experience in formal settings. In terms of assessment experience, both Morgan and Jamie were full-time signed language interpreter educators at a university in Australia, and they were experienced raters for national NAATI accreditation exams. Ashley was an experienced Auslan/ English interpreter and mentor of interpreters, who worked occasionally as a tutor in interpreter education programmes. Ashley had less assessment experience than Morgan and Jamie. Although it would be ideal to secure a large team of raters who meet all the selection criteria to the same degree, only a small pool of Auslan/English interpreters in Sydney possessed the requisite expertise. 3.3.2 Developing an Assessment Rubric An assessment rubric was developed specifically for this study based on a comprehensive review of the spoken and signed language interpreting studies literature as well as particular competencies being assessed in some teaching units of the Postgraduate Diploma in Auslan-English Interpreting at Macquarie University, Australia.5 This assessment tool suited the goals of evaluating participants’ Auslan-to-English interpreting performance and justifying the scores. The process of designing the assessment rubric (see Appendix D) entailed the following four steps: selecting assessment criteria, determining weightings of the assessment criteria, defining a rating scale, and writing level descriptors. 3.3.2.1 Selecting Assessment Criteria The assessment rubric includes the following four macro-level assessment criteria: accuracy, target text features, delivery features, and processing skills. This section

66 Methodology briefly defines these macro-level assessment criteria and their sub-criteria. It is important to note that each of these criteria is variably defined in the interpreting studies literature and warrants further research. Accuracy, also referred to as fidelity, faithfulness, equivalence, and completeness of message, is widely recognised as an essential aspect of spoken and signed language interpreting performance (see Bontempo & Hutchinson, 2011; Cokely, 1992; Gile, 2009; Lee, 2008; Liu, 2013; Pöchhacker, 2001; Russell & Malcolm, 2009; Taylor, 2002). Accuracy requires not only message equivalence at the ­lexico-semantic level, but also cohesion and coherence at the discourse level. In the present study, accuracy in interpretation was evaluated in terms of the following five sub-criteria: minimal unjustifiable additions, minimal unjustifiable omissions, minimal unjustifiable substitutions, minimal unjustifiable intrusions (source language interference), as well as cohesion and coherence. These sub-criteria generally resemble Cokely’s (1992, p. 75) miscue types (unjustifiable additions, omissions, substitutions, intrusions, anomalies) and Wang and Fang’s (2019, pp. 42–44) meaning-unit-based quality assessment framework. The sub-criteria are defined as follows. Unjustifiable addition refers to the interpreter’s addition of a considerable amount of new and meaningful information that cannot be found in the source text, causing the target language rendition to be inappropriately expanded in meaning. Unjustifiable omission refers to the interpreter’s omission of important information in the source text, causing serious meaning loss in the target text. Unjustifiable substitution refers to the interpreter’s distortion of a crucial source language message, resulting in a significant change in meaning. Unjustifiable intrusion refers to the interpreter’s strict adherence to source language syntactic structures, lexical choices, and/or cultural expressions, resulting in incomprehensible, ungrammatical, unacceptable, and unidiomatic target language utterances. Cohesion refers to the interpreter’s use of transition words (e.g., ‘and,’ ‘but,’ ‘because’) to connect adjacent ideas. Coherence indicates logical links that connect various ideas throughout an entire text or speech so as to form a unified whole. Target text features refer to the interpreter’s target language use being grammatically correct, idiomatic, natural, and appropriate to the specific context. In this study, target text features were evaluated in terms of the following three sub-criteria: grammaticality, suitable vocabulary, and conveying the signer’s register and affect. Register is a complex and dynamic phenomenon; it is conveyed through a combination of linguistic features (e.g., lexicon, syntax, phonology), paralinguistic characteristics (e.g., intonation, pausing, rate of speaking/signing), psychological factors (e.g., assertiveness, directness, nervousness), and sociological variables (e.g., politeness, respect, social status, conventions) (Shaw, 1987; Taylor, 2002). The register of a signed-to-spoken language interpretation needs to suit the Deaf signer’s message and intent, the context, and the audience (Taylor, 2002). A relatively formal register in spoken English was expected for the mock conference setting in the current study. Formal register entails salient features such as (i) fewer false starts, (ii) carefully constructed complete sentences, (iii) extended pauses, (iv) clear articulation and pronunciation, (v) more complex syntax, (vi)

Methodology

67

longer sentences using compound structures and subordinate clauses, and (vii) speaking in a deliberate manner (Taylor, 2002, pp. 115–116). Efficacy in conveying and shifting register not only requires an interpreter to have generic skills such as broad general knowledge, significant exposure to different settings and various people, and high bilingual and bicultural competencies, but also demands the interpreter to consider the unique facets of an interpreting assignment such as the speaker’s/signer’s communicative goal, the context, the rapport between the speaker/signer and the audience, and the audience’s needs (their comprehension level, familiarity with the subject matter, knowledge about technical terms, etc.) (Taylor, 2002). Delivery features refer to public speaking skills that reflect whether the target language interpretation is pleasant to listen to (Lee, 2008; Taylor, 2002). Gile (2009), Lee (2008), and Taylor (2002) maintain that good voice projection, clear articulation, vivid intonation, appropriate pace, smooth delivery, and pleasant style have a confidence-assuring effect and successfully convey the source ­speaker’s/signer’s messages to an audience. In the present study, delivery features were assessed according to the following three sub-criteria: clarity of articulation, fluency, and natural prosody. Processing skills or strategies refer to the interpreter’s strategic management of information during the simultaneous interpreting process. In this study, processing skills were assessed according to the following five sub-criteria: processing time, a free-literal interpretation continuum with appropriate shifts in between, strategic additions, strategic omissions, and successful anticipation. For a definition and previous findings regarding processing time, see Section 2.6 in Chapter 2. In regard to a free-literal interpretation continuum (for further information, see Section 2.8.2 in Chapter 2), Napier (2016) found that skilled Auslan/English interpreters strategically switched between a free interpretation style and a literal interpretation style while interpreting an English university lecture in order to provide accurate and effective Auslan renditions. Conscious strategic omissions are defined by Napier (2016, p. 64) as ‘omissions made consciously by an interpreter, whereby a decision is made to omit information in order to enhance the effectiveness of the interpretation.’ The interpreter uses his or her linguistic and cultural knowledge to determine what source language information is meaningful in the target language, what information is culturally relevant, and what information is repetitive (Napier, 2016). Strategic additions refer to additions made deliberately by the interpreter in order to ensure the effectiveness of meaning transfer, enhance cultural understanding, maximise utterance design, and fulfil particular communicative goals (Angelelli, 2012; Baraldi, 2012; Braun, 2017; Jacobsen, 2002; Kohn & Kalina, 1996; Major & Napier, 2012; Merlini & Favaron, 2005; Wadensjö, 1998; Wang & Fang, 2019). Successful anticipation refers to the interpreter making linguistic and extralinguistic prediction by virtue of high linguistic proficiency, preassignment preparation, familiarity with speakers and subject matters, domain knowledge, encyclopedic knowledge, and relevant prior interpreting experience (Gile, 2009).

68 Methodology 3.3.2.2 Determining Weightings Once the macro-level assessment criteria and sub-criteria were decided, the next step was to allocate weightings across them.Weightings in the current rubric were partially informed by Lee’s (2008) study. Lee assigned a weighting of 40% to accuracy, 40% to target language quality, and 20% to delivery; however, it is not clear why those particular weightings were initially decided. Five out of nine professional interpreter raters in Lee’s study later commented that accuracy should have been given a weighting of 50%. The majority of raters also recommended that a higher weighting should have been distributed to target language quality than to delivery. Therefore, I decided to allocate a weighting of 50% to accuracy, 20% to target text features, 15% to delivery features, and 15% to processing skills. 3.3.2.3 Defining the Rating Scale Once the weightings were assigned, five performance levels were nominated. Two of the three proposed raters were academics in an Australian university, and thus were familiar with their institution’s standard grading scale: Fail (0–49), Pass (50–64), Credit (65–74), Distinction (75–84), and High Distinction (85–100).To maximise the feasibility of the rubric in this study, the five performance levels for each macro-level assessment criterion were matched as follows: Poor (less than 50%, Fail), Average (50–64%, Pass), Good (65–74%, Credit), Very Good (75–84%, Distinction), and Excellent (85–100%, High Distinction). 3.3.2.4 Writing Level Descriptors Once the rating scale was defined, the final step was to write descriptors for the five levels of performance against the four macro-level assessment criteria. Descriptors that are too brief and general allow for greater ambiguity, while descriptors that are too long and specific are time-consuming for raters to follow (Lee, 2008). 3.3.3 External Raters’ Analytic Assessment Process In my doctoral research, when the three raters were initially contacted, only Ashley had sufficient time to complete the analytic assessment of all 31 participants’ simultaneous interpreting performance in both language directions. Ashley completed the evaluation work in July 2012. For inter-rater reliability purposes, Jamie replicated the analytic assessment on 16 randomly selected participants during July and August 2012. As remarkable variability between Ashley’s grades and Jamie’s scores regarding participants’ English-to-Auslan interpreting performance was observed, Morgan was subsequently invited to assess all 31 participants’ interpreting performance in both language directions during August and November 2012. As mentioned earlier in this chapter, the Auslan-to-English interpretation data of only 20 out of the 31 interpreters was selected for the current study. As a result, both Morgan and Ashley assessed all 20 interpreters’

Methodology

69

Auslan-to-English interpreting performance, and Jamie evaluated 11 (55%) of the 20 interpreters’ performance (see Table 4.2 in Chapter 4 for the raters’ scores). McNamara (1996, 2000) recommends that raters should receive sufficient rater training in the form of moderation meetings to ensure optimal inter-rater reliability and intra-rater reliability; however, for logistical reasons and the different timelines of the raters’ involvement in the assessment process, group moderation for all three raters before the analytic assessment process was not possible (a limitation of the present study). Instead, the raters underwent the same individual training from the researcher, in which they were informed about the purpose of the assessment (i.e., to evaluate participants’ competency to provide accurate and effective simultaneous interpretations for the intended conference audience), the Auslan source text, the English model translation, the Auslan-toEnglish interpretation videos, and the assessment rubric. Each rater followed the same analytic assessment procedure. When evaluating participants’ Auslan-to-English simultaneous interpreting performance, each rater listened to participants’ English renditions while taking notes on the English model translation (Appendix B) of the Auslan source text, only occasionally watching the Auslan presentation video to check particular signs or other aspects. It is worth noting that the raters’ over-reliance on the English model translation in terms of lexical selection and style might have influenced their expectations of participants’ interpretations. The raters assigned a score to each of the four macro-level assessment criteria on the rubric. The four sub-scores were then added up to represent the total score for each participant, with the maximum possible total score being 100. Due to their limited availability, each rater double-rated only one or two participants; therefore, discussion of intra-rater reliability is beyond the scope of this book, even though it is an ideal method. In each assessment session, the researcher sat next to the rater, monitoring the rating process and recording the rater’s oral comments for later analysis. Audio-recordings of all three raters’ oral comments made during the analytic assessment were fully transcribed to inform microanalysis of the experimental corpus. As noted earlier in this chapter, the current study focuses on only 20 participants who scored higher than 65 on the Auslan-to-English simultaneous interpreting task. In other words, these interpreters fell into the Good (65–74, Credit), Very Good (75–84, Distinction), and Excellent (85–100, High Distinction) categories. After obtaining the external raters’ scores and transcribing their oral comments, the next step was to conduct microanalysis of the interpretation corpus.

3.4 Microanalysis of the Interpretation Corpus Using ELAN and Excel I analysed the experimental corpus of Auslan-to-English simultaneous interpretation data by drawing on the three external raters’ oral and written comments made during the analytic assessment and on consultation provided by a NAATIaccredited professional Auslan/English interpreter who did not participate in

70 Methodology the study.The consultant had used ASL for 13 years since the age of 12; she then moved from the United States to Australia and acquired Auslan from Deaf signers in a signing work environment, where it took her approximately 10 years to achieve near-native proficiency in Auslan. She had 36 years of paid Auslan/ English interpreting experience, completed formal interpreting training at a university in Australia, and held NAATI Professional Auslan/English Interpreter accreditation for 15 years.This research assistant’s key roles in the current research project included enhancing my understanding of the Auslan monologue and responding to my queries about the quality of particular English renditions. The microanalysis of the interpretation corpus comprised six steps. First, all 20 participants’ Auslan-to-English interpretations and interview recordings were transcribed verbatim. Second, based on prosodic markers and utterance boundaries in signed languages (e.g., signs such as so and but, pauses, held handshape, hand clasp, hands drop, head tilt, head nod, headshake, eye aperture, body lean, body movement; Nicodemus, 2009), the Auslan monologue was segmented into 114 sentences (see Appendix B, in which double slashes [//] represent sentence boundaries), with each Auslan sentence serving as a unit for my microanalysis. It merits attention that concepts such as sentences, utterance boundaries, and discourse markers in signed languages are severely underexplored, appear far more complex than those in spoken languages, and warrant further research (Nicodemus, 2009). Segmentation of the Auslan source text was checked by the research assistant. Subsequently, Auslan sentences were matched with their corresponding English model translations (see Appendix B) to facilitate the microanalysis of the corpus. Third, the video footage of each participant’s Auslan-to-English interpreting was imported into ELAN (a video annotation software program) for in-depth analysis.Time-coded annotations were inserted into the video footage in ELAN to investigate both the quality of participants’ Auslan-to-English interpretations and the cognitive process that underlay effective and ineffective renditions. As shown in Figure 3.1, tiers regarding the aforementioned macro-level assessment criteria and sub-criteria were used to capture various positive and negative aspects of participants’ Auslan-to-English interpretations such as strategic decision-making (e.g., strategic additions, strategic omissions), literal versus free interpretation style, processing time, miscues (e.g., unjustifiable omissions, unjustifiable distortions, source language intrusion), target speech characteristics (e.g., lexical choices, ungrammatical sentences), and public speaking skills (e.g., self-­ correction, fluency). Apart from the macro-level assessment criteria and sub-criteria in the rubric (see Appendix D), new tiers were incorporated into the ELAN analysis (see Figure 3.1), including, among others: (i) direct equivalence, an interpretation that by and large conveys the meaning of a source language message literally and faithfully, similar to close interpretation (Wang & Fang, 2019) and close rendition (Wadensjö, 1998); (ii) minor addition, which is an interpreter’s inappropriate or unnecessary addition or repetition of a small piece of meaningful information that results in the target rendition being slightly expanded in content and/or

Methodology

71

Figure 3.1  ELAN working file showing annotation tiers

strengthened in illocutionary force; (iii) minor omission, that is, the interpreter’s omission of a small meaning unit in the source text, thus causing slight reduction in meaning; and (iv) minor distortion, an interpretation that slightly differs in meaning from the original message and thus may be deceptive for the audience. This sentence-level comparison of the source text and the target text in ELAN systematically captured the strengths and weaknesses of each participant’s Auslanto-English interpreting performance; however, this method did not allow a direct comparison of the 20 interpreters’ English renditions of the same Auslan sentence. Fourth, to resolve that problem, Excel spreadsheets were created to map each Auslan sentence with their 20 corresponding English interpretations (see Figure  3.2). This analysis not only revealed how different interpreters coped with the same problem triggers in the Auslan source text (e.g., numbers, dense information, syntactic differences between Auslan and English, ambiguity), but also facilitated the evaluation of sentence-level accuracy, namely whether an interpreter faithfully transferred the core meaning (sense, proposition, or who does what to whom) of an Auslan sentence into the target language. My research assistant and I considered both the context and communicative goals when determining interpreters’ sentence-level accuracy. Nevertheless, simultaneous interpreting is not sentence-level translation. This approach overlooked discourse-level coherence and provided limited insights into potential reasons for (in)accurate renditions. Hence, it was necessary to return to ELAN files to analyse the interpreting process of particular renditions. Fifth, after identifying the common issues in interpreters’ English renditions of the Auslan monologue, I used ELAN to analyse their cognitive process of interpreting extremely challenging Auslan sentences or segments, in an attempt

72 Methodology

Figure 3.2  Excel spreadsheet showing microanalysis of Auslan sentence 36 and its 20 English interpretations

to identify whether the interpreters’ inaccurate renditions resulted from problems regarding their comprehension of the Auslan source text, working memory functions, speech production, or coordination, among others. Finally, to address the disadvantages of sentence-level analyses of the Auslanto-English simultaneous interpretation data, I investigated interpreters’ English renditions of adjacent Auslan sentences, large sections of the Auslan source text, and logical links throughout the Auslan monologue. Combining the sentencelevel analyses with the discourse-level analyses shed light on whether some omissions were simply errors or strategic decisions. In particular, I conducted local analysis to investigate cognitive overload and coping strategies regarding salient challenges in the source text. Local analysis refers to an in-depth analysis of the cognitive processing of a sequence of two or three adjacent sentences in order to shed light on how the processing of the current sentence influences the processing of its neighbouring sentences (Gile, 2008). If an interpreter is finishing off his/her interpretation of the previous sentence while comprehending the beginning of the current sentence, this overlapping of sub-processes generates imported cognitive load; processing of the current sentence causes current cognitive load; and if the interpreter is finishing off rendering the current sentence into the target language while attending to the first few words of the subsequent sentence, this phenomenon creates exported cognitive load (Gile, 2008). Gile (2008) maintains that it is important to investigate the impact of the interpreter’s processing time in completing the processing of one sentence (i.e., tail-to-tail span, the time interval between the ending point of the source language sentence and the ending point of its

Methodology

73

target-language rendition) on the processing of the subsequent sentence, as excessive tail-to-tail span may culminate in cognitive saturation. Gile (2008, p. 59) points out that local analysis may provide explanations that overall analysis of a whole speech may fail to reveal, and therefore it ‘may increase the power of empirical research on interpreting.’ The process of local analysis in the present study involved the following steps: 1 select particular source language segments whereby most interpreters made significant errors (e.g., inaccurate renditions of numbers, distortions of key messages, omissions of whole sentences), 2 decide which participants’ interpretations will be used for the detailed analysis, depending on the researcher’s purpose of the analysis, 3 put the selected source language segment and the selected interpreter’s target language renditions in a Microsoft Word document, 4 number the source language sentences, add double slashes between the sentences, and insert a blank line after a pair of source language line and target language line in order to make the transcript easy to read, highlight the issues for discussion (e.g., numbers) by making them bold, 5 watch the selected segment of the participant’s interpretation video in ELAN to figure out how the source language lexical items align with the target language lexicons temporally in ELAN and spatially in the Microsoft Word document, 6 measure the processing time of particular items in ELAN (e.g., numbers, technical terms, sentence beginnings, sentence ends), 7 gauge the length of notable pauses both in the source language speech/ signed monologue and target language speech in ELAN, as they may provide insight into the interpreter’s processing speed or cognitive processing, 8 read the aligned transcript to compare the source text with the target text in order to pinpoint the interpreter’s processing units (e.g., phrase, clause, sentence), to identify interpretation errors and strategic decisions (e.g., using strategic additions, making strategic omissions, anticipating incoming information, summarising, switching between a free interpretation style and a literal interpretation style), and to posit potential reasons behind the errors and strategic decisions, 9 examine the effects of strategic additions and long processing time on both the cognitive processing and interpreting quality of the subsequent sentence components or sentences, 10 investigate what strategies the interpreter has adopted in order to catch up with the source language speaker/signer or to overcome interpreting difficulties, and 11 describe the local analysis of the interpreting process sentence by sentence. I used this local analysis to present and discuss the examples in Chapters 5, 6, and 7.

74 Methodology

3.5 Thematic Analysis of the Retrospective Interview Data I analysed participants’ retrospective interview data by using thematic analysis in NVivo.6 Thematic analysis involves identifying the recurring themes of textual data, coding and classifying the data according to those themes, and interpreting the resulting thematic structures by looking for commonalities, relations, and overarching patterns. The thematic analysis of the post-task interview data focused on two main areas: interpreters’ self-perceived challenges in working simultaneously from Auslan into English and their coping strategies. Themes that emerged from the qualitative interview data in turn informed the microanalysis of the Auslan-to-English simultaneous interpretation corpus by narrowing down areas of interest.

Notes 1 In Australia, all spoken and signed language interpreters and translators were accredited (before 1 January 2018) and are certified (after 1 January 2018) through the National Accreditation Authority for Translators and Interpreters (NAATI). At the time of the data collection (year 2011) there were only two accreditation levels available for Auslan/English interpreters – Paraprofessional Interpreter and Professional Interpreter, equivalent to NAATI’s current Certified Provisional Interpreter and Certified Interpreter, respectively. 2 It is important to note that not all hearing children with deaf parents are signers or native signers, because some deaf parents speak with their hearing children. 3 Johnston and Schembri (2007, p. 34) define fingerspelling as ‘the use of hand configurations to represent the letters of a writing system.’ Regarding fingerspelling, Deaf Auslan signers use a two-handed alphabet to spell selected words out manually, letter by letter. English words are often borrowed when there are no Auslan lexical equivalents, and so they need to be spelled on the fingers. In addition, Deaf signers who are bilingual in a signed language and a written language sometimes choose to use particular spoken language lexical items and therefore fingerspell them. 4 It was each interpreter’s choice to just watch the three-minute introduction silently or to interpret it for practice. 5 The particular competencies were determined through an iterative process of trialling assessment rubrics against student performance outcomes and observations. 6 NVivo software is widely used to analyse qualitative interview data in qualitative research and mixed methods research. NVivo: Visited on 26 October 2019, https:// www.qsrinternational.com/nvivo/nvivo-products.

References Angelelli, C. V. (2012). Challenges in interpreters’ coordination of the construction of pain. In C. Baraldi & L. Gavioli (Eds.), Coordinating Participation in Dialogue Interpreting (pp. 244–260). Amsterdam: John Benjamins. Baraldi, C. (2012). Interpreting as dialogic mediation: The relevance of expansions. In L. Gavioli (Ed.), Coordinating Participation in Dialogue Interpreting (pp. 297–326).Amsterdam: John Benjamins. Bontempo, K., & Hutchinson, B. (2011). Striving for an ‘A’ grade: A case study in performance management of interpreters. International Journal of Interpreter Education, 3, 56–71.

Methodology

75

Braun, S. (2017). What a micro-analytical investigation of additions and expansions in remote interpreting can tell us about interpreter’s participation in a shared virtual space. Journal of Pragmatics. Special Issue ‘Participation in Interpreter-Mediated Interaction’ 107, 165–177. doi: 10.1016/j.pragma.2016.09.011. Cokely, D. (1992). Interpretation: A Sociolinguistic Model. Burtonsville, MD: Linstok Press. Gile, D. (2008). Local cognitive load in simultaneous interpreting and its implications for empirical research. Forum, 6(2), 59–77. doi: 10.1075/forum.6.2.04gil. Gile, D. (2009). Basic Concepts and Models for Interpreter and Translator Training (Rev. ed.). Amsterdam/Philadelphia: John Benjamins. Haualand, H., & Allen, C. (2009). Deaf people and human rights. Retrieved from https:// www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=2ahUKEw jMsc7KtOTnAhWCYisKHWfiBGAQFjABegQIBRAB&url=https%3A%2F%2F www.rasit.org%2Ffiles%2FDeaf-People-and-Human-Rights-Report.pdf&usg=AOv Vaw1JAQTTPOPRAQV7ZVcQ6X_E. Accessed on February 22, 2020. Jacobsen, B. (2002). Pragmatic meaning in court interpreting: An empirical study in consecutively interpreted question-answer dialogues. (Unpublished doctoral dissertation). The Aarhus School of Business, Aarhus, Denmark. Johnston, T., & Schembri, A. (2007). Australian Sign Language (Auslan): An Introduction to Sign Language Linguistics. Cambridge, UK: Cambridge University Press. Kohn, J., & Kalina, S. (1996). The strategic dimension of interpreting. Meta, 41(1), 118– 138. doi: 10.7202/003333ar. Lee, J. (2008). Rating scales for interpreting performance assessment. The Interpreter and Translator Trainer, 2(2), 165–184. doi: 10.1080/1750399X.2008.10798772. Liu, M. (2013). Design and analysis of Taiwan’s interpretation certification examination. In D. Tsagari & R. van Deemter (Eds.), Assessment Issues in Language Translation and Interpreting (pp. 163–178). Frankfurt, Germany: Peter Lang. Major, G., & Napier, J. (2012). Interpreting and knowledge mediation in the healthcare setting: What do we really mean by ‘Accuracy’? Linguistica Antverpiensia, 11, 207–225. McNamara, T. (1996). Measuring Second Language Performance. New York: Longman. McNamara, T. (2000). Language Testing. Oxford, UK: Oxford University Press. Merlini, R., & Favaron, R. (2005). Examining the ‘Voice of Interpreting’ in speech pathology. Interpreting, 7(2), 263–302. doi: 10.1075/bct.9.08mer. Napier, J. (2016). Linguistic Coping Strategies in Sign Language Interpreting.  Washington, DC: Gallaudet University. Nicodemus, B. (2009). Prosodic Markers and Utterance Boundaries in American Sign Language Interpretation. Washington, DC: Gallaudet University Press. Pöchhacker, F. (2001). Quality assessment in conference and community interpreting. Meta, 46(2), 410–425. doi: 10.7202/003847ar. Russell, D., & Malcolm, K. (2009). Assessing ASL-English interpreters: The Canadian model of national certification. In C. V. Angelelli & H. E. Jacobson (Eds.), Testing and Assessment in Translation and Interpreting Studies (pp. 331–376). Amsterdam/Philadelphia: John Benjamins. Shaw, R. (1987). Determining register in Sign-to-English interpreting. Sign Language Studies, 57(Winter), 295–322. doi: 10.1353/sls.1987.0019. Taylor, M. M. (2002). Interpretation Skills: American Sign Language to English. Edmonton, Alberta: Interpreting Consolidated. Wadensjö, C. (1998). Interpreting as Interaction (1st ed.). London/New York: Longman. Wang, J. (2013). Working memory and signed language interpreting. (Unpublished doctoral dissertation). Macquarie University, Sydney, Australia.

76 Methodology Wang, J. (2016). The relationship between working memory capacity and simultaneous interpreting performance: A mixed methods study on professional Auslan/English interpreters. Interpreting, 18(1), 1–33. doi: 10.1075/intp.18.1.01wan. Wang, J., & Fang, J. (2019). Accuracy in telephone interpreting and on-site interpreting: A comparative study. Interpreting, 21(1), 36–61. doi: 10.1075/intp.00019.wan. Wang, J., & Napier, J. (2013). Signed language working memory capacity of signed language interpreters and deaf signers. Journal of Deaf Studies and Deaf Education, 18(2), 271–286. doi: 10.1093/deafed/ens068. Wang, J., & Napier, J. (2015a). Directionality in signed language interpreting. Meta, 60(3), 518–541. doi: 10.7202/1036141ar. Wang, J., & Napier, J. (2015b). Measuring bilingual working memory capacity of professional Auslan/English interpreters: A comparison of two scoring methods. The Interpreters’ Newsletter 19, 45–62. Wang, J., Napier, J., Goswell, D., & Carmichael, A. (2015). The design and application of rubrics to assess signed language interpreting performance. The Interpreter and Translator Trainer, 9(1), 83–103. doi: 10.1080/1750399X.2015.1009261.

4 Quality, Self-perceived Challenges, and Self-reported Strategies

This chapter addresses the first research question of this study, that is, ‘What are the challenges in simultaneous interpreting from a signed language (Auslan) into a spoken language (English) at a formal setting?’ The chapter first presents findings about the three external raters’ analytic assessment of the 20 professional interpreters’ Auslan-to-English simultaneous interpreting performance, and then reports on findings regarding the participants’ self-perceived challenges in this interpreting task and their self-reported coping strategies. It is noteworthy that the identification of challenges in the Auslan-to-English simultaneous interpreting task carried a positive connotation because it paved the way for subsequent analyses of how the interpreters actually dealt with those difficulties.

4.1 External Raters’ Scores for the Interpreters The three external raters’ scores for the participants’ performance on the Auslanto-English simultaneous interpreting task are shown in Figure 4.1. The wavy lines reveal the raters’ marked variability in rating the interpreters in terms of the quality of their interpretations. For example, Morgan’s ratings ranged from 66.6 (for Monica) to 95 (for Emily). While Ashley’s ratings of three interpreters (Tiffany, Kay, and Liz) were lower than 65, both Morgan and Jamie’s scores for the three interpreters were higher than 65; therefore, these three interpreters were still included in the sample. This figure shows remarkable homogeneity across all three raters for some interpreters (e.g., Linda) but noticeable differences in their ratings for others (e.g., Kay). Inter-rater reliability was then calculated for the raters. Although inter-rater reliability of 0.70 represents the minimum level of acceptable consistency between a pair of raters, inter-rater reliability of 0.90 is a higher level of consistency between them, as it represents 81% agreement and 19% disagreement (McNamara, 2000). For comparability reasons, inter-rater reliability in the current study – represented by Pearson correlation coefficients1 – focused only on the 11 participants who had been assessed by all three raters. Table 4.1 shows inter-rater reliability for Auslan-to-English simultaneous interpreting performance. Surprisingly, with regard to the total score, the

78 Quality, Challenges, and Strategies Morgan (educator)

100

Jamie (educator)

Ashley (praconer)

95 90 85 80 75 70 65 60 55

Zoe

Bernie

Debbie

Alex

Annie

Claire

Sophia

Liz

Jane

Kay

Lauren

Tiffany

Monica

Shannon

Helen

Linda

Molly

Emily

Wendy

Amber

50

Figure 4.1  Three external assessors’ scores for participants’ Auslan-to-English simultaneous interpreting performance

Table 4.1  Inter-rater reliability regarding Auslan-to-English simultaneous interpreting performance Inter-rater reliability Total score

Sub-scores Accuracy Target text features Delivery features Processing skills

Morgan and 0.75** 0.78** Jamie (N = 11) Morgan and 0.82** 0.77** Ashley (N = 11) Jamie and Ashley 0.50 0.52 (N = 11)

−0.02

0.63*

0.64*

0.65*

0.75**

0.94***

−0.32

0.59

0.66*

N = 11 refers to the 11 participants who were evaluated by all three external raters. *** p < 0.001. ** p < 0.01. * p < 0.05.

inter-rater reliability between Morgan (an interpreter educator) and Ashley (an interpreter practitioner) (0.82) was higher than that between the two interpreter educators Morgan and Jamie (0.75), which in turn was higher than that between Jamie and Ashley (0.50). This pattern held true for assessment criteria such as target text features and delivery features, but not for accuracy and processing skills. The consistency between Morgan and Ashley was satisfactory for the total score as well as all four sub-scores. Interestingly, the consistency between the two interpreter educators – Morgan and Jamie – was acceptable for both the total

Quality, Challenges, and Strategies 79 score and accuracy, but not for the other three sub-scores. In contrast, the ­consistency between Jamie and Ashley was satisfactory only for processing skills. The poor inter-rater reliability between Jamie and Ashley may have resulted from the small number of participants (11 out of 20) assessed by all the three raters; in my larger study in which the three raters evaluated 16 out of 31 participants, the inter-rater reliability was typically above 0.80 (see Wang, Napier, Goswell, & Carmichael, 2015). It is recognised that raters’ scores may be influenced by such factors as their educational and professional background, assessment experience, perspectives on the rating task, and scoring techniques (Wang et al., 2015), but the current study focused on examining the process and product of professional interpreters’ signed-to-spoken language simultaneous interpreting rather than on accounting for inter-rater variability.

4.2 Native Signers Versus Non-native Signers This section presents findings from a comparison between the 10 native signers and the 10 non-native signers in terms of Auslan-to-English simultaneous interpreting performance. Table 4.2 shows each participant’s total score assigned by the external raters as well as his/her average total score. Participants’ average total scores (see the far-right column in Table 4.2) were used for comparing the native signers with the non-native signers in terms of

Table 4.2  Participants’ Auslan-to-English simultaneous interpretation scores Pseudonyms

Native or non-native signer

Morgan’s scores

Jamie’s scores

Ashley’s scores

Average total score

Amber Emily Wendy Linda Molly Helen Shannon Tiffany Monica Lauren Kay Liz Jane Sophia Claire Alex Annie Debbie Zoe Bernie

Native Native Non-native Non-native Non-native Non-native Native Native Native Native Native Non-native Non-native Non-native Non-native Native Native Non-native Native Non-native

92.5 95 87.5 78.5 92 87 85.3 66.8 66.6 89 70.8 78 93.5 85.5 77 83 82.9 82 92 80

89.5 89 86.5 77.5 91 88 85 74.5 81.8 77.5 78

88 89 76 77 76 78 90 62 70 89 64 54 90 79 70 82 71 66 83.5 74

90 91 83 78 86 84 87 68 73 85 71 66 92 82 74 83 77 74 88 77

80 Quality, Challenges, and Strategies Table 4.3  Native signers versus non-native signers in terms of Auslan-to-English simultaneous interpreting performance Analytic assessment

Total score Accuracy Target text features Delivery features Processing skills

Native signers (N = 10)

Non-native signers (N = 10)

Mean (SD)

Mean (SD)

81.30 (8.39) 40.83 (4.88) 16.21 (1.07) 12.09 (1.15) 12.04 (1.49)

79.60 (7.37) 39.58 (4.22) 16.21 (1.20) 12.08 (1.15) 11.75 (1.22)

t

df

p

η2

0.48 0.61 -0.003 0.03 0.47

18 18 18 18 18

0.64 0.55 0.997 0.98 0.64

-

Note: All p values were two-tailed. Effect size η2 was calculated only when there was a significant difference (p ≤ 0.05).

performance quality. Results of independent-samples t-tests are presented in Table 4.3.The native signers performed similarly to the non-native signers on the Auslan-to-English simultaneous interpreting task in terms of the total score (overall performance), t(18) = 0.48, p = 0.64. In addition, the native signers were similar to the non-native signers in terms of accuracy, target text features, delivery features, and processing skills, with all p values larger than 0.05. Intriguingly, the same result patterns persisted when only Morgan’s scores were used for the comparison, and when only Ashley’s scores were used for the statistical analysis. In view of these results, hereafter participants will be simply referred to as interpreters rather than native or non-native signers unless the distinction is necessary for emphasis.

4.3 Self-perceived Challenges This section reports on findings of the interpreters’ perceived challenges in simultaneous interpreting from a signed language (Auslan) into a spoken language (English) in the mock conference setting. It draws on a thematic analysis in NVivo of the retrospective interview data that formed an important part of the study. Inspired by Gile’s (2009) Effort Model of simultaneous interpreting, the challenges that emerged from the qualitative interview data were grouped into four categories (see Table 4.4): (i) signed language comprehension problems, (ii) cognitive overload, (iii) target speech production difficulties, and (iv) the artificial t­ esting situation. As typical in studies on interpreting, these themes are multidimensional and interconnected. Representative comments by participants are incorporated into the sections to illustrate the challenges they faced. Each interpreter’s pseudonym, educational, and professional information is given in Appendix C. 4.3.1 Signed Language Comprehension Problems As shown in Table 4.4, a key finding is that as many as 80% of interpreters (7 native signers and 9 non-native signers out of 20 participants) reported that they had varying degrees of difficulties understanding this particular Deaf professional’s signed

Quality, Challenges, and Strategies 81 Table 4.4  Interpreters’ self-perceived challenges in the Auslan-to-English simultaneous interpreting task Signed language comprehension problems (16, 80% of 20) • Unfamiliarity with the Deaf presenter (11, 55%) • Unfamiliarity with his particular signing style (8, 40%) • Unknown/unfamiliar signs and sign variation (13, 65%) • Unfamiliarity with the subject matter (9, 45%) • Unclear information in the Auslan monologue (4, 20%) Cognitive overload (15, 75%) • Numbers (14, 70%) • Long processing time (9, 45%) • Syntactical differences between Auslan and English (6, 30%) • Information density in the Auslan monologue (2, 10%) Target speech production difficulties • Trying to produce comprehensible English sentences (6, 30%) • Using a formal register in English (e.g., lexical choices, syntactic structures) (10, 50%) • Maintaining coherence in the target language speech (7, 35%) The artificial testing conditions • Two-dimensional video rather than a live presentation (8, 40%) • No chance to meet the Deaf presenter beforehand (4, 20%) • No opportunity to cooperate with the Deaf presenter during interpreting (3, 15%) • No team interpreter (8, 40%) • No PowerPoint slides on show (9, 45%) • No real audience (3, 15%)

monologue. The interpreters’ comprehension problems primarily stemmed from their unfamiliarity with the Deaf presenter, unfamiliarity with his specific signs and considerable sign variation, unfamiliarity with his particular signing style, unfamiliarity with the subject matter, and some ambiguous segments in the Auslan source text. 4.3.1.1 International Sign, Unfamiliar Signs, and Sign Variation Thirteen participants reported that they did not understand or were ‘thrown’ by the Deaf presenter’s use of specific lexical signs taken from International Sign,2 unfamiliar jargon signs that were specific to the presentation topic, signs that they had never seen before, and sign variations. Although this Deaf professional had been using Auslan since birth and was working in Sydney at the time, he had worked overseas for many years, had travelled around the world, had been deeply involved in the World Federation of the Deaf (WFD), and had contributed to the United Nations Convention on the Rights of Persons with Disabilities (UNCRPD).Therefore, he had substantial contact with other signed languages (e.g., ASL, BSL) and International Sign. An analysis of the Auslan

82 Quality, Challenges, and Strategies source text revealed that the Deaf presenter incorporated some lexical items from International Sign (e.g., international sign, world federation of the deaf [wfd], u-n convention3 [followed by fingerspelling], articles [in UNCRPD], volunteer, human rights), ASL signs (east, west, south, southamerica, basic4 (preceded by the synonymous Auslan sign minimum), access5 (followed by fingerspelling), total communication [followed by ­fingerspelling]), depicting signs (e.g., using iconic gesture to reference the location of southamerica before using the ASL sign, using his hands to outline the shape of the Caribbean Sea after fingerspelling c-a-r-i-b-b-e-a-n-s), signs in Auslan that may be used more typically by individuals of his generation (e.g., both for bilingual), and many sign variations, which may be due to his use of the northern dialect of Auslan and his particular signing style (e.g., concept produced with a varied initial location and movement, education produced in front of his chest rather than forehead, s interpreter rather than sign language interpreter). Many of the interpreters in this study reported not being familiar with the Deaf presenter, a factor that influenced their ability to understand his signed discourse. As a result, the interpreters needed to adjust to his sign choices and signing style, and to draw on the context in order to understand signs that were unfamiliar to them. Some interpreters reported that they were unfamiliar with the Deaf presenter’s sign choices or sign variations for lexical items such as articles, approve, access, government, concept, country, and convention. Interpreters such as Annie reported feeling frustrated about not understanding these recurrent signs until figuring them out from the context later on: Signs like articles, I made an error at the start. I thought he was saying something else and then I realised he meant ‘articles.’ Liz reported seeing unfamiliar signs (from ASL) and only deciphered the sign approve when consulting the relevant printed PowerPoint slide: east, west, and south [were new to me]. There is a lot of International Signs thrown in there. … At one point, I actually had no idea of what he was saying. He was using a sign that means to ‘approve something, or to pass an exam.’ And it wasn’t till I used the notes here to realise he was talking about levels of legal recognition of sign language. Moreover, Tiffany commented on suppressing her first response when seeing the Deaf professional’s sign for government, which was her sign for minister: So, the first word that pops into my head is ‘minister,’ and I know he is not talking about ‘minister,’ he is talking about ‘government.’ So you get sort of caught on. The Deaf professional’s particular signing style (idiolect) also posed as a challenge for many interpreters. For example, Sophia did not work out the meaning

Quality, Challenges, and Strategies 83 of his use of the non-standard sign for concept until almost the end of his presentation: I wasn’t sure exactly what he was signing, because I signed concept a little bit differently. And it wasn’t until right near the end that I realised, ‘Oh, that’s concept!’ Furthermore, Alex commented on the Deaf professional’s use of signs from International Sign, foreign signs, and sign variation regarding education: He is idiosyncratic. … He threw me from time to time. … And because he has travelled all around the world and lived in different countries, it has different influences on his production of Auslan. He catches himself doing it, and then he will remedy what he has just signed in International Sign or some other sign language, and then he will spell the word or use the Auslan sign for that. … He’s using a sign for education, but it looks like swimming. It’s most unusual. When I saw that, I laughed because I understand. But had I not known that, I would have to try to work it out from context. Helen also highlighted the need for her to adjust to the Deaf presenter’s sign variations and sign choices: He definitely has a quite distinct style. For example, signs like country 1, he does it as country 2, so it’s almost like a completely different handshape and it’s in front of his face, more than it is by the side. And that’s fine once you work out ‘Oh, that’s his sign for country.’ But it is that variation that you have to get your head around for the whole monologue. And things like the u-n convention. When he first did that, I stumbled over what that was until I worked it out. And I was fine every time I saw it. But it’s an adjustment that you have to make. Another interesting finding is considerable individual differences among the 20 interpreters. Some interpreters were confused by many unfamiliar signs used by the Deaf professional, while others were less affected. For example, Debbie, an interpreter of similar age to the Deaf presenter, stated, ‘I don’t think the International Sign is a problem.’ In addition, Emily, an interpreter who worked at the same workplace as the Deaf professional and had interpreted for him overseas, reported that the only unfamiliar sign she noticed was access (an ASL sign that is not only iconic but also same as the Auslan sign in), which ‘was easy to comprehend.’ 4.3.1.2 The Deaf Presenter’s Signing Style Eight interpreters reported that they were not used to this Deaf professional’s particular signing style, which encompassed multiple aspects such as sign choice,

84 Quality, Challenges, and Strategies way of producing signs, clarity, pace, syntactical structures, and formal register. Helen commented that interpreters who were familiar with this Deaf presenter would find his particular signing style easy to understand: His signing style is very precise. And if you are used to him, I’m sure it’s incredibly clear. And he is very clear. But he definitely has a quite distinct style. Shannon, a Melbourne-based interpreter who was not very familiar with this Deaf professional, commented that the best practice would be for signed language interpreters who knew the Deaf professional very well to interpret for him in real-life conference settings: I don’t think I’ve actually interpreted for him before. But I’ve seen him in a lot of video clips. … I think he has an idiosyncratic signing style. I  believe that, if he was presenting at conferences or workshops in Australia, he would generally use interpreters who are familiar with him. And that’s something I would certainly advocate, because he has a particular style. … I think I understood most of it [the Auslan monologue], but there were still signs that I wasn’t sure in terms of his signing choices that he uses for particular words. And also, because he has worked ­overseas, I believe that he has a lot of International Sign influence in his signs. … Some of the flow and movement of International Sign influence the way he produces his signs. … I think, in some ways, he is quite Auslan in his signing, but the specific terminology is quite English-based. … He uses sort of English word order in some of his signs, although generally he is still quite Auslan. … I believe that he does have International [Sign] influence in his signing structure. … Location, we need to have regular contact with a Deaf professional before we can feel confident to do an effective job. Similarly, Liz, a Sydney-based interpreter, pointed out that, although she had interpreted for this Deaf presenter before, this time he was using ‘a completely different language structure to the way he would normally be signing to me’ in that he was ‘presenting in a very WFD (World Federation of the Deaf) international format,’ used many rhetorical questions, and signed very slowly. Debbie from Melbourne also commented that she sometimes made conscious intentional omissions (Napier, 2004) when she could not understand the Deaf professional’s Auslan structures: Sometimes I just couldn’t figure out his Auslan structure, [so] I just missed the point. … Because I just didn’t comprehend it, I couldn’t interpret it.

Quality, Challenges, and Strategies 85 It is plausible that the Deaf professional aimed to show formal register at the conference setting by (i) deliberately using fingerspelled words to introduce English terminology (code mixing), (ii) creating redundancy in the signed monologue (e.g., combining signs with iconic natural gestures, using two signs to express the same concept), (iii) signing slowly and clearly, (iv) using relatively small mouth movements, (v) incorporating many rhetorical questions, and (vi) occasionally switching from typical Auslan structures to English-like syntactic structures (code switching) (Johnston & Schembri, 2007; Napier, McKee, & Goswell, 2010; Shaw, 1987). Unfamiliarity with the Deaf professional also made it difficult for some interpreters to understand his discourse structure and prosodic markers. Zoe from Melbourne reported that, despite trying to put herself in his shoes, she could not get into his thought world: I just find him really difficult to interpret. He jumps around between concepts. … For me, he doesn’t flow naturally from one idea to the next. … I get thrown when I work with him, because I never know where he is going. It’s never clear to me where his point is. He seems to emphasise things, but everything seems to be emphasised instead of one thing. … I find it hard to get into his head. Geographic location appears to be a crucial factor that influences signers’ signed language production and impacts on interpreters’ comprehension of signed discourse. More interpreters in Melbourne (73%, 8 out of 11) and Perth (50%, one out of two) who used the southern dialect of Auslan (Johnston & Schembri, 2007) reported comprehension difficulties than interpreters in Sydney (33%, two out of six) who used the northern dialect. Monica (based in ­Melbourne) reported, ‘I’ve never seen him [this Deaf professional] before.’ Ten of the 20 (50%) interpreters had met the Deaf presenter before, but such sporadic contact proved insufficient for them to fully understand his signing on the particular subject matter. While three (27%) interpreters in Melbourne and one (50%) in Perth stated that they had interpreted for this Deaf professional in real life many years before, all six interpreters in Sydney had interpreted for him before, three of whom were working in the same community service organisation as he did and had interpreted for him quite often. Although some interpreters considered the Deaf presenter’s signing style as idiosyncratic, eight interpreters reported high satisfaction with the Deaf presenter’s signing style. For example, Bernie from Melbourne remarked, ‘He is [was] exceptionally clear and he was very well-paced. He delivered [the Auslan presentation] in a way that just, it’s almost a dream job to interpret for him.’ This again highlights the individual differences among the professional interpreters. In summary, unfamiliarity with the Deaf presenter created difficulties for some interpreters in understanding his formal signed monologue. Different interpreters expressed diverse views on the Deaf professional’s sign choice and signing style.

86 Quality, Challenges, and Strategies 4.3.1.3 Unfamiliarity with the Subject Matter While five of the interpreters (Liz, Emily, Molly, Sophia, and Helen) reported that they were familiar with the content of the source text, nine interpreters (e.g., Linda, Jane, Claire, Alex, Annie, Debbie, Lauren, Kay) stated that they were not very familiar with it, and the remaining six interpreters made no comment on their familiarity. Interpreters’ varying degrees of familiarity with the subject matter is largely related to (i) their domain-specific or encyclopedic knowledge, (ii) knowledge about the Deaf professional and his previous work experience, (iii) pre-task preparation, (iv) no briefing with the Deaf presenter, (v) limited exposure to Deaf people talking about this subject area, and (vi) the amount of previous relevant interpreting experience. Alex had seen the Deaf professional previously giving a similar presentation in an authentic situation in Australia. Like most participants, Claire read the preparation materials (contextual information of the interpreting task, PowerPoint slides, and a glossary of fingerspelled words) and had an overall picture, but she could not remember all the details, which hindered her from using anticipation during the interpreting task: When working from Auslan to English, it takes me longer to get the meaning. … I wasn’t always 100% sure of the topic. … I had a really good broad idea that he was going to talk about interpreting services, accessibility, [and] education. I saw these figures about how many Deaf schools, but I obviously didn’t memorise them. I forgot about the different approaches [of Deaf education]. So, I wasn’t predicting very well what was coming. Another finding is that many interpreters reported having prior knowledge about some parts of the source text, as illustrated by Jane’s remarks: I had seen this diagram [which represents the four fundamental human rights for Deaf people] before, but I hadn’t memorised the names of all the regional secretariats [of the World Federation of the Deaf]. … I knew about the articles [of the United Nations Convention on the Rights of Persons with Disabilities], and how many times sign language was mentioned in the Convention. But I didn’t actually know how the project had run or what the questionnaires covered.Yeah, so, not very familiar. I didn’t know any of the statistics about how many countries recognise sign language in their constitution or legislation. That’s all new. Interestingly, Annie stated that her Auslan-to-English interpreting performance would have been better if she was familiar with the Deaf professional’s research project, had done more preparation, and had more experience interpreting this subject area: I’m familiar with things like WFD (World Federation of the Deaf) and Deaf people around the world. But [regarding] all those statistics, I haven’t seen the survey myself. If I had seen the survey myself, that might have helped a

Quality, Challenges, and Strategies 87 bit more, because of the way he was presenting scientific data. … This sort of data isn’t talked about all the time in sign language.You hear it in English a lot and constant[ly] interpret [English speeches on] human rights at conferences. But, not so much with Deaf people. … Interpreters don’t work [from Auslan] into English enough with this sort of concepts. And Deaf people are rarely given the opportunity to speak about their human rights. … I think it’s just more sort of opportunities of practice. … I need more practice and more preparation. Sophia struck a similar chord with her remarks: We don’t get a huge opportunity to do things like this [interpreting from Auslan into English]. There [are] only a number of Deaf professionals [in my city] who give [Auslan] presentations, and they tend to be the same presentations. So, there is not a huge opportunity to practise those skills. So, whenever there is an opportunity, I jump at it. Annie and Sophia’s comments indicate that, since signed language interpreters have more opportunities to practise interpreting from a spoken language (e.g., English) into a signed language (e.g., Auslan), they are familiar with many subjects (e.g., disability discrimination) discussed in spoken language rather than in signed language. Ten interpreters (e.g., Annie, Claire, Molly) reported that more preparation would have improved their interpreting performance. If interpreters had searched for the title of the Deaf presenter’s PowerPoint slides (already provided as preparation materials) on the Internet, they could have found the final report of his research project which served as the basis of his Auslan monologue. Molly explained how she prepared for this interpreting assignment: I try to understand what the point of the whole presentation is, like what’s he doing there, what’s he trying to achieve, who is he talking to, and then I will pick out things I think might be difficult. So, do I understand what this word means or this sign means in this context? But, there is no point trying to memorise numbers. That’s why I want someone [my team interpreter] going, ‘[laughter] Pointing to it.’ Due to unfamiliarity with the Deaf professional and his subject area, some interpreters such as Lauren had to allocate most of their attentional resources to signed language comprehension and feared about using long processing time: Even though I read the preparation material and I knew what he was going to be talking about, I wasn’t too sure as to what he was going to present here. … So, I was concentrating on comprehending his signed discourse more than finding the right words in English. … Generally, I’m quite close [my processing time is quite short]. In this situation, I find that if I get a bit behind, it’s harder for me to catch up, because I’m not familiar with the content.

88 Quality, Challenges, and Strategies However, three interpreters (Tiffany, Monica, and Lauren) pointed out that if it was an authentic conference interpreting assignment, they would have arranged a time to sit down with the Deaf professional to ask him for clarification and rehearse the signed presentation. Monica underscored the importance of rehearsal and long-term familiarity: I would sit with him beforehand and get him to do the whole presentation. Also, I would want to work with him in small settings first before [interpreting for him at] a big conference. At small meetings, I can say ‘Stop, do that sign again.’ So, I would like to work with him over time. Obviously, familiarity is very important, so then I would know what the work it is he does, and what he is likely to talk about, what the signs that he uses. In addition, Tiffany explained how familiarity contributes to trust and high quality of interpretation: I would not take licence and put in my own words with this Deaf professional. I don’t know him well enough. I don’t have his trust. He doesn’t know me at all. … However, when I work with somebody that I know very well, a Deaf person [that] I feel I’ve got a rapport with and I have got their trust, I feel that I can get into their head. … If you get into a Deaf person’s head and they trust you to speak for them, that then gives you freedom to be able to use your words, fill in those gaps, and structure your sentences. In summary, the interpreters reported that their unfamiliarity with the Deaf professional, his particular sign choices, his signing style, and the subject matter caused comprehension difficulties in the Auslan-to-English interpreting task. Poor comprehension then affected the interpreters’ interpreting process and performance. 4.3.2 Cognitive Overload As many as 15 participants reported that they experienced cognitive overload when simultaneously interpreting the Auslan monologue into spoken English. As Table 4.4 shows, the primary causes of cognitive overload identified include: (i) interpreting numbers, (ii) long processing time, (iii) syntactical differences between Auslan and English, and (iv) dense information in the source text. These factors are often intertwined. Amber stated that a combination of long processing time, numbers, and dense information caused heavy cognitive pressure: Time lag [processing time]. I think, with this particular Deaf professional, you need to have much more time lag [processing time] than in most cases. … Numbers and the density of information. There are just lots of statistics. Ninety-three responses. Of the 93, 83 were this. Of the 83, this many were

Quality, Challenges, and Strategies 89 this. So, the information is very detailed, very dense. So, trying to put that information into understandable English was difficult. … I felt like I needed to make sure I was remembering particularly the numbers and what they related to. But at the same time, I didn’t want to jump ahead of myself, because he would probably go somewhere else unattended. So, just trying to wait to get the full concept of what he was trying to say, but at the same time try not to forget the numbers or get caught up in the next part too soon. By the same token, Jane commented on high cognitive load, which resulted from a mixture of numbers, dense and complex information expressed concisely by the Deaf professional, and signed concepts that required elaboration or explanation in the target language to meet the needs of the non-signing hearing audience: He is such a clear signer, and such a visual user of Auslan. His Auslan is quite dense. The way he gets across concept is very economical. His facial expression and spatial placement are very important. … And some of the signs he uses are very meaningful, but it’s quite hard to put it [them] into English. … I think the ‘total communication’ thing [concept] was tricky, because describing Deaf education in Auslan is lovely and economical, and very easy to describe. [Describing] Deaf education in English for hearing audience[s] who don’t necessarily understand what that is, is quite wordy. So, I had to describe in English what he was doing very economically in Auslan, and then straightaway deal with a whole lot of numbers. It was, this number of countries had this type of education plus others, this number had this type of education plus others, and then there is a whole new concept of whether or not these types of education happened in dedicated Deaf schools. So, it’s quite conceptually complex, and there were numbers in the middle as well. Syntactical differences between Auslan and English required interpreters to use long processing time. Helen described how she coped with the Deaf presenter’s rhetorical questions: Those rhetorical questions, if you watch him carefully, as he is signing it, he starts to raise his eyebrows, so you know that it’s not a statement. But you have to watch for that, because otherwise you end up saying ‘Oh, blah blah blah. No, they are not!’ Laugh. And it sounds ridiculous. I don’t think I got tripped up by any of those, but there were a couple of them. Sometimes my lag time [processing time] is a little long, so I see those things [i.e., cues of rhetorical questions] before I have to say them. I know that they are coming, and so, I don’t have to repair an incorrectly interpreted statement like that. Similarly, Sophia perceived that her processing time when working from Auslan into English was longer than her processing time in the opposite direction

90 Quality, Challenges, and Strategies due to syntactical asymmetry between Auslan and English and difficulty in finding the appropriate English words, and that her excessively long processing time sometimes led to cognitive overload: I definitely have a longer time lag [processing time] with Auslan-to-English interpreting than with English-to-Auslan interpreting. … Quite often Auslan is sort of back to front, so it’s better to wait till the whole thing (Auslan meaning unit or sentence) comes out and then say something coherent [in English] rather than start talking and change mind halfway through. … There were a couple of times where I just had trouble getting words out. … I’m still speaking three minutes later and he has moved on. So, there were a couple of moments when I thought, ‘Oh, there is definitely a bit of cognitive overload happening.’ Interestingly, Annie vividly recounted her cognitive process of dealing with numbers in adjacent Auslan sentences: Yeah, numbers, those numbers. I have memory overload. I was holding on, ‘Ok, quick, get the idea across, and see how many.’ … Get out [Express the information in spoken English] as quickly as possible. Laugh. So, I will go, ‘Okay 34’ and then he [the Deaf presenter] tells me what it is for, and then I have to think ‘34 respondents’ or ‘34 countries’ to keep the ideas connected. The number, the concept, what, and who. … [I was telling myself,] ‘Don’t forget the number, don’t forget the number.’ And he threw another number, so I had to let go of the 34 and hold on to the other number and throw it. It’s like catching and throwing a ball. You catch it, you hold it, quickly throw it again while you are catching another one. In stark contrast, five interpreters stated that numbers did not present a challenge because the Deaf professional’s Auslan presentation was well-paced and he sometimes repeated his numbers. In terms of memory strategies, some interpreters focused on memorising the meaning of concepts rather than individual signs, remembered the English equivalents of signs, or used visual memory (handshape of signs, reference points in space). Sophia recalled using generalisation and remembering the Deaf presenter’s use of space to deal with numbers: Any team interpreter working with me knows to write down numbers, because I am not very good with remembering numbers. I have developed strategies to cope with numbers. Even though I miss the number, I usually get a rough idea of it, so I say,‘many or about 50 respondents said this.’ Also, if that was an actual Deaf professional presenting, the PowerPoint would be here, so I would have the visual prompts. … When he was talking about respondents from World Federation of the Deaf (WFD) who were positioned on this side, and respondents from countries that weren’t from WFD, I definitely ­remembered who was positioned where.

Quality, Challenges, and Strategies 91 In summary, a suite of factors caused cognitive overload in signed-to-spoken language simultaneous interpreting, including numbers, dense information, syntactical differences between a signed language and a spoken language, and long processing time.The interpreters reported that the use of strategies (such as generalisation and using visual memory), a team interpreter, and having visible PowerPoint slides could have reduced their cognitive load during simultaneous interpreting. 4.3.3 Target Speech Production Difficulties In this mock conference setting, interpreters were expected to produce an accurate, coherent, and formal English speech fluently for a non-signing hearing audience. Participants reported the following difficulties in target speech production: (i) trying to produce comprehensible English sentences, (ii) using an appropriate register in English that would accurately represent the Deaf professional in this formal situation, and (iii) maintaining coherence in the target speech. 4.3.3.1 Trying to Produce Comprehensible English Sentences Six interpreters reported that they felt it was challenging to produce meaningful, grammatical, and idiomatic English sentences while managing other concurrent sub-tasks in Auslan-to-English simultaneous interpreting. Three participants stated that the Deaf professional’s use of code mixing and code switching impacted their lexical and syntactical choices in the target language, as illustrated by Helen’s remarks: When I interpret for a Deaf person who is so bilingual and is using so many English references such as fingerspelled words, signs that have been created around particular English concepts that [are] related to this convention [UNCRPD], and English word order, as an interpreter, sometimes I have less room to move and I feel a bit more constrained in what my word options are in English, because I feel I have to stay closer to that.Whereas if he has a more ­Auslan-like delivery, I have more room to move in English, and I have more confidence a lot of the time in English. If it comes to me more Auslan-like, I feel I can do a better job of this. Shannon, who grew up in a signing Deaf family, reported that it was cognitively challenging to constantly switch between form-based and meaning-based information processing, transfer meaning between modalities, and create an equivalent communicative effect: I need to be confident to follow his English that he was [is] producing in his signs, but also remember that occasionally I still need to work the

92 Quality, Challenges, and Strategies grammar around so that it still [makes] sense in English. … I’ve understood his signs in a visual [image] of how he signs it, and [I have] worked out the meaning, and I’m simultaneously thinking about, ‘Ok, how do I restructure it so that it makes sense in spoken English?’ … I need to really think about putting it [the meaning of signed discourse] into coherent grammatical structures in English.’ … And [I am thinking about] when I’ve heard people present information on the UN and the Convention [UNCRPD], how they speak and produce coherent [English] sentences that I guess is the ‘norm’ in terms of how they produce the jargon when they speak, and [I  am] trying to create that same effect when I am watching him. … Mentally, I do find interpreting from Auslan into English more tiring than the other direction. The Australian Deaf presenter is a multilingual of Auslan, written English, International Sign, and other sign languages, due to his life experiences, professional roles in international organisations regarding Deafness and disabilities, and many years of working overseas. Helen’s and Shannon’s remarks suggest that they had to flexibly switch between a literal interpretation approach and a free interpretation approach in order to cope with the Deaf presenter’s code mixing and code switching in the formal signed monologue. In addition, the Deaf presenter’s use of fingerspelling and signs that represent specialised English terminology suggests his preference for the interpreters to incorporate those precise English words in the target language renditions. However, due to syntactical differences between Auslan and English, those specific English words may not have fit in the corresponding English interpretations to create a natural speech output, or that the interpreters increased their cognitive efforts to search for particular English words, phrases, clauses, or syntactic structures that could work in the target language renditions. As per Gile’s (2009) Effort Model of simultaneous interpreting, such intentional efforts require extra processing capacity for target language production. 4.3.3.2 Using a Formal Register in English Ten participants stated that matching – or maintaining – the Deaf professional’s formal register in this conference setting was a challenge, as illustrated by Lauren’s comments: ‘I didn’t get the right register, so that was really annoying me. I had it, and then I lost it.’ Bernie consciously refrained from retaining the Deaf presenter’s use of rhetorical questions in his English interpretation and also struggled with speaking in a higher register: ‘How many countries responded in this way? Oh, there were 32.’That [rhetorical question] is a conversation that two people might have. That’s not what he [the Deaf presenter] was doing. I didn’t want to do that, because that’s not the register that he was delivering in Auslan.The register required

Quality, Challenges, and Strategies 93 a completely different delivery. … I found it hard to match the register, and I found it hard to deliver it in a way that I was happy with. Debbie also expressed dissatisfaction with her spoken English interpretation and stated that having a team interpreter would have improved her register: I wouldn’t have been happy with that [my English interpretation] if I had been interpreting for an audience, because he [the Deaf presenter] is such a professional. He would want a really smooth, just unblemished English interpretation. … I would have wanted a tandem [team interpreter] prompting me to seek a better register of English in some cases. Annie imagined herself in an academic context to help her select appropriate English words for the Deaf professional’s particular signs: I was struggling with trying to find the right register, and trying to get across concepts to an audience who don’t sign. … In this interpreting task, I was setting up Auslan ideas in English and trying to have the correct discourse model that an academic would use. … ‘Graphs,’ ‘statistics,’ ‘results,’ and ‘respondents’ – these are really typical words that I was just thinking as I was interpreting. Helen, an interpreter who was a non-native signer with 21 years of interpreting experience and a PhD, stated that generally she felt more confident working from Auslan into English (from L2 into L1) because of effective self-monitoring and native language production: I feel like I have greater facility in English, because it’s my native language. I feel that I can monitor myself more accurately, because I can hear what it is that I’m saying. … I have a mental filing cabinet of more options for words, because it’s my native language. So, I just feel more confident working from Auslan into English. … I have an academic English register. … Working from Auslan into English, when I knew what it was [I understood the Auslan messages], it [my English interpretation] came out fluently and smoothly, and had good prosody, and was at the right register. And I made some really good word choices at different times during this interpreting task. … I think, when working from Auslan into English, I have better compensation strategies. So, if I’m not sure of something, I still have ways to work around it in English. However, when I’m working from English into Auslan, I have fewer compensation strategies. Helen’s comments highlight the advantages for non-native signer interpreters to work from their non-native language (signed language) into their native language (spoken language) as long as they understand the signed source text

94 Quality, Challenges, and Strategies thoroughly. Interpreters’ native language production increases the chance of ­representing a Deaf professional appropriately in terms of using a formal register. 4.3.3.3 Maintaining Coherence in the Target Speech Seven interpreters commented that maintaining coherence in the target speech was a challenge. For instance, Debbie occasionally devoted most of her attention to source message comprehension and did not have enough attention left to produce a coherent English speech: Sometimes I struggled to make coherent sentences in English, because I was struggling to figure out what he was trying to say. So, I kept rephrasing things all the time. … I waited a little bit longer to see if what he’s going to say next would refer back to the previous point that I had missed, which he did a lot of times. In addition, Amber, an interpreter who was a native user of Auslan, attributed the difficulty in maintaining coherence in her English interpretation to discourse-level differences between Auslan and English and to directionality: He [the Deaf presenter] put the end first and then the beginning. It’s very different from English. For example, rather than saying ‘The way we did the survey was,’ he says, ‘We did a survey,’ and he talks about the survey, and then he goes back, ‘How did we do it?’ That [the question at the end] should have come first. I was just trying to make coherent, complete sentences and concepts. I found it really difficult. … Working into the English direction, I feel like my processing speed is not as quick. I can’t fix mistakes easily. And if the Deaf person goes into another direction, I need to practise the strategies of linking that [the previous] concept to the new concept rather than just stopping that and moving into something new. In English I have to really think about it [target language production], whereas in Auslan I can do that much quicker and easier [more quickly and easily]. Jane devoted much attention to re-ordering information and considering the hearing audience in order to ensure coherence in her target language speech: I was spending a lot of effort on making sure that my English [speech] is coherent. I need to be responding to his [the Deaf presenter’s] idea of coherence in Auslan first, and then take that into coherent English. … It’s an extra effort to connect what you are seeing and how you’ve got to translate that with what you’ve just said in English.That created a bit of memory overload. … I think the audience is going to get confused if they are not feeling the links between things. … I do a lot of re-ordering of information working into English, to make it coherent at that sort of sentence and discourse level to make it sound like a presentation in English. … There were

Quality, Challenges, and Strategies 95 structures that he had in there that helped me to achieve coherence. Like, when he was talking about the regions, he did go with this region, this, and then next, next, next, and finally. So, he gave me everything I needed to make that bit coherent. He was beautifully coherent in Auslan. I use phrases such as ‘Moving onto the next thing’ and ‘And the next thing we looked at was,’ so that it felt like there was a link between the different concepts. In summary, the interpreters reported that maintaining coherence in the target speech was related to source language comprehension, interpreters’ knowledge about discourse-level differences between a signed language and a spoken language, and their self-monitoring of the target speech. 4.3.4 Artificial Testing Conditions As shown in Table 4.4, some interpreters commented that the working conditions under this research setting were inferior to those at actual conferences, and this artificiality affected their interpreting process and performance. Eight interpreters stated that the use of a two-dimensional video rather than a live presentation as the source language stimuli deprived them of an opportunity to meet the Deaf professional before the Auslan-to-English simultaneous interpreting task for briefing, clarification, rehearsal of the entire presentation, and building a good rapport with him. The briefing could have enabled them to familiarise with him, his sign choices, signing style, signing speed, and content, to develop rapport or trust, and to establish cooperation strategies. Moreover, the use of the Auslan video made it impossible for the interpreters to apply the look-pausenod cooperation strategies (Napier, Carmichael, & Wiltshire, 2008) to work with the Deaf professional during the simultaneous interpreting task. Many of these issues are captured in Alex’s comments: It was an artificial situation, that was a difficulty as well. Because this is a two-dimensional video, I had some difficulties sometimes looking at this screen and seeing his hand-shapes for numbers. I was unsure sometimes if it was three or two. So, some of the finer details are lost. And also, because I was not live, I [didn’t] have the opportunity to work him and develop that rapport so that he knows where I’m at in my interpreting. So, if he is too far ahead, I’d like to give him a signal to say, ‘Could you please just slow a little down?’, without informing the rest of the audience. That assists me with my time lag [processing time], my processing, and my interpretation. That didn’t happen, so that was a difficulty which was a small stress. The fact that no PowerPoint slides were visible behind or next to the Deaf presenter made it difficult for nine interpreters to follow his train of thoughts (change of slides), understand his particular signs, decipher his visual representation of the diagram made up of four overlapping circles (sign language, bilingual education, signed language interpreting training, and accessibility), check

96 Quality, Challenges, and Strategies English terminology, and retain numbers or details in his signed presentation. In addition, Liz used the printed PowerPoint slides to check a few numbers and unfamiliar signs that she had missed, whereas Helen revealed that looking at the slides distracted her from watching the Auslan video and led to her missing more information. Furthermore, eight participants commented that a team interpreter could have ameliorated their signed language comprehension problems by pointing to specifics on the printed PowerPoint slides, providing numbers or target language words when needed, and enhancing their confidence. Finally, the absence of an audience affected some interpreters’ self-motivation. In view of the aforementioned challenges, it is not surprising that six interpreters stated that they felt stressed during the Auslan-to-English simultaneous interpreting task and felt tired during or at the end of the task. In summary, the main challenges in signed-to-spoken language simultaneous interpreting reported by professional Auslan/English interpreters in this experimental study include: signed language comprehension problems, cognitive overload, target speech production difficulties, and the artificial testing environment.

4.4 Self-reported Strategies This section presents interpreting strategies that participants reported that they had adopted during the Auslan-to-English simultaneous interpreting task.These strategies include: (i) waiting, (ii) stalling, (iii) making strategic additions, (iv) making strategic omissions, (v) paraphrasing or rephrasing, (vi) summarising, (vii) using compression, (viii) generalising, (ix) switching between a literal interpretation style and a free interpretation style, and (x) using long-term memory. Some of these strategies such as waiting, making strategic additions, and generalising have been mentioned in the previous section and thus will be discussed only in a cursory manner here. Representative quotes from participants are integrated to illustrate the interpreting strategies. Although some strategies are primarily associated with source speech comprehension (e.g., waiting) or target speech production (e.g., switching between a literal interpretation style and a free interpretation style), others are linked to both (e.g., generalising). Firstly, it is necessary to wait long enough so as to understand Auslan structures correctly and render them into grammatical English sentences in a formal register, as evidenced by Claire’s comments: I was trying to keep it smooth and not wait too long. But when I waited, I actually got better information. … In sign language, he [the Deaf professional] usually puts the positive [at the beginning and the negation at the end] and says, ‘Do Deaf people have equal rights to everyday living and access? No, they don’t.’ When I interpret it, I don’t really want to say it like that, because it doesn’t sound appropriate in the target language. So, I try and wait, and then I hope that I can fit it all in and won’t miss the next bit.

Quality, Challenges, and Strategies 97 However, waiting can be a double-edged sword. Long processing time sometimes led to cognitive overload and conscious intentional omissions, as commented by Molly: ‘There [were] a couple of things where conceptually I just didn’t quite get it, and so I was waiting, and then by the time I got it, I missed the next thing.’ Hence, it is important for interpreters to adjust their processing time appropriately, as explained by Bernie: One strategy was to use a long enough time lag [processing time] to avoid being sucked into using rhetorical questions just to cope. … Another strategy was to shorten the time lag [processing time], because sometimes I find that if I’m really close with my time lag [processing time], particularly with someone who is delivering a lot of information like this [the Auslan monologue], I can catch up, I can keep there. But as soon as I let the time lag [processing time] lengthen a little, I started to use techniques of paraphrasing, and there was too much detail and it was too dense for me to paraphrase. So, there are [were] times when I was a bit behind, and then I managed to catch up and then I stayed with it, and then I fell behind a little bit and I caught up again. So, my time lag [processing time] was short and long, short and long, short and long, all the way through. Secondly, stalling was used as a strategy to buy time and receive more information, as explained by Monica, ‘I use longer lag time [processing time] for [interpreting from] Auslan to English. I buy time, so I put in clichés or just standard kind of wording to buy time to cover up the lag.’ Thirdly, making strategic additions includes elaborating Deafness-related concepts for the non-signing audience (see also Jane’s earlier comments on explaining ‘Deaf education’), adding contextual information, and explicating the Deaf presenter’s intended meaning, among others. Unpacking the Deaf presenter’s concepts that are unfamiliar to the hearing audience is a strategy to maximise audience design and enhance communicative effect, as illustrated by Annie’s comments: When he was talking about Deaf community, oralism, and what sign language is, I had to add information to help listeners, like, guide listeners on a message, because it’s quite abstract talking about Deaf community with words, especially if people have never met a Deaf person, it’s very surreal. Familiarity with the Deaf professional contributes to making strategic additions, as evidenced by Amber’s remarks: He talks in a very broad manner and says ‘access to government.’ … So, I would say ‘access to government information and services.’ I added that in because I know him and I know that’s what he means.

98 Quality, Challenges, and Strategies Fourthly, the opposite of making strategic additions is making conscious strategic omissions, which was adopted by interpreters such as Monica: There were a few redundancies, so I just chopped them out. Yeah. Laugh. Save, save myself [my attention] for the really hard bits. Fifthly, rephrasing was employed when some interpreters figured out a previously missed sign or concept by using the context.This is illustrated by Debbie’s comments, ‘If I didn’t get it [understand a signed message], I sometimes worked it out later and then had to rephrase it into a sentence differently.’ Furthermore, when interpreters lagged quite far behind the Deaf presenter, they employed summarising, used compression (condensing), or spoke faster in order to catch up. For example, Amber used summarising to bring concepts together quickly: Because I’m a very short time lag [processing time] person, when I get really behind, I would try and construct one sentence that is going to encompass everything that he has just said. So, I’m in front again. In addition, Linda used summarising and compression to gloss over a source message that she had missed and then carried on: There was a little bit of condensing. There was one point that I had overlooked, and then I tried to give a very brief summary of what that meant and continued on. Both Emily and Sophia reported that they used generalisation to deal with a missed number. Emily commented: ‘I tried to catch all the numbers. I think I missed one. So I just said “a number of ” instead of the number.’ Moreover, interpreters reported switching between a literal interpretation style and a free interpretation style as an interpreting strategy. As mentioned in Section 4.3.3.1, interpreters stated that they had to restructure Auslan sentences in order to produce comprehensible English sentences, which constituted free interpretation. In addition, Lauren perceived that his predominantly literal interpretation style led to an improper representation of the Deaf professional: I tried to be as literal as possible. I knew I got stuck in certain areas. … My main strategy was just to stop the negative self-talk. I just need to get the information across, and it may not be register-correct or word-correct. As long as I get the information out there, that’s the best I can do. And that’s sort of what got me through it. But that was making me feel guilty at the same time. So, I thought, I’m not properly representing him. Finally, a few interpreters reported that they tried to recall details from the PowerPoint slides that they had read for preparation and to retrieve relevant

Quality, Challenges, and Strategies 99 information from their long-term memory to perform as best as they could on the Auslan-to-English simultaneous interpreting task. Alex commented, ‘I tried to recall the presentation, particular data, statistics, and terminology. Also, I used previous learning and knowledge to supplement my interpretation.’ In summary, interpreters reported adopting a wide range of strategies such as adjusting processing time, making strategic additions and strategic omissions, rephrasing, summarising, and condensing to cope with a constellation of challenges in simultaneous interpreting from Auslan into English.

4.5 Discussion This section discusses the aforementioned quantitative and qualitative results in relation to the pertaining literature. 4.5.1 Native Signers and Non-native Signers This study found that the 10 native signers were similar to the 10 non-native signers in terms of the total score and each of the four sub-scores of Auslan-toEnglish simultaneous interpreting performance. Overall, both groups performed quite well on this interpreting task, achieving an average score of approximately 80 (out of 100). These results corroborate my previous findings from a comparison of a larger sample – 14 native signers and 17 non-native signers – in terms of performance on the same interpreting task (Wang & Napier, 2015). These results are also consistent with van Dijk, Boers, Christoffels, and Hermans’s (2011) findings, showing no significant differences between two groups of experienced SLN (Sign Language of the Netherlands)/Dutch interpreters – 10 native signers and 15 non-native signers – in terms of SLN-to-Dutch simultaneous interpreting performance. These findings indicate that age of acquiring signed language did not have a significant impact on professional interpreters’ signed-to-spoken language simultaneous interpreting performance.These results may be attributable to the following factors. First, it appears that the native signers and the non-native signers were almost equally affected by signed language comprehension problems in this signed-tospoken language simultaneous interpreting task (see Section 4.3.1). The key difference between these two groups is early acquisition of signed language from signing Deaf parents, which may influence an interpreter’s signed language comprehension (receptive) and production (productive) skills as well as level of involvement in the Deaf community, but not always. Previous research has shown some marked differences between interpreters who are native signers and interpreters who are non-native signers in spoken-to-signed language simultaneous interpreting performance, particularly in signed language production aspects such as target text features (e.g., grammaticality, vocabulary, conveying register and affect) and delivery features (e.g., clarity of articulation, fluency, prosody) (Wang & Napier, 2015) and in the use of expansions as interpreting strategies (Lawrence, 2003). It is possible that native signers have greater

100 Quality, Challenges, and Strategies advantages over non-native signers in the signed language production stage of spoken-to-signed language interpreting than in the signed language comprehension stage of signed-to-spoken language interpreting, because interpreters have more control over their own signed (target) language production than the signed (source) language input produced by other (Deaf or hearing) signers. In terms of directionality, when working from Auslan into English, the native signers were interpreting from one native language into another native language (A1 to A2), whereas the non-native signers were interpreting from their nonnative language into their native language (B to A). Given that the interpreters who are native signers still interact with the general public predominantly by using a spoken and/or written language and generally have considerably more opportunity to listen to and use spoken language across all domains, settings, and registers than signed language, they are possibly more proficient and resourceful in their speaking skills than (receptive and productive) signing skills, though both are their native languages. In other words, interpreters who are native signers possibly have spoken language as their dominant (relatively stronger) native language and signed language as their non-dominant (relatively weaker) native language, when they use the languages to discuss some topics in some contexts. However, they may have signed language as their dominant native language and spoken language as their non-dominant native language, when they use the languages to discuss other subject matters (e.g., Deafness-related topics) in other contexts.These assumptions warrant further investigation. However, unfamiliarity with the Deaf presenter, his specific signs, and his signing style may have ­mitigated the native signers’ signed language proficiency advantage over the non-native signers. Given that source language comprehension is the pre-­requisite of accurate interpretation (Cokely, 1992; Taylor, 2002), both groups’ signed language comprehension problems probably had an equally negative impact on their signed-to-spoken language simultaneous interpreting performance. Second, even though the native signers may understand the signed language monologue better than the non-native signers, the former may not be better at expressing the understood signed messages appropriately in spoken language; thus, each group’s strengths might have been offset by their weaknesses. As shown in Table 3.1 in Chapter 3, in this study more non-native signers held a university degree and completed formal (postgraduate-level) interpreting training than native signers, indicating that the former is probably more capable of using a formal register in English and employing interpreting skills than the latter. Bontempo, Napier, Hayes and Brashear (2014) found that, among signed language interpreters globally, native signers rated themselves more highly than non-native signers on signed language proficiency, but the non-native signers rated themselves more highly than the native signers on spoken language proficiency. In other words, when it comes to signed language comprehension and spoken language production in signed-to-spoken language simultaneous interpreting, the two groups may perform similarly on the whole. Third, there are considerable individual differences within each group (either native signers or non-native signers) in terms of numerous aspects such as signed

Quality, Challenges, and Strategies 101 language proficiency, spoken language proficiency, general education level (see Table 3.1 in Chapter 3), domain-specific knowledge, world knowledge, intelligence, formal interpreting training, interpreting experience in general, experience of working from a signed language into a spoken language in formal settings in particular, working full-time or part-time as a signed language interpreter, and preparation for this interpreting task. As a consequence, it is almost impossible to generalise the common features and competencies of either the native signers or the non-native signers. Thus, the mean value of each group’s signed-to-spoken language interpreting performance disguises these individual differences and shows only a summary statistic. In addition, signed-to-spoken language simultaneous interpreting performance in a formal setting may be more relevant to other factors such as interpreters’ general educational background, formal interpreting training, English vocabulary size, interest in lifelong learning, years of interpreting experience, and experience of working in this language direction. Note that the two groups in this study were similar in terms of average years of interpreting experience (16 years, with a standard deviation of 8) and in terms of the number of interpreters with a minimum of 10 years of interpreting experience (8 native signers and 7 non-native signers; see Table 3.1). It is plausible that extensive interpreting experience has obscured the possible differences between the native signers and the non-native signers. These assumptions merit further research. Interpreters with similar backgrounds regarding the aforementioned aspects need to be recruited in future studies on directionality and interpreting. Fourth, among signed language interpreters, both native signers and nonnative signers received more formal training and had more real-life experience of interpreting from a spoken language (e.g., English) into a signed language (e.g., Auslan, ASL) than vice versa (Kalata-Zawłocka, 2016; Napier et al., 2010; Napier, Rohan, & Slatyer, 2005; Nicodemus & Emmorey, 2013; Nilsson, 2016; van Dijk et al., 2011). Deaf people are still given less opportunity to express their views than to receive information from hearing people via signed language interpreters (Napier et al., 2010). Moreover, both formal interpreting training and real-world signed language interpreting work primarily involve dialogic interpreting rather than monologic interpreting. The disproportion of interpreting directions in both formal training and real-life interpreting work appears to have resulted in many signed language interpreters preferring and feeling more confident about working from a spoken language into a signed language than vice versa (Napier et al., 2005; Nicodemus & Emmorey, 2013, 2015; Wang & Napier, 2015). Taken together, both the native signers and the non-native signers have limited training, opportunity, and experience in interpreting signed monologues into spoken language in formal settings; therefore, they are likely to be on a par in signed-tospoken language simultaneous interpreting performance. Fifth, both meaning-based information processing and effective self-­monitoring in signed-to-spoken language simultaneous interpreting may have contributed to the native signers and the non-native signers having achieved similar performance. As shown by the retrospective interview data, nearly all interpreters in this study were aware of the need to deal with syntactical differences between a signed language and

102 Quality, Challenges, and Strategies a spoken language, produce spoken language sentences that would make sense to the non-signing audience, and consciously monitor their target speech for a formal register, coherence, fluency, and prosody. Nicodemus and Emmorey (2013, 2015) pointed out that while signed language interpreters may rely on fingerspelling and transcoding (producing signs with spoken language syntactic structure) as default strategies when working from a spoken language into a signed language, such strategies are not as readily available when working from a signed language into a spoken language, as transcoding in the latter direction often leads to ungrammatical constructions in spoken language. The fact that both the native signers and the nonnative signers could hear their own voice in signed-to-spoken language interpreting probably helped them to actively monitor their performance quality. Additional factors such as feeling under a lot of pressure when simultaneously interpreting for a high-profile Deaf professional and being filmed for a research study may have affected both the native signers’ and non-native signers’ performance to some extent. Given that native signer interpreters probably can attain a better understanding of the signed messages produced in their native language, and well-educated nonnative signer interpreters probably are better at using a formal register and selecting appropriate vocabulary to express understood messages in their native spoken language, pairing up a native signer interpreter (for source language comprehension) and a non-native signer interpreter (for target language production) in signed-tospoken language simultaneous interpreting assignments in formal settings in real life is likely to provide highly accurate and effective interpretations. 4.5.2 Improving Signed Language Comprehension The retrospective interview data provided insights into interpreters’ perspectives on challenges, coping strategies, and cognitive processing during signed-to-spoken language simultaneous interpreting. A key finding is that the vast majority of the professional Auslan/English interpreters – seven who acquired Auslan from signing Deaf parents and nine who learned it as adolescents or young adults – reported experiencing varied degrees of difficulties in understanding the Deaf professional’s signed monologue, which hindered them from producing highly accurate spoken interpretations.This finding is consistent with previous research which has found that signed language interpreters experience great difficulties with signed language comprehension (e.g., fingerspelling, sign variation) (Napier et al., 2010; Nicodemus & Emmorey, 2015; Patrie, 1992, 2009; Taylor, 2002; van Dijk et al., 2011; Wang & Napier, 2015). It is highly likely that the interpreters’ signed language comprehension problems have resulted from sign language being markedly less standardised than spoken language, considerable sign variation within the Deaf community, a lack of personal contact with the particular Deaf professional, and little exposure to formal signed monologues, each of which is discussed briefly here. Although English does change gradually across time and has contact with other spoken languages, it is quite standardised in terms of phonology, vocabulary, grammar, pragmatics, and usage across communication situations. Compared

Quality, Challenges, and Strategies 103 with spoken languages, signed languages are much less standardised in terms of vocabulary, syntax, register, and other aspects. Deaf signers’ production of a sign can be influenced by internal linguistic factors such as the signs before or after it in a sentence and by register (Johnston & Schembri, 2007). There is also considerable variation in Deaf signers’ use of a signed language (e.g., Auslan, ASL) in relation to external social factors such as age, gender, region, religion, education background, social networks, socio-economic factors, age of signed language acquisition, and individual usage (Johnston & Schembri, 2007; Lucas & Valli, 1992; Napier et al., 2010). In their comments, the interpreters in this study frequently cited the ‘idiosyncratic language’ produced by the Deaf presenter.6 Especially for readers unfamiliar with the Deaf community, it is important to note the social, educational, and linguistic factors that underlie this comment. In most parts of the world, signed languages are not yet formally recognised as legitimate human languages. This lack of understanding of, awareness about, and recognition for signed languages has led to the devaluation of Deaf people in a number of ways. Critically, historically, Deaf people have rarely been given the opportunity to be informed about, much less study, their natural visual language. Rather, they have been told that they must learn the spoken language of ‘the larger world’ in which they reside and, further, that their use of signed language should be suppressed. Deaf people have been placed in educational systems designed to teach them spoken languages through a variety of devices and approaches (sound amplification systems, body aids, Cued Speech, artificially invented sign systems, cochlear implants, simultaneous communication, total communication, speech therapy, Rochester method, and oral education, among others). Some of these devices and approaches may serve their intended purposes, but too many others delay natural language acquisition in Deaf children. In this situation, many Deaf people grew up piecing together their means of communication, which resulted in a wide variety of language use and varying levels of signed language proficiency among Deaf people. Today, there are numerous efforts to investigate and document the linguistic structures of signed languages and further efforts to install signed language instruction in schools for Deaf students. However, these initiatives are slow, and to date, there has not been widespread acceptance about the validity of signed languages. As a result, the pattern in which Deaf children do not learn a standard form of a signed language continues. Thus, it must be acknowledged that Deaf people worldwide having ‘idiosyncratic’ sign language production is the result of many social forces which are typically orchestrated by hearing people. As a result, signed language interpreters work with Deaf individuals who bring a variety of signing styles. Further, the majority of signed language interpreters learn a signed language as a non-native language and, as such, do not bring language ‘instincts’ and experiences with a wide variety of Deaf consumers to their work.While it may be true that Deaf signers’ signed language production can be idiosyncratic, it is also true that signed language interpreter education is often lacking in providing diverse linguistic models (i.e., a broad range of Deaf

104 Quality, Challenges, and Strategies signers) to interpreting students. Often, individuals are awarded signed language interpreter certifications and degrees prior to being fully fluent in a signed language, especially one that has such a wide variety of users. The huge demand for signed language interpreting services has far outstripped the supply, which often results in signed language interpreters not having true bilingual skills, not having adequate preparation to manage the signed language diversity they face in their work. Furthermore, the great availability of work for signed language interpreters may result in a situation in which practitioners become fossilised in their interpreting skills and, further, unable to evaluate their proficiency in signed language comprehension and production. Again, these situations in signed language interpreter education are the result of large social forces. This discussion is not intended to implicate either Deaf people or signed language interpreters but rather to recognise the complex social conditions, which impact the lives and work of Deaf individuals and signed language interpreters. As may be seen in the linguistic samples, this inequity in fact was the topic of the Deaf presenter’s lecture (see Appendix B for a full transcript of his Auslan presentation and an English translation). All these social, historical, linguistic, and educational factors have contributed to signed language being a non-standardised language, Deaf individuals signing differently from one another, and signed language interpreters stating that it is difficult to understand signed language input. With the closure of many Deaf schools and Deaf clubs, new technologies such as the online dictionary Auslan Signbank, FaceTime, YouTube, and online videos of Deaf native signers using signed languages may hold great potential for signed language transmission in the twenty-first century. The retrospective interview data also revealed that the Sydney-based Deaf native signer used some signs from the northern dialect that were unfamiliar to interpreters who lived in Melbourne, Perth, and Adelaide and used the southern dialect. In his monologue, the Deaf professional also incorporated some old signs that were specific to his generation, several signs that were influenced by preceding and subsequent signs in a sentence, and some foreign signs due to his prior overseas work experience, all of which posed as comprehension challenges to some interpreters. These findings indicate that the considerable sociolinguistic variation in sign language due in large part to oppressive linguistic constraints on Deaf people makes it difficult for signed language interpreters to understand signed language source texts. While Auslan has core vocabulary, grammatical structures (topic-comment statements, constructed action, constructed dialogue, rhetorical questions, etc.), and unique non-manual features that are independent from English, both the vocabulary and grammar of Auslan have been influenced by spoken and written English (Johnston & Schembri, 2007; Napier et al., 2010). Deaf signers also use depicting signs in their signing to manually and vividly describe objects and/or their movement. As a result, Deaf signers of the same sign language differ considerably from one another in terms of vocabulary size, fingerspelling, syntactical structure, communication behaviours, discourse features, and signing styles

Quality, Challenges, and Strategies 105 (Napier et al., 2010). In addition, sometimes Deaf signers use both code mixing (integrating spoken language words and phrases into signing) and code switching (changing from a more ‘deaf ’ or sign language-structured signing style to a more spoken language-structured signing style) to show a higher register of signed language in formal settings such as signed lectures and conference presentations, or when referring to texts (Napier et al., 2010). Some interpreters in the current study (e.g., Helen, Shannon) reported a lack of flexibility and versatility in coping with the Deaf professional’s use of code mixing and code switching in his signed presentation. Signed language interpreters’ limited exposure to formal register in signed language (due to the relatively small number of Deaf professionals and Deaf leaders compared with their hearing counterparts, and Deaf people having less opportunity to express themselves in signed language than be informed by hearing people via signed language interpreters), and their fewer opportunities to practise interpreting from a signed language into a spoken language in formal settings than vice versa, probably have affected their signed-tospoken language interpreting performance and created some internal fear about working in this language direction (see also Napier et al., 2010; Nicodemus & Emmorey, 2013, 2015;Taylor, 2002; van Dijk et al., 2011;Wang & Napier, 2015). There has also been contact among Auslan, other signed languages (ASL, BSL, NZSL, etc.), and International Sign (Johnston & Schembri, 2007). The Deaf professional in the current study has worked and travelled extensively overseas and has been deeply involved in international organisations and events pertaining to Deaf and disabled people. As a result of his life experiences, he used quite a few ASL signs, signs from other signed languages (ASL, etc.), and International Sign in this Auslan monologue, which taken together made it challenging for some interpreters to understand his signed source text. Partly due to the above constellation of factors, signed language interpreters’ (both native signers’ and non-native signers’) repertoire of signing often does not align perfectly with their Deaf clients’ or Deaf consumers’ repertoire of signing, thus engendering signed language comprehension problems. Most signed language interpreters in numerous countries are unbalanced bilinguals who have a spoken language as their native language and a signed language as their nonnative language (Nicodemus & Emmorey, 2013). Non-native signer interpreters’ signed language proficiency varies from one another and partly depends on when, from whom, how, and where they learned the signed language. Given the prevalence of Deaf non-native signers, some signed language interpreters may have acquired signed language from Deaf non-native signers rather than Deaf native signing models. Even among interpreters who are native signers, their signing skills (both receptive and productive skills) also differ from one another and may be influenced by home signs. To resolve signed language comprehension problems, signed language interpreters – both native signers and non-native signers – need to continuously consolidate their signed language skills, expand their signed vocabulary, understand linguistic, structural, and discourse-level differences between a signed language and a spoken language, and have exposure to a broad spectrum of Deaf signers’ signing

106 Quality, Challenges, and Strategies styles (Napier et al., 2010). Signed language interpreters or interpreting students who have primarily learned signed language in the classroom may not understand different Deaf people’s particular signing styles, comprehend the various discourse styles of signed language, make themselves easily understood by Deaf people from different educational backgrounds, or have in-depth knowledge about Deaf culture (Napier et al., 2010). By actively involving oneself in the Deaf community, signed language interpreters will have exposure to a wide range of Deaf signers, improve their own signed language receptive and productive skills, keep abreast of new signs, familiarise with regional signs, learn from Deaf signers’ use of various grammatical structures, observe how Deaf signers create coherence and logic at discourse level (e.g., through the use of space), see how Deaf people conceptualise and idiomatically express ideas in signed language in different contexts and registers, identify the discourse-level and syntactical differences between a spoken language and a signed language, observe how Deaf people adjust their signing to suit different audiences as well as informal and formal settings, know more about Deaf culture, and gradually earn Deaf people’s trust (Napier et al., 2010; Taylor, 2002). By participating in professional development activities and attending sign language interpreting conferences (e.g., ASLIA [Australian Sign Language Interpreters’ Association] national conferences), signed language interpreters can also learn new signs and/or specialised terminology, meet different Deaf and hearing signers who use regional dialects, observe diverse signing styles, familiarise with sign variations, and practise signing with a broad range of signers. Alternatively, meeting the particular Deaf presenter before an interpreting assignment can also help bridge gaps between the Deaf presenter’s linguistic repertoire and the interpreters’ linguistic repertoires. During the briefing, the Deaf presenter and the team of interpreters can discuss the big picture, context, goals, key points, gist, and structure of the signed presentation, what signs the Deaf presenter prefers for particular English concepts and jargons, and the exact English words that s/he wants to be expressed for specific signs.The interpreters can also use the opportunity to become familiar with the Deaf presenter’s sign choices, signing style, prosody, and pace. The interpreters can establish rapport, trust, and look-pause-nod cooperation strategies (Napier et al., 2008) with the Deaf professional in this joint preparation process. Results from the retrospective interview data indicate that the best practice is for the Deaf presenter to rehearse the entire signed presentation and for the interpreter to practise simultaneously rendering it into a spoken language in order to familiarise with the signed source text and build confidence. To improve interpreting students’ signed language adaptability skills, signed language interpreting training programs need to consider including more Deaf interpreters as interpreter educators, and inviting various Deaf signers as role players, presenters, or consumers. 4.5.3 Enhancing Native Language Proficiency Another important finding from the qualitative interview data is that quite a few professional signed language interpreters struggled with producing the target speech in their native language (English). In particular, the interpreters reported

Quality, Challenges, and Strategies 107 that it was difficult to interpret visual meaning from a multi-layered visual language into more linear spoken English sentences in a way that would make sense, maintain coherence within the target speech, and express a formal register in spoken language to accurately and effectively represent the high-profile Deaf professional at the conference setting.This finding indicates that signed language interpreters – both native signers and non-native signers – also need to thoroughly understand the linguistic and cultural differences between a signed language and a spoken language and continually refine their spoken (native) language proficiency. Signers and speakers use very different ways to convey, explain, and link information (Napier et al., 2010). People in the Deaf community share considerable contextual knowledge and many similar life experiences; therefore, they tend to assume that their audience knows the people, places, and events that they are referring to and thus do not provide great details about those entities (Napier et  al., 2010). The high-context communication in a signed language and the low-context communication in English require signed language interpreters to constantly decide what source language information to retain, leave out, or add to their interpretations when working in each language direction based on linguistic needs, cultural needs, and differences between the original speaker/signer and the audience in terms of life experiences (Finton & Smith, 2004; Napier & Leeson, 2015; Napier et al., 2010). People may use cohesion and coherence markers to express logical relationship between ideas (e.g., cause and effect, contrast), temporal ordering, spatial relationship, and their attitudes towards topics or people (Taylor, 2002). It appears that signed languages such as Auslan and ASL achieve coherence differently from spoken languages such as English. Deaf signers usually set up the timeframe (when, e.g., past tense) and people (who) at the beginning and rely on this initial contextual set-up to develop narratives (Napier et al., 2010;Taylor, 2002). To make ideas hang together in order to form a coherent whole, Deaf signers often use a discourse structure (e.g., introduction, body, and conclusion), full noun phrases (e.g., signs, fingerspelling, name signs), body shifting, reference points in space (pointing to a particular location where a person was previously set up in the signing space), non-manual features (e.g., direction of eye gaze), list buoys (referring to fingers on the non-dominant hand to keep track of referents mentioned in the discourse), classifier handshapes, role shift (which shows that a part of the signed discourse is presented from a particular person’s [the signer’s or someone else’s] perspective), verb modifiers, ellipsis and substitution, discourse markers (e.g., conjunctions [so, but, etc.], fillers [um, hold, etc.]), reiteration, synonyms, superordinate terms, and collocation (Finton & Smith, 2004; Johnston & Schembri, 2007; Napier et al., 2010; Taylor, 2002). In contrast, speakers of English maintain coherence in speeches or texts by using verbs that reflect tense and timeframe, pronouns that specify the gender of people, a linear sequence of words that describe how an event unfolds across time, as well as cohesive markers (such as pauses, transitional words, and phrases) (Finton & Smith, 2004; Johnston & Schembri, 2007; Napier et al., 2010; Taylor, 2002), among others.

108 Quality, Challenges, and Strategies As to the consolidation of native language skills and knowledge, signed language interpreters particularly need to expand their spoken language vocabulary, broaden their specialised knowledge and world knowledge, master a wide variety of grammatical structures, speak clearly, fluently, confidently, and effortlessly in a formal register in high-level settings, and link ideas at a discourse level to form a coherent speech (Napier et al., 2010; Taylor, 2002; Wang et al., 2015). Pursuing university education is a way to expand one’s vocabulary, develop the skills of using a formal register (Napier et al., 2010), broaden knowledge base, and acquire autonomous learning skills. Native or near-native proficiency in each of an interpreter’s working languages constitutes the foundation of interpreting skills (Napier et al., 2010; Taylor, 2002). Cultivating habits such as reading across a broad array of topics, developing innate curiosity in learning, creating bilingual glossaries about specialised fields, preparing for interpreting assignments, and enhancing conscious awareness are also likely to contribute to smooth and automatic speech production in one’s native language. 4.5.4 Cognitive Overload and Interpreting Strategies Many interpreters in this research commented that they experienced cognitive overload when working simultaneously from Auslan into English mainly due to numbers, dense information in the source text, long processing time, and syntactical asymmetry between Auslan and English. These results corroborate previous findings regarding problem triggers in spoken and signed language simultaneous interpreting (Gile, 2009; Seeber & Kerzel, 2012; Timarová et al., 2015; Wang, 2016). In addition, some interpreters reported using strategies such as adjusting processing time, making strategic additions and strategic omissions, generalising, summarising, and switching between a literal interpretation style and a free interpretation style to deal with challenges and cognitive overload in the Auslan-toEnglish simultaneous interpreting task.These findings are consistent with previous research on spoken and signed language interpreters’ linguistic coping strategies, coping tactics, and strategies in simultaneous interpreting (Bartłomiejczyk, 2006; Gile, 2009; Napier, 2004, 2016; Wang, 2016). The findings highlight the fact that simultaneous interpreting is a complex and conscious decision-making process. The aforementioned findings indicate that it is important for signed language interpreters to receive formal interpreting training, especially postgraduate-level interpreting training, in order to acquire a full range of simultaneous (and consecutive) interpreting skills, coping strategies, linguistic and extralinguisitic knowledge, an academic register, and autonomous learning skills. Most spoken language conference interpreters typically have completed postgraduate-level interpreting training to master an array of consecutive and simultaneous interpreting skills, familiarise with technical terms of specialised areas, and use a full range of registers (e.g., informal, formal registers) in their working languages across a variety of professional settings. Conference-level simultaneous interpreting is a highly complex language processing task that requires many years of specialised training and deliberate practice. Postgraduate-level interpreting training enables interpreting students to better understand discourse-level aspects of both a spoken language and a

Quality, Challenges, and Strategies 109 signed language, consolidate their bilingual skills, acquire interpreting skills for working in different settings and modes, learn strategies for coping with challenges in interpreting, enhance self-monitoring skills, and develop versatility for interpreting for a broad range of Deaf signers and hearing people. Interpreting training programs also need to strike a balance between spoken-to-signed language interpreting training and signed-to-spoken language interpreting training. 4.5.5 Familiarity and Designated or Preferred Interpreters While most interpreters in the present study encountered Auslan comprehension problems due to unfamiliarity with the Deaf presenter, three interpreters who were highly familiar with the Deaf presenter (Amber, Emily, and Jane) produced highly accurate Auslan-to-English interpretations, probably due to working together in the same workplace. These findings support the argument by Hurwitz (2008) (a Deaf professional himself) that even qualified and highly skilled signed language interpreters face challenges when interpreting Deaf professionals’ signed messages into a spoken language if they have never worked together before. Hurwitz also underscored the importance for Deaf professionals to work closely with their interpreters if they themselves want to ensure the high quality of the spoken interpretations of their formal signed presentations. For example, when an interpreter conveys his signed messages in spoken English, Hurwitz watches the interpreter to make sure that the interpreter can understand him well, and speech-reads the interpreter to gauge if the interpreter can follow him. Deaf professionals prefer highly skilled signed language interpreters who can interpret for them well without excessive requests for repetition or clarification; and familiarity between the Deaf professionals and the interpreters can reduce or minimise such requests (Hurwitz, 2008). In relation to role and responsibility, Hurwitz (2008) stated that he, as a Deaf professional, assumed the responsibility to ensure that he was understood by his hearing audience. Familiarity with both the source language speaker and the subject matter often increases the chance for interpreters to anticipate incoming information in the source speech and thus contributes to effective interpretation (Gile, 2009). Unsurprisingly, many interpreters’ unfamiliarity with the Deaf professional and his presentation topic prevented them from using anticipation as a strategy in the Auslan-to-English simultaneous interpreting task. These findings support Hauser, Finch, and Hauser’s (2008) advice that Deaf professionals should work with designated interpreters to make sure that their signed messages are understood completely and rendered accurately into a spoken language, and that their unique personality, professional image, and communication style are appropriately represented by the interpreters. While designated interpreters may be common in the United States, they are quite rare in other countries such as Australia and the United Kingdom. Therefore, an alternative approach is giving the choice for Deaf professionals to select a pool of their preferred highly skilled signed language interpreters with whom they can work regularly and to allow for sufficient time for the

110 Quality, Challenges, and Strategies interpreters to become familiar with the Deaf professionals’ sign choices, signing style, professional background, and communicative goals (De Meulder, Napier, & Stone, 2018; Harrelson, 2019; Napier et al., 2008). Familiarity with the Deaf professional and the subject matter, regular personal and professional contact, mutual trust, rapport, allocating sufficient time for preparation, information sharing for pre-assignment preparation, joint preparation, use of look-pause-nod cooperation strategies, teamwork, and partnership between the Deaf professional and the designated or preferred highly trained and skilled signed language interpreters, all contribute to the interpreters’ optimal performance of interpreting from a signed language into a spoken language in formal settings (Clark & Finch, 2008; De Meulder et al., 2018; Haug et al., 2017; Hauser & Hauser, 2008; Napier et al., 2008). If the professional interpreters in the current study had been in the room with the Deaf presenter, had been briefed by him beforehand, or had rehearsed the entire signed presentation with him, they would have had the opportunity to negotiate, agree on, and implement similar cooperation strategies before and during the signed-to-spoken language simultaneous interpreting task and would have accomplished better interpreting performance.

4.6 Summary This chapter has presented the results of the three external raters’ analytic assessment of 20 professional Auslan/English interpreters’ Auslan-to-English simultaneous interpreting performance at the mock conference setting. Results showed no significant differences between the 10 native signers and the 10 non-native signers in terms of Auslan-to-English simultaneous interpreting performance. In addition, the retrospective interview of individual interpreters has tapped into their cognitive processing during signed-to-spoken language simultaneous interpreting and their use of interpreting strategies.A thematic analysis of the i­ nterview data revealed that most interpreters in this study encountered source language (signed language) comprehension problems mainly due to unfamiliarity with the Deaf presenter and his subject matter. Many interpreters also reported that they experienced cognitive overload during the signed-to-spoken language simultaneous interpreting task. Moreover, some interpreters stated that they struggled with native language (English) production, in particular, achieving coherence and using a formal register. Interpreters also pointed out that the artificial testing conditions hindered them from using cooperation strategies with the Deaf presenter that would be available in real-life conference interpreting assignments. Finally, interpreters reported that they employed a range of strategies to deal with those challenges in signed-to-spoken language simultaneous interpreting. Although interpreters’ self-reports have shed interesting light on their cognitive processes of simultaneous interpreting, perceived challenges, and perspectives on coping strategies, I acknowledge that their perceptions are subjective and may differ from their actual interpreting behaviour. Therefore, Chapters 5, 6, and 7 will report on findings of cognitive overload and strategies in signedto-spoken language simultaneous interpreting that are based on detailed analyses of the Auslan-to-English simultaneous interpretation data.

Quality, Challenges, and Strategies 111

Notes 1 Pearson correlation coefficients (r) describe the direction and strength of the linear relationship between two continuous variables. The values of r range from −1 to +1. The sign at the front indicates whether there is a negative correlation (as one variable increases, the other decreases) or positive correlation (as one variable increases, so does the other).The bigger the absolute value of r (ignoring the sign at the front), the stronger the relationship between the two variables. Another important statistic to consider is significance level (p), which indicates how much confidence we should have in the correlation results obtained. If the p value is smaller than 0.05, it means that there is a significant correlation between the two variables. If the p value is larger than 0.05, it means that the correlation results do not reach statistical significance. For example, as shown in Table 4.1, there was a significant, positive, and strong correlation between Morgan’s total score and Ashley’s total score, r = 0.82, N = 11, p < 0.01, indicating that there was a close relationship (i.e., a high level of consistency) between the two raters’ scores. 2 Johnston and Schembri (2007, p. 73) define International Sign as ‘a form of signed language using a special lexicon devised to assist communication between deaf people who had no language in common (either signed or written).’ 3 When this Deaf professional gave a similar formal Auslan presentation on the UNCRPD at a conference in Australia many years ago, he actually introduced these innovated signs before starting his presentation to inform the audience that those were the signs that he was going to use for the particular English concepts (personal communication with Ms Teresa Cumpston-Bird, who attended that presentation). 4 https://aslsignbank.haskins.yale.edu/dictionary/gloss/2092.html. 5 https://aslsignbank.haskins.yale.edu/dictionary/gloss/2046.html. 6 I would like to thank Professor Brenda Nicodemus for initiating a discussion on this issue and for contributing ideas that are incorporated into the following five paragraphs.

References Bartłomiejczyk, M. (2006). Strategies of simultaneous interpreting and directionality. Interpreting, 8(2), 149–174. doi: 10.1075/forum.4.2.01bar. Bontempo, K., Napier, J., Hayes, L., & Brashear, V. (2014). Does personality matter? An international study of sign language interpreter disposition. Translation & Interpreting, 6(1), 23–46. Clark, P., & Finch, K. L. (2008). Interpreting in the work-related social setting: How not to trump your partner’s ace. In P. C. Hauser, K. L. Finch, & A. B. Hauser (Eds.), Deaf Professionals and Designated Interpreters: A New Paradigm (pp. 58–65). Washington, DC: Gallaudet University Press. Cokely, D. (1992). Interpretation: A Sociolinguistic Model. Burtonsville, MD: Linstok Press. De Meulder, M., Napier, J., & Stone, C. (2018). Designated or preferred? A deaf academic and two signed language interpreters working together for a PhD defense: A case study of best practice. International Journal of Interpreter Education, 10(2), 5–26. 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: Still Shining After 25 Years (pp. 125–143). Conference of Interpreter Trainers. Gile, D. (2009). Basic Concepts and Models for Interpreter and Translator Training (Rev. ed.). Amsterdam/Philadelphia: John Benjamins. Harrelson, P. (2019). Deaf employees’ perspectives on effective American Sign Language-English interpreting in the workplace:An investigation using the Critical IncidentTechnique. (Unpublished doctoral dissertation). Gallaudet University, Washington, DC, United States.

112 Quality, Challenges, and Strategies Haug,T., Bontempo, K., Napier, J., Nicodemus, B., van den Bogaerde, B., &Vermeerbergen, M. (2017). Deaf leaders’ strategies for working with signed language interpreters: An examination across seven countries. Across Languages and Cultures, 18(1), 107–131. doi: 10.1556/084.2017.18.1.5. Hauser, A. B., & Hauser, P. C. (2008). The deaf professional-designated interpreter model. In P. C. Hauser, K. L. Finch, & A. B. Hauser (Eds.), Deaf Professionals and Designated Interpreters: A New Paradigm (pp. 3–21). Washington, DC: Gallaudet University Press. Hauser, P. C., Finch, K. L., & Hauser, A. B. (Eds.). (2008). Deaf Professionals and Designated Interpreters: A New Paradigm. Washington, DC: Gallaudet University Press. Hurwitz, T. A. (2008). Foreword. In P. C. Hauser, K. L. Finch, & A. B. Hauser (Eds.), Deaf Professionals and Designated Interpreters: A New Paradigm (pp. vii–x). Washington, DC: Gallaudet University Press. Johnston, T., & Schembri, A. (2007). Australian Sign Language (Auslan): An Introduction to Sign Language Linguistics. Cambridge, UK: Cambridge University Press. Kalata-Zawłocka, A. (2016). Introduction. In A. Kalata-Zawłocka & B. van den Bogaerde (Eds.), To Say or Not to Say – Challenges of Interpreting from Sign Language to Spoken Language. Proceedings of the 23rd efsli Conference in Warsaw, Poland, 11th–13th September 2015 (pp. 7–14). European Forum of Sign Language Interpreters (efsli). Lawrence, S. (2003). Interpreter discourse: English to ASL expansion. In E. A. Winston (Ed.), Mapping Our Course: A Collaborative Venture. Proceedings of the Tenth National Convention of the Conference of Interpreter Trainers (pp. 205–216). Conference of Interpreter Trainers. Lucas, C., & Valli, C. (1992). Language Contact in the American Deaf Community. San Diego, CA: Academic Press. McNamara, T. (2000). Language Testing. Oxford, UK: Oxford University Press. Napier, J. (2004). Interpreting omissions: A new perspective. Interpreting, 6(2), 117–142. doi: 10.1075/intp.6.2.02nap. Napier, J. (2016). Linguistic Coping Strategies in Sign Language Interpreting.  Washington, DC: Gallaudet University. Napier, J., Carmichael, A., & Wiltshire, A. (2008). Look-pause-nod: A linguistic case study of a deaf professional and interpreters working together. In P. C. Hauser, K. L. Finch, & A. B. Hauser (Eds.), Deaf Professionals and Designated Interpreters: A New Paradigm (pp. 22–42). Washington, DC: Gallaudet University Press. Napier, J., & Leeson, L. (2015). Signed language interpreting. In F. Pöchhacker (Ed.), Routledge Encyclopedia of Interpreting Studies (pp. 376–381). New York: Routledge. Napier, J., McKee, R., & Goswell, D. (2010). Sign Language Interpreting:Theory and Practice in Australia and New Zealand (2nd ed.). Sydney, NSW: The Federation Press. Napier, J., Rohan, M., & Slatyer, H. (2005). Perceptions of bilingual competence and preferred language direction in Auslan/English interpreters. Journal of Applied Linguistics, 2(2), 185–218. doi: 10.1558/japl.2005.2.2.185. Nicodemus, B., & Emmorey, K. (2013). Direction asymmetries in spoken and signed language interpreting. Bilingualism: Language and Cognition, 16(3), 624–636. doi: 10.1017/S1366728912000521. Nicodemus, B., & Emmorey, K. (2015). Directionality in ASL-English interpreting: Accuracy and articulation quality in L1 and L2. Interpreting, 17(2), 145–166. doi: 10.1075/intp.17.2.01nic. Nilsson, A.-L. (2016). Interpreting from signed language into spoken language: The skills and knowledge needed to succeed. In A. Kalata-Zawłocka & B. van den Bogaerde (Eds.), To Say or Not to Say – Challenges of Interpreting from Sign Language to Spoken Language. Proceedings of the 23rd efsli Conference in Warsaw, Poland, 11th–13th September 2015 (pp. 15–48). European Forum of Sign Language Interpreters (efsli).

Quality, Challenges, and Strategies 113 Patrie, C. J. (1992). Fingerspelled word recognition skills in sign language interpreters: A comparison between novice and experienced interpreters. Journal of Interpretation, 5, 51–90. Patrie, C. J. (2009). Cognitive Processing Skills in ASL. San Diego, CA: DawnSignPress. Seeber, K. G., & Kerzel, D. (2012). Cognitive load in simultaneous interpreting: Model meets data. International Journal of Bilingualism, 16(2), 228–242. doi: 10.1177/1367006911402982. Shaw, R. (1987). Determining register in Sign-to-English interpreting. Sign Language Studies, 57(Winter), 295–322. doi: 10.1353/sls.1987.0019. Taylor, M. M. (2002). Interpretation Skills: American Sign Language to English. Edmonton, Alberta: Interpreting Consolidated. Timarová, Š., Cenková, I., Meylaerts, R., Hertog, E., Szmalec, A., & Duyck, W. (2015). Simultaneous interpreting and working memory capacity. In A. Ferreira & J. W. Schwieter (Eds.), Psycholinguistic and Cognitive Inquiries into Translation and Interpreting (pp. 101–126). Philadelphia: John Benjamins. van Dijk, R., Boers, E., Christoffels, I., & Hermans, D. (2011). Directionality effects in simultaneous language interpreting: The case of sign language interpreters in the Netherlands. American Annals of the Deaf, 156(1), 47–55. doi: 10.1353/aad.2011.0013. Wang, J. (2016). The relationship between working memory capacity and simultaneous interpreting performance: A mixed methods study on professional Auslan/English interpreters. Interpreting, 18(1), 1–33. doi: 10.1075/intp.18.1.01wan. Wang, J., & Napier, J. (2015). Directionality in signed language interpreting. Meta, 60(3), 518–541. doi: 10.7202/1036141ar. Wang, J., Napier, J., Goswell, D., & Carmichael, A. (2015). The design and application of rubrics to assess signed language interpreting performance. The Interpreter and Translator Trainer, 9(1), 83–103. doi: 10.1080/1750399X.2015.1009261.

5 Processing Time Regarding Sentence Beginnings, Numbers, and Negation

Given that many interpreters stated in the post-task interview that numbers, long processing time, and syntactical differences between Auslan and English sometimes led to cognitive overload and interpretation errors (see Chapter 4), the present chapter focuses on quantitative analysis of the 20 professional Auslan/English interpreters’ processing time of 40 selected Auslan sentences that contained numbers or negation. Some of the Auslan sentences in the presentation featured multiple challenges (such as numbers and syntactical differences between Auslan and English) within one sentence.While a multitude of studies (e.g., Collard & Defrancq, 2019; Defrancq, 2015; Timarová et al., 2015; Timarová, Dragsted, & Hansen, 2011) have investigated spoken language interpreters’ processing time during simultaneous interpreting, scarce scholarly attention (e.g., Cokely, 1986; McKee & Napier, 2002; Wang, 2020) has been devoted to exploring processing time in signed language simultaneous interpreting.The current study bridges this gap and contributes novel knowledge regarding (i) professional interpreters’ average onset processing time in signed-to-spoken language simultaneous interpreting, (ii) the interpreters’ number processing time (also referred to as processing time regarding/for numbers), (iii) processing time for accurate renditions of numbers in comparison with processing time for inaccurate renditions of numbers, (iv) onset processing time (also referred to as onset time lag, namely the time interval between the beginning of a source language sentence and the beginning of its corresponding target language sentence) versus number processing time, (v) the impact of overly short or long processing time on the interpreting process and interpreting output, as well as (vi) the relationship between processing time and interpreting strategies. First, this chapter briefly introduces the Auslan sentences for which onset processing time was measured, reports on the interpreters’ average onset processing time, and presents findings regarding variability in processing time both within and between interpreters.The chapter then compares processing time for correct interpretations of numbers with processing time for incorrect interpretations of numbers, and presents findings from a comparison of onset processing time and processing time regarding numbers in order to investigate if the interpreters shortened their processing time in an attempt to render the numbers accurately in the target language. Apart from the quantitative analysis, qualitative analysis of representative examples of Auslan-to-English simultaneous

Processing Time 115 interpretation data is presented to illustrate that either excessively short or long onset processing time severely affects simultaneous interpreting process and performance. Another example is used to illustrate that both external and internal factors cause the length of onset processing time to fluctuate from sentence to sentence. By measuring and analysing interpreters’ processing time in the Auslan-toEnglish simultaneous interpreting task, this chapter contributes to the answering of the second and third research questions of the study: 1 What causes cognitive overload in simultaneous interpreting from a signed language into a spoken language? 2 What strategies do signed language interpreters employ to cope with the challenges and cognitive overload in simultaneous interpreting from a signed language into a spoken language?

5.1 Measuring Processing Time on 40 Auslan Sentences A total of 40 Auslan sentences (see Table 5.1) scattered throughout the signed presentation were selected for measuring interpreters’ onset processing time, namely the time interval between the start of an Auslan sentence and the beginning of its corresponding English rendition. Nearly all these Auslan sentences included numbers or negation, which have been identified as challenges for some signed language interpreters when working from a signed language into a spoken language (Taylor, 2002). As to 29 Auslan sentences that contained numbers,1 processing time for numbers (the temporal interval between the beginning of an Auslan number and the beginning of its corresponding English interpretation) was measured in ELAN in preparation for comparing onset processing time with number processing time. Regarding syntactical differences between Auslan and English, 9 of the 29 Auslan sentences (28, 85, 86, 87, 88, 97, 98, 99, and 100) contained numbers near or at the end of the sentence which need to be re-ordered to the beginning or middle of idiomatic English sentences. Numbers in the remaining 20 Auslan sentences containing numbers do not need to be moved syntactically when transferred into English. Ten Auslan sentences featured negation (note that Auslan sentence 30 contained both a number and negation), with seven sentences (29, 89, 105, 107, 108, 110, and 111) ending with a negating sign (e.g., not, nothing) or a headshake, and the remaining three sentences having a negating sign in the middle of the sentence. Another two Auslan sentences (106 and 109) did not contain numbers or negation but were adjacent to some of the Auslan sentences characterised by end negation and therefore were included in the analysis of processing time. It would be ideal to measure processing time on more Auslan sentences. However, since measuring processing time was an extremely time-consuming process (largely due to the need to manually identify the measuring points), I decided to gauge processing time only on these 40 Auslan sentences for quantitative analysis, then to conduct detailed qualitative analysis of how the processing

116 Processing Time Table 5.1  Forty Auslan sentences selected for measuring and analysing processing time Syntactical features

Sentence Auslan sentences No.

Middle negation (MN1)

6

Number at the end (NE1)

28

End negation (EN1) Middle negation (MN2) and number Middle negation (MN3) Number

29

Number

53

Number

54

Number

55

Number

62

Number

63

30

31 51

other too some country nothing national a-s-s-o-c deaf link world-federation-of-the-deaf still post maybe have group deaf people in that country// (11.8 seconds) In those countries where there is no national association of the Deaf affiliated to WFD, the package was still sent to any Deaf groups that were in existence. so altogether how many responses we receive survey (research) responses altogether how many? ninetythree// (11.3 seconds) So, altogether we received in total 93 responses to our survey. but ninety-three those full link w-f-d? no// (5.2 seconds) However, not all of these were from organisations affiliated to the WFD. they group sixty-eight member w-f-d other group those twenty-five not member w-f-d// (4.6 seconds) Sixty-eight responses were from members of the WFD, 25 were from non-member groups. s-o mean those twenty-five countries nothing national a-s-s-o-c o-f deaf// (5.6 seconds) So those 25 countries do not have a National Association of the Deaf. u-n convention passed finish point u-n before 2006// The UN convention was passed at the UN in 2006. australia government finish sign r-a-t-i-f-i-e-d in 2008// The Australian government ratified the convention in 2008. one important part o-f that (hesitation) agreement convention that have five different article article article article article relate what? sign language out of thirty article article article// In one important part of the convention there are five articles that specifically relate to signed languages out of a total of 30 articles. but have eight times mention mention mention sign language put-in put-in put-in// However, signed languages are mentioned eight times. s-o well like have how many o-f those country finish recognise their deaf people out-there? only seventyseven respond from ninety-three country say yes deaf people recognise equal// So, how many of those countries that responded to our survey replied that they recognised Deaf people in their country? Only 77 out of the 93 countries replied that Deaf people were recognised as equals to hearing people. but what about sign language what? finish they have approve confirm point sign language? only forty-four country nod say government finish approve sign language o-u-t o-f ninety-three country hmmm// But, what about the formal recognition of the national signed language itself? Only 44 countries replied that their governments have formally recognised their national signed language, out of 93 countries. (Continued)

Processing Time 117 Table 5.1  (Continued) Syntactical features

Sentence Auslan sentences No.

Number

64

Number

65

Number

67

Number

69

Number

70

Number

71

Number

72

Number

73

Number

77

Number

78

that surprise under half altogether// That equates to less than half of the total number of respondents. what about sign language research? have have have what what what? only have fifty-four finish research sign language make that book lick-finger-flick-page flickpage flick-page d-i-c-t-i-o-n-a-r-y// And, how many countries had signed language research happening? Only 54 of the respondents reported that research had been undertaken and that dictionaries of their signed languages had been produced. doesn’t mean all countries those forty-four have same level approval// Of the 44 countries that have recognised their signed language, not all have done so to the same extent. s-o those forty-four country say yes my government finish recognise my sign only ten of those forty-four yes where? constitution level// Of those 44 respondents that answered yes to their government recognising their signed language, only 10 of those 44 had done so at the constitution level, other group nineteen finish approve through legislation// whereas 19 had recognition within legislation, other nineteen country those finish approve their sign language through what? policy// with another 19 countries having done so within policy. so that means only ten country round world actually recognise sign language ten// So that means that only 10 countries around the world have actually, formally recognised their signed languages. nineteen point well legislation also high level o-f approve// However, the 19 countries that recognised signed languages within their legislation can be seen as also having ‘ratified’ the language at a high level. group those country respond say eighty-eight country finish say yes they recognise their children right t-o access their education good eighty-eight wow// From the respondents, 88 countries answered that Deaf children’s rights to access an education had been recognised, which seems an impressive number. what about legislation level what? well relate t-o deaf education only sixty-one country say nod have have legislation relate t-o deaf education// But, what existed at a legislative level regarding Deaf education? Only 61 countries replied that they had legislation pertaining to Deaf education. (Continued)

118 Processing Time Table 5.1  (Continued) Syntactical features

Sentence Auslan sentences No.

Number

81

Number near the end (NE2)

85

Number near the end (NE3)

86

Number at the end (NE4)

87

Number at 88 the end (NE5) End negation 89 (EN2) Number 96

Number near the end (NE6)

97

Number at the end (NE7)

98

Number at the end (NE8)

99

Number at the end (NE9)

100

them how many? eighty-eight country have deaf schools// Of the respondents, 88 replied that they had Deaf schools in their country. bilingual have nod plus have other altogether countries have what twenty-three countries have that// (8.6 seconds) There were 23 countries that reported having a bilingual education option alongside other approaches. other group say have total communication plus other methods education thirty-five countries// (8.8 seconds) Another group of countries, numbering 35, described having a total communication approach with other methods of education available. those only total communication speak-sign altogether thirty-one// (7.6 seconds) Thirty-one respondents replied that their country had a total communication approach only, which incorporates speaking and signing at the same time. other group only oral sign ban five// (5.6 seconds) A further five countries reported an oral-only approach that bans signed language. means bilingual education high level not// (4.6 seconds) This meant that bilingual education methods were not in high use. respond eighty-two country say yes deaf people can access t-o government wow// Eighty-two countries replied positively that Deaf people could access government – an impressive response. but what about sign language interpreter services have have have? what? provide have? sixty-two country// (8.1 seconds) But what about signed language interpreting services? Sixty-two countries replied that they had signed language interpreting services. so sixty-two interpreting services have but how many those finish have training qualified professional interpreters qualified have how many? only fortythree// (15.2 seconds) Of those 62, how many had interpreters qualified to a professional level? Only 43 responded positively. that forty-three group those how many have c-o-d-e-of-e-t-h-i-c-s how many? only thirty-one// (12.3 seconds) And of those 43, how many had a code of ethics in their country? Only 31. important part that group look-over-there look-here government responsible pay interpreters have? how many countries what? thirty-two// (10.4 seconds) An important part of the research is whether governments take responsibility for paying for interpreting services. Of those surveyed, only 32 countries replied that they did. (Continued)

Processing Time 119 Table 5.1  (Continued) Syntactical features

Sentence Auslan sentences No.

Number

101

End negation (EN3)

105

106

End negation (EN4)

107

End negation (EN5)

108

109

End negation (EN6)

110

End negation (EN7)

111

s-o talk about sign language interpreting service have sixty-two those thirty-two government responsible pay// So, then you have to consider that 62 countries replied that they had signed language interpreting services, but only 32 responded that their government paid for interpreting services. s-o altogether research talk many many block (prevent) deaf people no few countries those put-down for deaf people access to education government or perception equal same same other people// (20.9 seconds) So, in summary, what does the research tell us? Firstly, we found that only in a few countries were Deaf people marginalised in their societies, and that their situation was not good in terms of access to education and government and also in terms of how well Deaf people were seen as equal to other people. b-u-t have what l-a-c-k o-f recognise approve sign language also l-a-c-k o-f bilingual education// (13.2 seconds) But, there was a distinct lack of countries where the national signed language had been formally recognised. Also, there was a lack of responses regarding bilingual education. many many many have sign language interpreting services? not really// (5.4 seconds) Not many countries reported having signed language interpreting services. people community know about deaf people? no// (5.9 seconds) We found that the hearing community was not very aware of the situation of Deaf people. means many those deaf people their lives can access t-o services really oppressed// (8.3 seconds) This, in turn, meant that for many of these Deaf people, they were experiencing oppression in terms of accessing services. so those deaf people true enjoy enjoy enjoy equal human rights? no// (9.5 seconds) So, all in all, it seems that Deaf people do not yet enjoy the full and equal human rights. so summarise that say deaf people equal like other round world have have have? no// (9.5 seconds) So, in closing, it is fair to say that Deaf people the world over are not yet equal to their hearing counterparts.

Note: The seconds in the brackets indicate the length of the specific Auslan sentence.

120 Processing Time time of adjacent Auslan sentences impacted on both the simultaneous interpreting process and the performance. A combination of the quantitative analysis and the qualitative analysis of processing time regarding challenging sentences and segments of the source text appeared to have yielded a better understanding of the cognitive process of simultaneous interpreting than measuring processing time of the whole source text alone. The following tiers were added to ELAN for measuring different types of processing time: 1 Onset processing time (i.e., onset time lag): On this tier, an annotation was created to mark the temporal delay between the beginning of an Auslan sentence and the beginning of its corresponding English interpretation. Onset processing time sheds light on the interpreter’s processing unit and the amount of information that he or she needs in order to be able to start interpreting a source language sentence. 2 Tail-to-tail span: An annotation was created to measure the time interval between the end time of an Auslan sentence and the end time of its corresponding English interpretation. Tail-to-tail span indicates how far the interpreter lags behind the Deaf presenter in finishing up the English rendition of an Auslan sentence.Tail-to-tail span was measured for only 21 Auslan sentences (the 10 Auslan sentences containing negation, 9 Auslan sentences with numbers, and another 2 Auslan sentences without these features but adjacent to sentences with negation). 3 Number processing time (i.e., number time lag, processing time regarding numbers, processing time for numbers): An annotation was created to mark the onset of an Auslan number and the onset of its English equivalent in the interpretation. Number processing time reflects the length of time that the interpreter waits before interpreting the specific number. Number processing time was measured for all the 29 Auslan sentences containing numbers. 4 Processing time for specific lexical items (e.g., technical terms, key signs): In a similar vein, an annotation was added to signal the time period between the beginning of a particular sign and the beginning of its English equivalent. This is especially useful for speculating on potential reasons for interpretation errors when an interpreter distorts an important lexical item in the source text. Processing time in this study was represented as a number of seconds that an interpreter lagged behind the Deaf presenter. Figure 5.1 illustrates onset processing time, tail-to-tail span, and processing time for the number five by using Auslan sentence 88 (other group only oral sign ban five) and an interpreter Jane’s English interpretation as an example. The rationale for measuring onset processing time, tail-to-tail span, and number processing time was to compare findings of the current study with the findings of previous studies (e.g., (Cokely, 1986; Collard & Defrancq, 2019, 2021; Lee, 2002;Timarová et al., 2014) in order

Processing Time 121

Figure 5.1  An example from ELAN of measuring three types of processing time (onset time lag, tail-to-tail span, and number time lag)

to assess whether there is any consistency in result patterns.While corpus-based studies on spoken language simultaneous interpreting (e.g., Collard & Defrancq, 2019, 2021; Defrancq, 2015) typically measured professional interpreters’ processing time by attaching time tags to the onset of source language words and their target language equivalents, I measured onset processing time as the time interval between the onset of an Auslan sentence and the onset of its equivalent English sentence because onset processing time was measured in many previous studies (e.g., Cokely, 1986; Timarová et al., 2011), and interpreter educators who are also experienced interpreters (e.g., Jones, 2002; Nilsson, 2016; Setton & Dawrant, 2016;Taylor, 2002) often advise students on waiting to truly understand the speaker’s/signer’s message before starting to interpret during simultaneous interpreting. It is important to discuss some prominent methodological issues related to the measurement of processing time in signed-to-spoken language simultaneous interpreting, which may also be encountered by other researchers in conducting further studies on processing time in signed language interpreting. While it is possible to rely on prosodic markers in order to divide signed language discourse into signed sentences (Nicodemus, 2009), sentence boundaries in signed language discourse are under-researched and seem not as clear-cut as those in spoken language discourse. Thus, there was still a degree of subjectivity in determining the start (e.g., the hand clasp position versus the recognisable handshape of a particular sign) and end of Auslan sentences in the signed presentation of this study. The same subjectivity also applies to the start and end of Auslan signs. As in Timarová et al.’s (2011) study, processing time was measured when an Auslan sentence was either completely or partially rendered into English, with only Auslan sentences omitted in their entirety being marked as missing values. Due to the use of software (ELAN 5.2) that can precisely measure processing time in simultaneous interpreting, the length of processing time was measured in milliseconds (e.g., 1.331 seconds), which was then used for quantitative analysis. However, given the unavoidable imprecision in the human detection of measurement points (the start of a sentence, etc.), the precision level of processing time for qualitative analysis of interpretation data was set at 0.1 second, a demarcation point that is in line with some previous studies (e.g., Lee, 2002; Timarová et al., 2011). Another issue to acknowledge is that the average length of processing time is highly likely to change if more source language sentences were included for measuring interpreters’ processing time.

122 Processing Time

5.2 Variability in Onset Processing Time Table 5.2 shows the descriptive statistics of individual interpreters’ onset processing time in relation to the aforementioned 40 Auslan sentences. The 20 professional Auslan/English interpreters rendered the vast majority of these Auslan sentences into spoken English, with some interpreters omitting a few of the entire sentences. Based on individual interpreters’ mean onset processing time in the fourth column of Table 5.2, the overall mean onset processing time for the 20 interpreters was 4.12 seconds (i.e., mean of the 20 interpreters’ mean onset processing time, SD = 0.88) and the overall median onset processing time was 3.9 seconds (i.e., mean of the 20 interpreters’ median onset processing time). The individual interpreters’ mean onset processing time ranged from 2.43 seconds (Lauren) to 6.36 seconds (Liz), indicating marked between-subject variability in terms of mean onset processing time. The length of onset processing time ranged between 0.56 (see Annie’s lower limit) and 15.40 seconds (see Tiffany’s upper limit).The results showed no negative onset processing time for these 40 Auslan sentences, implying that none of the interpreters was ahead of the Deaf presenter when processing the beginning of these Auslan sentences. Given the propensity for prediction, this finding may seem surprising; however, interpreters likely tend to predict a source message

Table 5.2  Descriptive statistics of each interpreter’s onset processing time regarding the 40 Auslan sentences Pseudonym Valid Missing Mean Standard Range values values (seconds) deviation (seconds)

Median Mean–Median (seconds) (seconds)

Liz Amber Emily Wendy Linda Jane Molly Sophia Helen Claire Alex Annie Debbie Shannon Zoe Tiffany Monica Lauren Bernie Kay

5.48 4.19 5.07 3.29 4.85 3.37 3.15 3.83 2.98 2.65 3.91 3.64 3.58 3.50 2.89 3.52 3.89 2.12 3.47 3.10

36 40 39 40 36 40 39 40 40 40 40 40 38 40 40 38 37 40 40 38

4 0 1 0 4 0 1 0 0 0 0 0 2 0 0 2 3 0 0 2

6.36 4.52 5.16 3.85 5.38 3.5 3.85 4.5 3.65 3.13 4.22 4.07 3.95 3.65 3.38 4.50 4.91 2.43 3.7 3.59

3.21 1.93 1.90 1.83 2.82 1.74 2.15 2.48 2.00 1.48 1.99 2.40 1.92 1.68 1.73 2.97 2.89 1.27 1.99 1.63

0.67–13.50 1.64–10.43 2.26–11.19 1.66–9.23 1.41–12.05 1.13–8.66 1.07–10.01 1.70–11.64 1.08–9.60 1.10–6.74 2.06–10.55 0.56–9.30 1.66–11.17 1.16–7.46 1.23–7.83 1.53–15.40 1.37–11.18 0.90–7.30 1.13–9.47 1.76–9.70

0.88 0.33 0.09 0.56 0.53 0.13 0.70 0.67 0.67 0.48 0.31 0.43 0.37 0.15 0.49 0.98 1.02 0.31 0.23 0.49

Processing Time 123 that is in the middle or at the end of a sentence rather than at the beginning of the sentence. For each interpreter, the range of onset processing time regarding the 40 Auslan sentences was rather large (e.g., 1.53–15.40 seconds for Tiffany), indicating remarkable intra-subject variations in onset processing time. In other words, each interpreter’s onset processing time varied considerably from one sentence to another instead of remaining stable throughout the Auslan-toEnglish simultaneous interpreting task. As can be seen in the far-right column of Table 5.2, interpreters’ mean onset processing time was generally larger than median onset processing time; a pairedsamples t-test confirmed that the 20 interpreters’ mean onset processing time (Mean  = 4.12, SD = 0.88) was significantly longer than their median onset processing time (Mean = 3.62, SD = 0.81), t(19) = 8.18, p < 0.001, effect size = 0.78 (large) (the Wilcoxon Signed Rank Test, namely the non-parametric alternative of the paired-samples t-test, revealed the same result pattern). In addition, Figure 5.2 illustrates three top-performing interpreters’ (Amber, Jane, and Zoe) onset processing time of these 40 Auslan sentences.The wavy line indicates that the interpreter’s onset processing time fluctuated from one measurement point (sentence beginning) to another rather than remaining consistent throughout the Auslan-to-English interpreting task. It can be seen that different interpreters showed different processing time patterns. Furthermore, the three interpreters’ onset processing time for some Auslan sentences (e.g., 29, 30, 87) was similar; however, their onset processing time varied considerably for other Auslan sentences (e.g., 54, 63, 97). Taken together, these results indicate that processing time appears to be a dynamic, elusive, and complicated variable during simultaneous interpreting, and that processing time often varies from person to person, and from one measurement point to another.

11 10 9 Onset me lag

8

Amber

Jane

Zoe

7 6 5 4 3 2 0

6 28 29 30 31 51 53 54 55 62 63 64 65 67 69 70 71 72 73 77 78 81 85 86 87 88 89 96 97 98 99 100 101 105 106 107 108 109 110 111

1

40 Auslan sentences

Figure 5.2   Three interpreters’ onset processing time for the 40 selected Auslan sentences

124 Processing Time

5.3 Accuracy of Interpreting Numbers and Processing Time Regarding Numbers This section first presents findings regarding interpreters’ accuracy rate of rendering 32 Auslan numbers in the Auslan source text (see Table 5.1 and Table 5.4), then reports on results of comparing processing time for the accurate renditions of numbers with processing time for the inaccurate interpretations of numbers in order to investigate if interpreters shortened their processing time in an attempt to transfer numbers correctly into the target language. It is noteworthy that most of these 32 Auslan numbers were not repeated in their immediate context; therefore, interpreters’ erroneous renditions of these numerals (e.g., omission, substitution) would result in the hearing audience misunderstanding the Deaf presenter’s concrete and important information, and interpreters’ generalisation and paraphrasing of these numbers may be acceptable overall but probably still lead to ambiguity and a loss of precise meaning in the target text. 5.3.1 Error Taxonomy for Number Interpretations The number error taxonomy distinguishes the following four types of inaccurate or imprecise renditions of numbers. Omission refers to an instance when an interpreter left out a number altogether. Substitution refers to a situation in which an interpreter distorted the number into a different number (e.g., an interpreter changed ‘2006’ in Auslan sentence 51 to ‘2009’) or changed a large number to a general expression that indicated a small number (e.g., an interpreter replaced eighty-eight in Auslan sentence 81 with the expression ‘only a few of those’). Generalisation shows indications that an interpreter provided a potentially acceptable general expression (e.g., ‘some,’ ‘many,’ ‘quite a number,’ ‘a lot of ’) of the precise number (e.g., twenty-five). In addition, paraphrasing refers to a situation in which the interpreter used an acceptable but imprecise alternative to the specific number (e.g., in Auslan sentence 30, the Deaf presenter said that of the 93 countries that responded to the survey, 68 countries were members of the World Federation of the Deaf and the other 25 countries were non-members, but the interpreter replaced the specific number 25 with the acceptable expression ‘the others’ that required the audience to mentally calculate the number). From the perspective of strategies, generalisation can be considered as an interpreting strategy that is employed by an interpreter in order to cope with an interpreting difficulty, manage the interpreting process, and produce a smooth interpretation. In this chapter focusing on numbers, such generalisations of numbers are considered as errors because they do result in ambiguous and imprecise renditions in comparison with the precise numerical information provided by the Deaf presenter. It is worth noting that some previous studies on numbers in simultaneous interpreting (e.g., Braun & Clarici, 1996; Pinochi, 2009) defined omissions of numbers as leaving out a number altogether or replacing it with a generic expression such as ‘many’ or ‘few’; that is, the omissions of numbers

Processing Time 125 reported in these previous studies may include instances that would be coded as omissions and generalisations in the present study. 5.3.2 Accuracy Rate of Interpretations of Individual Numbers Table 5.3 summarises each professional Auslan/English interpreter’s total accurate English renditions of Auslan numbers. Only one interpreter (Shannon) transferred all the 32 Auslan numbers precisely into English; and three other interpreters (Jane, Annie, and Lauren) rendered 31 numbers correctly into the target language. The mean accuracy rate of the 20 interpreters’ English renditions of the 32 Auslan numbers was 89%. In addition,Table 5.4 shows the accuracy rate regarding English interpretations of each of the 32 Auslan numbers and summarises the number and type of number errors. It is worthy of attention that the accuracy rate for all but five numbers (under half in Auslan sentence 64, fifty-four in Auslan sentence 65, nineteen in Auslan sentence 71, nineteen in Auslan sentence 73, and thirty-five in Auslan sentence 86) was 80% or higher. It is also worth noting that the Deaf presenter reiterated some of the numbers (e.g., sixty-two in Auslan sentence 97), which may have refreshed the interpreters’ memory of the numbers and contributed to the high accuracy rate. In total, there were 29 omissions of numbers, 34 substitutions of numbers, eight generalisations of numbers, and one paraphrasing of a number. In other words, distortions and omissions were the most common types of incorrect or imprecise Table 5.3  Each interpreter’s total accurate English interpretations of 32 Auslan numbers Pseudonyms

Native or non-native signer

Total of accurate interpretations of 32 numbers

Percentage of accurate interpretations of numbers

Liz Amber Emily Wendy Linda Jane Molly Sophia Helen Claire Alex Annie Debbie Shannon Zoe Tiffany Monica Lauren Bernie Kay Mean

Non-native Native Native Non-native Non-native Non-native Non-native Non-native Non-native Non-native Native Native Non-native Native Native Native Native Native Non-native Native

26 30 28 27 26 31 28 27 30 26 29 31 29 32 27 27 27 31 30 26 28.4

81% 94% 88% 84% 81% 97% 88% 84% 94% 81% 91% 97% 91% 100% 84% 84% 84% 97% 94% 81% 89%

126 Processing Time Table 5.4  Accuracy rate regarding English interpretations of each of the 32 Auslan numbers Slide No. Auslan Number sentence No.

Accurate interpretations Inaccurate interpretations of individual of individual numbers numbers N

%

N

%

7

20

100%

0

0%

17

85%

15%

twenty- 17 five

85%

2006 2008

19 18

95% 90%

54 54 55

five thirty eight times

19 16 16

95% 80% 80%

62

seventy- 18 seven forty18 four under 14 half

90%

3: 1 omission, 1 generalisation (‘some’), and 1 paraphrasing (‘the others’) 3: 1 substitution (‘28’) and 2 generalisations (‘some,’ ‘quite a number’) 1 substitution (‘2009’) 2: 1 omission and 1 substitution (‘2006’) 1 substitution (‘3’) 4 omissions 4: 1 omission and 3 generalisations (‘quite a number of them,’ ‘a lot of them,’ and ‘many different comments’) 2 omissions

10%

65

fiftyfour

14

70%

67

95%

69 70

forty19 four ten 19 nineteen 18

2: 1 omission and 1 substitution (‘42’) 6: 2 omissions and 4 substitutions (‘only half of,’ ‘only about 50%,’ ‘about half of the world,’ and ‘probably about half of ’) 6: 3 omissions and 3 substitutions (‘55,’ ‘4,’ and ‘less than that’) 1 substitution (‘40’)

5% 10%

71

nineteen 15

75%

72 73

ten 20 nineteen 14

100% 70%

1 omission 2: 1 omission and 1 generalisation (‘many’) 5: 2 omissions and 3 substitutions (‘9,’ ‘9,’ and ‘the remainder’) 0 6: 3 omissions and 3 substitutions (‘a higher proportion,’ ‘those other 9, that’s 1/9,’ and ‘the others’)

28 30

30 No slide 51 53

10

63 64

11

ninetythree sixtyeight

90% 70%

95% 90%

15% 5% 10% 5% 20% 20%

10%

30%

30% 5%

25% 0% 30%

(Continued)

Processing Time 127 Table 5.4  (Continued) Slide No. Auslan Number sentence No.

Accurate interpretations Inaccurate interpretations of individual of individual numbers numbers N

%

N

%

13

18

90%

10%

20

100%

2 substitutions (‘89’ and ‘87’) 0

18

90%

2: 1 substitution (‘only a few of those’) and 1 generalisation (‘a high number of ’) 0

10%

6: 1 omission and 5 substitutions (‘32,’ ‘25,’ ‘25,’ ‘25,’ and ‘15’) 0

30%

0% 15%

77 78 81

14

85

eightyeight sixtyone eightyeight

0%

twenty- 20 three thirty- 14 five

100%

thirtyone five eightytwo sixtytwo fortythree thirtyone thirtytwo

20

100%

20 17

100% 85%

20

100%

0 3 substitutions (‘62,’ ‘92,’ and ‘83’) 0

19

95%

1 substitution (‘3’)

5%

17

85%

15%

17

85%

sixtytwo 101 thirtytwo Mean accuracy rate

19

95%

3: 1 omission and 2 substitutions (‘21,’ ‘21’) 3: 2 omissions and 1 substitution (‘a smaller number’) 1 omission

18

90%

2 omissions

10%

17.75

89% 29 Omissions 34 Substitutions 8 Generalisations 1 Paraphrasing

4% 5% 1% 0%

86 87 17

88 96 97 98 99 100 101

70%

Total number of inaccurate interpretations of numbers

0%

0%

0%

15% 5%

English renditions of Auslan numbers. Taken together, the four types of number errors amounted to 72. Theoretically, there could be 640 English renditions (i.e., 32 numbers multiply 20 interpreters) of the 32 Auslan numbers. Hence, the professional Auslan/English interpreters’ error rate of rendering the Auslan numbers was 11% (72 out of 640). Specifically, the 29 omissions accounted for 4% of the 640 possible English renditions, the 34 substitutions 5%, and the eight generalisations 1%.

128 Processing Time A closer look at the inaccurate interpretations of numbers reveal that the omissions, substitutions, generalisations, and paraphrasing may be attributable to the following factors: (i) working memory overload (i.e., the interpreter forgot the number possibly due to long processing time or to multitasking, among other factors), (ii) comprehension problems of the number due to the twodimensional screen rather than the Deaf presenter signing at a real-life conference (e.g., thirty-five in Auslan sentence 86 was misinterpreted by three interpreters as ‘25’), (iii) the absence of a team interpreter, and (iv) no visual access to projected PowerPoint slides for the interpreters, among other factors. 5.3.3 Accuracy Rate of Interpretations of Numbers in Clusters The 32 Auslan numbers occurred in adjacent Auslan sentences, whose key points such as numbers and referents appeared on the tables or charts of printed PowerPoint slides. Altogether, there were eight clusters of numbers, with seven clusters having PowerPoint slides. I was curious to examine the accuracy rate of a cluster of numbers in comparison with the aforementioned accuracy rate of individual numbers. As shown in Table 5.5, the results revealed that the success rate of interpreting a cluster of Auslan numbers in adjacent sentences (mean accuracy rate being 65%) was substantially lower than the success rate of rendering individual Auslan numbers (mean accuracy rate being 89%). For example, in relation to the six Auslan sentences (67, 69, 70, 71, 72, and 73) containing numbers that were presented on slide 11, the accuracy rate of English interpretations of the six individual numbers ranged from 70% (i.e., 14 interpreters rendered the number nineteen in Auslan sentence 73 correctly) to 100% (i.e., all 20 interpreters transferred the number ten in Auslan sentence 72 accurately); however, only nine (45%) interpreters rendered all the six Auslan numbers successfully. In other words, for neighbouring Auslan sentences that contained numbers, some interpreters rendered the number in one sentence accurately but missed the number in the following sentence. 5.3.4 Processing Time for Accurate Interpretations of a Number Versus Processing Time for Inaccurate Interpretations of the Number Furthermore, quantitative analyses were conducted to investigate if processing time regarding accurate renditions of a number was significantly shorter than processing time regarding inaccurate interpretations of the number. Table 5.6 provides the analysis of the number of accurate renditions of each of the 32 Auslan numerals, the number of inaccurate renditions of each numeral, both the mean and median of processing time regarding the accurate renditions of each number, as well as the mean and median of processing time for the inaccurate renditions of each number. It should be pointed out that when measuring processing time in relation to inaccurate interpretations of a number, only substitutions, generalisations, and paraphrasing of the number were included for the measurement and omissions had to be excluded, because processing time could

Processing Time 129 Table 5.5  Accuracy rate of individual numbers versus accuracy rate of a cluster of numbers in adjacent sentences Slide No.

7

Auslan sentence No.

28 30 30 No slide 51 53 54 54 55 10 62 63 64 65 11 67 69 70 71 72 73 13 77 78 81 14 85 86 87 88 17 96 97 98 99 100 101 101 Mean accuracy rate

Number

ninety-three sixty-eight twenty-five 2006 2008 five thirty eight times seventy-seven forty-four under half fifty-four forty-four ten nineteen nineteen ten nineteen eighty-eight sixty-one eighty-eight twenty-three thirty-five thirty-one five eighty-two sixty-two forty-three thirty-one thirty-two sixty-two thirty-two

Accurate interpretations of individual numbers

Accurate interpretations of a cluster of numbers in adjacent sentences

N

%

N

%

20 17 17 19 18 19 16 16 18 18 14 14 19 19 18 15 20 14 18 20 18 20 14 20 20 17 20 19 17 17 19 18 17.75

100% 85% 85% 95% 90% 95% 80% 80% 90% 90% 70% 70% 95% 95% 90% 75% 100% 70% 90% 100% 90% 100% 70% 100% 100% 85% 100% 95% 85% 85% 95% 90% 89%

14

70%

10

50%

8

40%

9

45%

16

80%

14

70%

13

65%

12

65%

not be measured if there was no target language rendition of a source language item. As can be seen from Table 5.6, the number of accurate renditions of each numeral was 14 or above and the number of inaccurate interpretations (omissions, substitutions, generalisations, and paraphrasing) of each numeral was six or smaller.The accurate and inaccurate renditions of a number could be considered as two different samples for statistical analysis. However, unequal sample sizes may lead to problems such as (i) unequal variances between samples which make

130 Processing Time Table 5.6  Processing time for accurate and inaccurate English interpretations of Auslan numbers Auslan Number sentence No. 28 30 30 51 53 54 54 55 62 63 64 65 67 69 70 71 72 73 77 78 81 85 86 87 88 96 97 98 99 100 101 101

Processing time for accurate interpretations of a number

Processing time for inaccurate interpretations of the number

N

Mean (SD)

Median

N

Mean (SD)

Median

20 17 17 19 18 five 19 thirty 16 eight times 16 seventy-seven 18 forty-four 18 under half 14 fifty-four 14 forty-four 19 ten 19 nineteen 18 nineteen 15 ten 20 nineteen 14 eighty-eight 18 sixty-one 20 eighty-eight 18 twenty-three 20 thirty-five 14 thirty-one 20 five 20 eighty-two 17 sixty-two 20 forty-three 19 thirty-one 17 thirty-two 17 sixty-two 19 thirty-two 18

1.95 (0.76) 2.74 (1.02) 1.85 (0.28) 2.34 (1.32) 2.13 (1.10) 2.46 (0.82) 2.64 (2.09) 3.56 (1.22) 3.68 (2.36) 2.35 (2.23) 1.81 (0.60) 2.44 (1.04) 3.05 (2.75) 2.70 (1.78) 2.33 (1.31) 2.93 (1.59) 2.50 (1.69) 2.67 (0.88) 1.93 (0.80) 1.80 (0.59) 3.02 (2.49) 2.55 (1.37) 3.21 (1.18) 2.13 (1.11) 2.59 (1.68) 2.98 (2.25) 2.00 (1.10) 2.07 (1.00) 2.66 (1.53) 3.55 (2.18) 3.25 (1.94) 4.41 (2.39)

1.82 2.77 1.90 1.90 1.91 2.43 2.29 3.54 2.73 1.41 1.59 2.31 2.35 1.99 1.76 2.46 1.89 2.70 1.69 1.74 2.24 2.33 3.45 1.80 2.46 1.89 1.63 1.80 2.23 3.23 2.29 3.63

0 2 3 1 1 1 0 3 0 1 4 3 1 0 1 3 0 3 2 0 2 0 5 0 0 3 0 1 2 1 0 0

NA 6.32 (3.42) 3.13 (0.67) 8.87 (NA) 4.65 (NA) 4.70 (NA) NA 3.90 (1.07) NA 2.76 (NA) 3.38 (0.89) 5.71 (1.43) 3.59 (NA) NA 3.82 (NA) 4.75 (1.67) NA 5.92 (0.71) 1.37 (0.77) NA 5.32 (0.43) NA 2.90 (2.15) NA NA 3.41 (1.22) NA 2.70 (NA) 3.97 (0.38) 8.03 (NA) NA NA

NA 6.32 3.20 8.87 4.65 4.70 NA 3.45 NA 2.76 3.70 6.40 3.59 NA 3.82 5.56 NA 6.13 1.37 NA 5.32 NA 2.00 NA NA 3.03 NA 2.70 3.97 8.03 NA NA

ninety-three sixty-eight twenty-five 2006 2008

Note: Regarding each of the 17 Auslan numbers marked in bold, the median processing time for accurate interpretations of the number was markedly shorter than the median processing time for inaccurate interpretations of the number.

it invalid to use parametric tests such as the independent-samples t-test, (ii) a general loss of statistical power, and (iii) issues with confounding variables.2 Consequently, I decided to use descriptive statistics rather than the independentsamples t-test or the Mann-Whitney U Test for comparing processing time for accurate interpretations of a number with processing time for inaccurate interpretations of the number. As Table 5.6 shows, for each of the 17 Auslan numbers marked in bold (53% of all 32 Auslan numbers), the median processing time for accurate renditions of the number was considerably shorter than the median processing time for inaccurate

Processing Time 131 renditions of the number. For example, regarding eighty-eight in Auslan sentence 81, the median processing time for 18 accurate English interpretations of this number (2.24 seconds) was remarkably shorter than the median processing time for 2 inaccurate English interpretations of this number (5.32 seconds). This finding indicates that using short processing time probably contributes to accurate interpretations of numbers during simultaneous interpreting. For this study, what is regarded as short processing time? The median processing time for correct English renditions of these 32 Auslan numbers ranged from 1.41 seconds (forty-four in Auslan sentence 63) to 3.63 seconds (thirty-two in Auslan sentence 101), with the overwhelming majority (88%, or 28 out of 32) being shorter than 3 seconds. In contrast, the median processing time regarding inaccurate renditions of these 32 Auslan numbers varied from 1.37 seconds (eighty-eight in Auslan sentence 77) to 8.87 seconds (2006 in Auslan sentence 51), with the vast majority (80%, or 16 out of 20) being more than 3 seconds. These findings indicate that number processing time exceeding 3 seconds is highly likely to result in inaccurate renditions of numbers in signed-to-spoken language simultaneous interpreting when interpreters work alone and have no visual access to projected PowerPoint slides. In addition, for three numbers (eight times in Auslan sentence 55, eightyeight in Auslan sentence 77, and thirty-five in Auslan sentence 86), the median processing time for their correct interpretations was larger than the median processing time for their incorrect interpretations, indicating that processing time alone does not guarantee the accuracy of interpreting numbers. It is also worthy of attention that 7 (22%) of the 32 Auslan numbers were interpreted accurately by all 20 participants, indicating that some numbers are easier to interpret than others.

5.4 Onset Processing Time Versus Number Processing Time Regarding Auslan sentences whose numbers were accurately rendered in the target language, I compared interpreters’ onset processing time of the respective sentences with their processing time for the numbers in order to further investigate if interpreters shortened their processing time to render the numbers accurately. Table 5.7 shows the Mean and Median onset processing time for these Auslan sentences, the Mean and Median processing time for the numbers, and the results of statistical analyses of comparing onset processing time with number processing time. Paired-samples t-test (see t, p, and effect size) involving the Mean onset processing time and the Mean number processing time was carried out if both groups of data showed normal distribution (according to the results of Kolmogorov-Smirnov statistic). Wilcoxon Signed Rank Test (see z, p, and effect size) involving the Median onset processing time and the Median number processing time was conducted if one or both groups of data showed abnormal distribution. Interestingly, regarding Auslan sentences whose numbers were interpreted accurately, for each of the 19 (59% of the 32) Auslan numbers marked in bold in Table 5.7, onset processing time for the respective Auslan sentence was

132 Processing Time Table 5.7  Onset processing time versus number processing time Auslan Number sentence No.

N Onset processing time for Auslan sentences whose numbers were interpreted accurately Mean (SD)

28

ninety-

Results of comparing onset Processing time for the processing time numbers with number processing time Median Mean (SD) Median

20 4.87 (2.01) 4.34

1.95 (0.76) 1.82

17 3.56 (0.96) 3.43

2.74 (1.02) 2.77

17 4.08 (1.84) 3.63

1.85 (0.28) 1.90

three

30

sixtyeight

30

twentyfive

51

2006

19 2.56 (1.41) 1.97

2.34 (1.32) 1.90

53

2008

18 2.21 (0.83) 1.97

2.13 (1.10) 1.91

54

five

19 3.06 (1.98) 2.35

2.46 (0.82) 2.43

54

thirty

16 3.26 (2.46) 2.17

2.64 (2.09) 2.29

55

eight times 16 2.80 (0.81) 2.87

3.56 (1.22) 3.54

62

seventy-

18 6.00 (2.35) 5.70

3.68 (2.36) 2.73

18 7.58 (3.27) 8.54

2.35 (2.23) 1.41

64

under half 14 1.83 (0.48) 1.96

1.81 (0.60) 1.59

65

fifty-four

14 4.29 (2.19) 3.92

2.44 (1.04) 2.31

67

forty-

19 4.31 (1.95) 4.46

3.05 (2.75) 2.35

19 3.57 (2.21) 3.08

2.70 (1.78) 1.99

seven

63

fortyfour

four

69

ten

z = −3.88, p < 0.001, effect size = 0.61 (large) t(16) = 10.62, p < 0.001, effect size = 0.88 (large) z = −3.57, p < 0.001, effect size = 0.61 z = −-0.85, p > 0.05 z = −0.50, p > 0.05 z = −1.05, p > 0.05 z = −0.67, p > 0.05 t(15) = −3.29, p = 0.005, effect size = 0.42 (large) z = −3.03, p = 0.002, effect size = 0.51 (large) z = −3.38, p = 0.001, effect size = 0.56 (large) z = −1.16, p > 0.05 t(13) = 3.05, p = 0.009, effect size = 0.42 (large) z = −2.72, p = 0.006, effect size = 0.45 (large) z = −1.61, p > 0.05 (Continued)

Processing Time 133 Table 5.7  (Continued) Auslan Number sentence No.

N Onset processing time for Auslan sentences whose numbers were interpreted accurately Mean (SD)

Results of comparing onset Processing time for the processing time numbers with number processing time Median Mean (SD) Median

70

nineteen

18 3.30 (1.34) 2.83

2.33 (1.31) 1.76

71

nineteen

15 2.69 (1.35) 2.36

2.93 (1.59) 2.46

72

ten

20 3.26 (1.67) 2.90

2.50 (1.69) 1.89

73

nineteen

14 2.39 (0.64) 2.39

2.67 (0.88) 2.70

77

eighty-

18 3.64 (2.27) 2.84

1.93 (0.80) 1.69

eight

78

sixty-one

20 3.03 (1.57) 2.64

1.80 (0.59) 1.74

81

eighty-

18 3.91 (2.18) 3.41

3.02 (2.49) 2.24

20 4.37 (1.90) 4.06

2.55 (1.37) 2.33

86

thirty-five 14 4.63 (2.12) 4.31

3.21 (1.18) 3.45

87

thirty-one 20 4.18 (1.34) 3.90

2.13 (1.11) 1.80

88

five

20 3.45 (0.98) 3.35

2.59 (1.68) 2.46

96

eighty-

17 3.35 (2.20) 2.59

2.98 (2.25) 1.89

eight

85

twentythree

two

z = −2.94, p = 0.003, effect size = 0.49 (large) z = −0.31, p > 0.05 z = −3.25, p = 0.001, effect size = 0.51 (large) t(13) = −3.41, p = 0.005, effect size = 0.47 (large) z = −3.64, p < 0.001, effect size = 0.61 (large) z = −3.17, p = 0.002, effect size = 0.50 (large) z = −2.94, p = 0.003, effect size = 0.49 (large) t(19) = 3.07, p = 0.006, effect size = 0.33 (large) t(13) = 1.91, p > 0.05 t(19) = 5.65, p < 0.001, effect size = 0.63 (large) t(19) = 1.85, p > 0.05 z = −2.06, p = 0.039, effect size = 0.35 (moderate) (Continued)

134 Processing Time Table 5.7  (Continued) Auslan Number sentence No.

N Onset processing time for Auslan sentences whose numbers were interpreted accurately Mean (SD)

Results of comparing onset Processing time for the processing time numbers with number processing time Median Mean (SD) Median

97

sixty-two

20 3.12 (1.69) 2.59

2.00 (1.10) 1.63

98

forty-

19 4.24 (2.54) 3.57

2.07 (1.00) 1.80

three

99

thirty-one

17 4.21 (2.31) 4.34

2.66 (1.53) 2.23

100

thirty-two 17 4.75 (1.89) 5.03

3.55 (2.18) 3.23

101

sixty-two

19 5.29 (1.87) 5.43

3.25 (1.94) 2.29

101

thirty-two 18 4.98 (1.35) 5.32

4.41 (2.39) 3.63

z = −3.55, p < 0.001, effect size = 0.56 (large) t(18) = 3.98, p = 0.001, effect size = 0.47 (large) z = −2.72, p = 0.006, effect size = 0.47 (large) t(16) = 1.73, p > 0.05 z = −3.46, p = 0.001, effect size = 0.56 (large) t(17) = 1.19, p > 0.05

Note: N refers to the number of accurate English interpretations of the number in an Auslan sentence. All p values are two-tailed. For each of the 19 Auslan numbers marked in bold, onset processing time was significantly longer than number processing time.

significantly longer than processing time for the number (p < 0.05, consistently with large effect sizes indicating the considerable magnitude of the statistically significant differences). In other words, many interpreters used sufficiently long onset processing time when beginning to interpret these Auslan sentences, then shortened their processing time when dealing specifically with the numbers in these sentences. A closer examination of the 18 Auslan sentences that contained these 19 numbers revealed that many of these sentences featured signs such as how many (e.g., Auslan sentence 28) and what about (e.g., Auslan sentence 65) at the front that would cue interpreters that the Deaf presenter would soon mention a number. Alternatively, some of these Auslan sentences (e.g., Auslan sentence 63) were preceded by sentences that also contained numbers (e.g., reporting research findings); thus, the context served as a cue that the Deaf presenter might develop his ideas and mention another number. It is highly likely that many interpreters realised these cues and then reduced their processing time to wait for or cope with the imminent number.

Processing Time 135 In addition, for some Auslan sentences such as 85 and 87, the finding of onset processing time being significantly larger than number processing time appears to be related to syntactical differences between Auslan and English. Many of the interpreters used long onset processing time, presumably to receive more information from the Deaf presenter in order to understand the signed message sufficiently before starting to interpret, and uttered the English equivalent of the Auslan number as soon as it appeared in order to produce idiomatic English sentences by putting the English number before its referent and pertaining concept. For example, Bernie rendered Auslan sentence 87 those only total communication speak-sign altogether thirty-one into an idiomatic English sentence ‘There were 31 countries who reco [interpreter’s false start] who mentioned that they had total communication methods.’ With onset processing time of 5.96 seconds, Bernie started interpreting this Auslan sentence after he saw the number thirty-one and his number processing time was only 0.93 second. See Section 4.4 in Chapter 4 for Bernie’s reflection on shortening his processing time in certain places in an attempt to catch up with the Deaf presenter. Furthermore, in some cases, strategically omitting sentence initial signs such as them how many (see Auslan sentence 81: them how many? eighty-eight country have deaf schools) and starting the English interpretation with the number (e.g., Molly’s rendition ‘Eighty-eight countries said “Yes,” they did have Deaf schools.’) contributed to the onset processing time being much longer than the number processing time. Another interesting finding in Table 5.7 is that the onset processing time for two Auslan sentences (55 and 73) was significantly shorter than the number processing time. A common characteristic of these two Auslan sentences is that they started with a number. Some interpreters mirrored the Auslan sentence structure by either starting their English sentences with the number or putting the number at the front of their English sentences. For example, Tiffany transferred Auslan sentence 55 but have eight times mention mention mention sign language put-in put-in put-in successfully into English as ‘There are eight mentions of sign language throughout that document,’ keeping the number at the front of her English rendition. However, other interpreters restructured the Auslan sentence and moved the number to the middle or rear part of their English renditions. An example is that Amber rendered Auslan sentence 55 into English as ‘And within the document, sign language is mentioned eight times,’ with the number near the end of her English sentence. In other words, interpreters’ different ways of reproducing the understood Auslan messages in the target language led to their different onset processing time and number processing time. As shown in Table 5.7, for 11 Auslan sentences containing numbers, onset processing time was similar to number processing time (p > 0.05). Analysis of these Auslan sentences revealed that the Deaf presenter’s use of English-like syntactic structures (e.g., Auslan sentences 51, 53, 54, and 69) and some of the Auslan sentences (e.g., Auslan sentences 64 and 88 in Table 5.1) being quite short may have contributed to these findings. For instance, Annie rendered

136 Processing Time Auslan sentence 53 australia government finish sign r-a-t-i-f-i-e-d in 2008 as ‘The Australian government has actually ratified this document, and uh that happened in 2008.’ Unsurprisingly, due to the remarkable resemblance between this Auslan sentence and Annie’s English rendition in terms of syntactical structure, her onset processing time (2.33 seconds) was very similar to her number processing time (2.44 seconds) in terms of length. Taken together, onset processing time and number processing time appear to be influenced by a multitude of factors such as Auslan sentence structure and its similarity or lack thereof to English rendition, length of Auslan sentences, position of the number in the Auslan sentence, interpreters’ working memory capacity, interpreters’ use of strategies, interpreters’ varying ways of expressing the same source message in the target language, among others. Processing time in signed language simultaneous interpreting warrants further investigation.

5.5 Consequences of Excessively Short Onset Processing Time Aside from measuring different types of processing time for carrying out the aforementioned statistical analyses, I also conducted qualitative analyses of interpreters’ processing time in relation to the sentence-level accuracy of their simultaneous interpretation in an attempt to explain why serious interpretation errors (e.g., source language intrusion, omission of Auslan sentences) occurred. As expected, the qualitative analysis revealed that excessively short onset processing time severely affected some interpreters’ simultaneous interpreting performance. Overall, the results suggested that exceptionally short onset processing time caused the following four types of problems: (i) misunderstanding the source text, (ii) source language intrusion, (iii) inaccurate interpretation, (iv) literal interpretation, (v) ungrammatical or unidiomatic target language sentences, and (vi) frequent self-repairs or self-corrections resulting in disfluency in the target language speech. Example 5.1 illustrates that one interpreter’s (anonymity was intentionally maintained) excessively short onset processing time led to insufficient understanding of source messages, diluted the Deaf presenter’s emphasis, gave rise to source language intrusion, and resulted in an ungrammatical target language sentence, thus affecting both the interpreting process and interpreting product. In Example 5.1, the Deaf presenter used Auslan sentences 109, 110, and 111 to summarise his signed presentation, thus it is quite important that interpreters convey these messages accurately to the non-signing audience. The interpreter in Example 5.1 used fairly short onset processing time consistently to cope with Auslan sentences 109, 110, and 111, which contained key information at the end (the verb oppressed and negation no). With onset processing time of 2.5 seconds, the interpreter began interpreting Auslan sentence 109 when the Deaf presenter finished producing the subject of the sentence – many those deaf people. She rendered the initial meaning unit many those deaf people their lives as ‘So, Deaf people in their lives.’ The interpreter then transferred the second meaning unit can access t-o services as a question – ‘can they access services?’ She expressed the Deaf presenter’s emphasised signs

Processing Time 137

Example 5.1  Source language intrusion and ungrammaticality due to exceptionally short onset processing time

really oppressed as an answer (‘Not always.They are often oppressed.’).Although the interpreter accurately conveyed the Deaf presenter’s semantic content, she mitigated his end-weight emphasis by employing the question-and-answer syntactic structure and by not using intonation to convey the Deaf presenter’s pragmatic meaning. With onset processing time of 2.4 seconds, the interpreter started rendering Auslan sentence 110 after understanding its subject so those deaf people, which she transferred literally into English as ‘So, those Deaf people.’ Following the Deaf presenter closely, she changed the following verb-object syntactic structure – true enjoy enjoy enjoy equal human rights? – into an attributive clause ‘who really enjoy like having equal human rights,’ which resulted in distorting the original meaning. The interpreter paused for 0.7 second after seeing the sentence-final negating sign no that negates the aforementioned verb-object phrase, probably pondering how to cope with the syntactical difference between Auslan and English. She then resumed interpretation by saying ‘and that is not (1.1-second pause) out there (0.7-second pause) in total’; the pronoun ‘that’ causes ambiguity; the negation ‘not out there in total’ is inaccurate; and both pauses indicate that the interpreter found it difficult to produce an equivalent and grammatical English rendition.

138 Processing Time With onset processing time of 2.3 seconds, the interpreter began interpreting Auslan sentence 111 after seeing the signs so summarise that say deaf people and rendered them in English as ‘So, the summary that Deaf people.’ However, this interpretation is problematic because the Deaf presenter actually meant ‘In summary,’ and was about to make a conclusion for his whole presentation. Then the interpreter paused for 0.8 second to receive more source language information or to plan her target language rendition. She misinterpreted equal like other round world as a modifier for the subject ‘Deaf people’ – ‘getting equal access out there around the world’ – while the Deaf presenter meant whether Deaf people are equal to other people (hearing people) around the world. Soon after seeing his sentence-final negating sign no, the interpreter rendered it as ‘they are not!’ Due to her excessively short processing time when dealing with Auslan sentence 111, her English interpretation is ungrammatical, does not make sense, and is plagued with source language intrusion. The sentence-final negator no negates the meaning of the previous meaning unit equal like other round world. The interpreter needs to wait till the end of the Auslan sentence to start interpreting, or follow the question-and-answer structure and then provide a complete answer (e.g.,‘In summary, are Deaf people equal to other people around the world? No, they are not equal to their hearing counterparts yet’) in order to reflect the Deaf presenter’s emphasis through using end negation. The interpreter may use other coping strategies to render these Auslan sentences with end negation accurately. In summary, due to overly short onset processing time, this interpreter was processing the Auslan source segment at lexical item level or phrasal level, which prevented her from grasping the meaning or key point of Auslan sentences that ended with a negating sign. In other words, the interpreter started interpreting Auslan sentences 110 and 111 too early, without fathoming the Deaf presenter’s complete meaning unit or idea. As a result, her English interpretations are quite literal or inaccurate, disjointed, severely influenced by Auslan syntactical structures, inflicted with false starts and self-repairs, and ungrammatical in English. These findings corroborate interpretation errors reported by Cokely (1986, 1992) and Taylor (2002), who found that signed language interpreters’ overly short processing time affected their source language comprehension and target language production, which led to many interpretation errors such as unjustifiable omissions, distortions, and source language intrusion.

5.6 Corollaries of Exceptionally Long Onset Processing Time and Tail-to-tail Span The qualitative analysis of processing time in Auslan-to-English simultaneous interpreting revealed that excessively long onset processing time and tail-to-tail span led to the following problems: (i) unjustifiable omissions of important details or whole sentences, (ii) interpretation breakdowns, (iii) overly generalised or summarised interpretations that might be uninformative or misleading for the hearing audience, and (iv) inaccurate renditions comprised of fragmental information from adjacent sentences.

Processing Time 139 Example 5.2 illustrates that an interpreter’s (anonymity was intentionally maintained) overly long onset processing time coupled with a self-correction in interpreting the first Auslan sentence resulted in long tail-to-tail span for the first sentence, which caused a chain effect of problems such as unjustifiable omission of the second Auslan sentence, unjustifiable substitution of the core meaning of the third Auslan sentence, and unjustifiable omission of the fourth Auslan sentence. As mentioned earlier, these Auslan sentences are vital because they serve as the conclusion of the Deaf presenter’s presentation. With overly long onset processing time of 7.5 seconds, the interpreter in Example 5.2 began rendering Auslan sentence 108 when the Deaf presenter was producing the third sign of the subsequent Auslan sentence (i.e., 109), and interpreted the first sentence as an English statement ‘However, many people in the wider community are not aware of … Deaf people in general.’ As indicated by her 0.8-second pause, she had difficulties in recalling the object (deaf people) of the verb know, probably because she lagged quite far behind the Deaf presenter. She first said ‘interpreters,’ then corrected herself by saying ‘or peo [interpreter’s false start], or Deaf people in general.’ Her excessively long onset processing time, lengthy expressions (e.g., unjustifiable additions such as ‘however,’ ‘many,’ ‘wider,’ and ‘in general’), within-sentence pauses, and self-correction resulted in quite long tail-to-tail span (10.2 seconds). Her English interpretation of Auslan sentence 108 overlapped with the Deaf presenter’s Auslan sentence 109. By the time the interpreter finished this interpretation, the Deaf presenter had started producing Auslan sentence 110. She omitted Auslan sentence 109 probably due to her overly long tail-to-tail span (10.2 seconds) regarding Auslan sentence 108, cognitive overload, and time pressure. With onset processing time of 6.7 seconds, the interpreter began interpreting Auslan sentence 110 when the Deaf presenter was signing rights, the

Example 5.2  Unjustifiable omissions and substitutions due to excessively long onset processing time

140 Processing Time second-to-last sign of this sentence. She added a padding phrase, ‘my thought is,’ when he was signing the end negation no. Her subsequent 1-second pause suggests that she was planning her target language production. She rendered the subject and verb correctly as ‘Deaf people do not’ and paused for 2.3 seconds to process the received information. She then rendered the verb phrase true enjoy enjoy enjoy effectively as ‘equally enjoy.’ However, she replaced the key point human rights with ‘freedom, dignity, and access to (0.5-second pause) the wider world and community.’ By the time she finished her English interpretation of Auslan sentence 110, the Deaf presenter had finished producing Auslan sentence 111 and had embarked on producing the first sign of Auslan sentence 112. This interpreter omitted Auslan sentence 111 primarily due to her long onset processing time for Auslan sentence 110 and her lengthy English rendition (e.g., addition of ‘my thought is’ and verbose object) of this Auslan sentence. In summary, due to her exceptionally long onset processing time, self-repair, verbose target language expressions, and rather long tail-to-tail span, the interpreter in Example 5.2 omitted two Auslan sentences (109 and 111) and distorted the meaning of another Auslan sentence (110). Her English interpretation is hit-and-miss: she interpreted only some details of the first Auslan sentence correctly but missed the second Auslan sentence; she produced an inaccurate rendition of the third Auslan sentence and omitted the fourth Auslan sentence. Thus, her onset processing time (i.e., almost as long as an entire Auslan sentence) was overly long and taxed her working memory capacity. As a result, the excessively long onset processing time severely affected both her cognitive processing and interpreting output. These findings coincide with Lee’s (2002) study showing that exceptionally long onset processing time of the current source language sentence affects both the cognitive processing and accuracy of interpreting the subsequent source language sentence(s).This example also highlights the importance for interpreters to be aware of and adjust their processing time, as well as express messages concisely in the target language (see also Gile, 2009).

5.7 Interrelation Between Processing Time and Interpreting Strategies The qualitative analysis of processing time in Auslan-to-English simultaneous interpreting provided evidence that interpreters’ processing time is closely related to their interpreting strategies. Example 5.3 illustrates that as Emily’s interpreting strategies changed from one sentence to another her processing time varied. The Deaf presenter said that his research team received a total of 93 responses to their (questionnaire) survey (Auslan sentence 28), that not all of the 93 respondents were members of the World Federation of the Deaf (WFD) (Auslan sentence 29), that 68 of the responses were from member countries of the WFD and the other 25 responses were from non-member countries of WFD (Auslan sentence 30), and that this meant those 25 countries did not have national associations of the Deaf (Auslan sentence 31). Overall, Emily’s onset processing time became shorter and shorter primarily because of

Processing Time 141

Example 5.3  Emily’s onset processing time changed according to interpreting strategies

her employment of different interpreting strategies and use of pronouns that enabled her to catch up with the Deaf presenter’s lecture. As for Auslan sentence 28, Emily waited for 9 seconds before starting to interpret it probably because she predicted from the Deaf presenter’s initial sign how many that he would produce a number later and she was aware that the topiccomment Auslan structure required her to restructure the sentence. She started interpreting this Auslan sentence when he was producing the sentence-final sign – ninety-three – and she produced a short and idiomatic English sentence (‘We received a total of ninety-three survey responses’), which shortened her lag. Starting out with the long onset processing time of 9 seconds, she then followed the Deaf presenter closer in order to catch the number ninety-three, the processing time for which was 1.8 seconds. Given that Auslan sentence 29 featured end negation, Emily did not start speaking until she saw the negating sign no at the end; thus, her onset processing time was 5.1 seconds. As a result, she produced an accurate and idiomatic interpretation for Auslan sentence 29. Note that her strategy of using the pronoun phrase ‘all of those survey responses’ for the repeated number ninety-three (which was excluded from number processing time analysis due to the repetition; see Table 5.1) was effective as it avoided redundancy, enhanced coherence, and saved her some attentional resources. Her use of the abbreviation ‘WFD’ for the Deaf presenter’s fingerspelling w-f-d also shortened her lag. With onset processing time of 3.8 seconds, Emily started interpreting Auslan sentence 30 after she understood its first meaning unit they group sixty-eight member w-f-d, which she reformulated as an elliptical clause (‘Sixty-eight of them were,’). Again, the use of the pronoun ‘them’ improved coherence, and the

142 Processing Time concise expression reduced her lag. Emily’s omission of member w-f-d can be considered as a strategic omission because it contributes to the succinctness of the target text and does not cause a loss of meaning. Her processing time for the number sixty-eight was 2.9 seconds, which is shorter than her onset processing time (3.8 seconds). Then she paused for 1.9 seconds to wait for the other number cued by the signs other group those. Once she understood the awaited information twenty-five, she transferred the number and strategically added its referent (‘and twenty-five responses’). She unpacked the Deaf presenter’s concisely expressed concept not member w-f-d into a comprehensible but long English rendition ‘were from countries that were not member countries,’ and she spoke quite fast in order to keep her processing time manageable. Emily’s processing time for the number twenty-five was 1.9 seconds, which is shorter than both her aforementioned onset processing time and number processing time for sixty-eight. These findings indicate that the interpreter predicted that numbers would be produced by the Deaf presenter and thus reduced her processing time in order to interpret the numbers accurately. With onset processing time of 3.9 seconds for Auslan sentence 31, Emily again used the pronoun ‘they’ to represent the Deaf presenter’s repeated number twenty-five (which was also excluded from number processing time analysis because it was acceptable to omit or paraphrase the repeated number; see Table  5.1) in order to catch up and save attentional resources. She faithfully conveyed the meaning of this Auslan sentence by using a succinct English rendition, ‘which means they didn’t have national associations of the Deaf.’ In summary, Emily used rather long onset processing time for Auslan sentence 28, presumably to cope with the syntactical difference between Auslan and English (external factor) and to produce a natural and accurate interpretation. She strategically used pronouns to deal with the Deaf presenter’s repeated numbers in order to buy time, reduce cognitive load, and avoid redundancy in her interpretation. This strategy reduced her processing time, saved her processing capacity which could be allocated to Signed Language Comprehension Effort and Working Memory Effort. In addition, Emily paused to receive more signed information, giving her time to unpack economically expressed signed concepts and produce grammatical and comprehensible English renditions, thus increasing her processing time. However, she then spoke faster, used an elliptical sentence, made a strategic omission, and used a pronoun in order to catch up with the Deaf presenter. Taken together, this interpreter’s processing time fluctuated according to the nature of the source material, the interpreting strategies that she applied, her own delivery rate, and her target language sentence structure.

5.8 Discussion This chapter has reported on results from quantitative analyses of the 20 professional Auslan/English interpreters’ processing time, specifically focusing on their onset processing time regarding 40 Auslan sentences containing numbers or negation and their processing time regarding 32 Auslan numbers. The results

Processing Time 143 revealed that the overall mean onset processing time for the 20 interpreters was 4.12 seconds (SD = 0.88) and the overall median onset processing time 3.9 seconds. The individual interpreters’ mean onset processing time varied from 2.43 seconds (Lauren) to 6.36 seconds (Liz).These findings are largely consistent with most results in the existing literature on processing time in spoken and signed language simultaneous interpreting (previous studies found an average processing time of 1–5 seconds and processing time in the existing literature ranged from −6.27 seconds to 19.17 seconds) (Anderson, 1994; Barik, 1973; Christoffels & De Groot, 2004; Cokely, 1986; Collard & Defrancq, 2019; Defrancq, 2015; DíazGalaz, Padilla, & Bajo, 2015; Gerver, 1969/2002; Kim, 2005; Lee, 2002, 2006, 2011; Oléron & Nanpon, 1965/2002; Paneth, 1957/2002; Podhajská, 2008; Timarová et al., 2011; Treisman, 1965). The corroboration of results is probably because the mean is a summative statistic, conceals individual differences, and is influenced by extreme values. Due to the high variability in processing time both between subjects and within subjects, it is possible that mean processing time is not very useful or informative in the analysis of cognitive aspects of spoken and signed language simultaneous interpreting. It is important to bear in mind that mean processing time is an average of various degrees of temporal delay throughout a source speech, and that processing time at a particular point in time depends on many external and internal factors (Napier, McKee, & Goswell, 2010). The length of interpreters’ onset processing time in this study ranged from 0.56 to 15.40 seconds, which is within the processing time range (−6.27 seconds to 19.17 seconds) reported in the previous studies (see Collard & Defrancq, 2019; Defrancq, 2015, p. 31; Lamberger-Felber, 2001; McKee & Napier, 2002, p. 42; Podhajská, 2008, pp. 103–104). This consistency in the upper limit of processing time may reflect the limits of human memory. Although interpreters may remember a large amount of source text information based on meaning, due to multitasking and time pressure in simultaneous interpreting, they cannot lag extremely far behind the speaker/signer and still produce highly accurate interpretations. An interpreter may lag quite far behind the Deaf presenter, but his/her interpreting performance may be poor due to cognitive overload. It could be revealing to conduct further research to investigate the accuracy of interpreting source language sentences with quite long onset processing time (e.g., over 6 seconds). Unlike previous research (e.g., Collard & Defrancq, 2019; Defrancq, 2015), the current study did not find negative onset processing time for the 40 Auslan sentences among the 20 professional interpreters. It could be that most of the 40 Auslan sentences were complex and featured numbers or negation and some of them were syntactically asymmetrical to English sentences and therefore had low predictability for the interpreters. Other possible explanations include that the interpreters did not have visual access to the Deaf presenter’s PowerPoint slides during the Auslan-to-English simultaneous interpreting task, and that most of the interpreters reported that they were unfamiliar with the Deaf presenter and his presentation content (see Chapter 4); these factors may have affected the interpreters’ confidence and ability to employ anticipation as a strategy.

144 Processing Time In addition, my research found that the 20 professional Auslan/English interpreters’ median onset processing time was significantly shorter than their mean onset processing time, which corroborates Timarová et al.’s (2011) finding regarding spoken language interpreters. Both findings indicate that researchers need to be cautious about methodological issues related to both the measurement and statistical analysis of processing time in simultaneous interpreting. It is highly recommended that researchers should report the mean, standard deviation, median, and range of processing time (Timarová et al., 2011), use both mean and median processing time for statistical analysis and see if the result patterns are the same, and compare their own results with previous findings to see if there is consistency. Moreover, this study found high variability in the professional interpreters’ onset processing time and number processing time during the Auslan-to-English simultaneous interpreting task. Specifically, onset processing time varied considerably from one interpreter to another. Moreover, an individual interpreter’s processing time varied remarkably from one measurement point to another within the interpreting task. These results corroborate previous findings (Timarová et al., 2011; Wang, 2020) that spoken and signed language interpreters do not maintain a constant, fixed processing time throughout a simultaneous interpreting task but move on a continuum. Some interpreters’ onset processing time for some Auslan sentences were similar in length but their onset processing time for other Auslan sentences were markedly different. Taken together, the high variability in processing time during simultaneous interpreting indicates that processing time is driven by a complex web of external factors (e.g., the difficulty level of incoming chunks of information, syntactical differences between the source language and the target language, information density) and internal factors (e.g., interpreters’ familiarity with the topic and terminology, working memory capacity, interpreting strategies). A key finding from the analysis of the 20 professional interpreters’ English renditions of the aforementioned 32 Auslan numbers is that the mean accuracy rate for the individual numbers was 89%. This finding of high accuracy rate of number interpretations is in line with previous findings regarding professional spoken language interpreters’ accuracy rate of rendering numbers in simultaneous interpreting (e.g., over 90% in Korpal & Stachowiak-Szymczak, 2020 and 79% in Collard & Defrancq, 2021). This finding indicates that professional signed language interpreters coped with numbers in the signed-to-spoken language simultaneous interpreting task quite well. The high level of accuracy rate of the professional interpreters’ number renditions probably has resulted from the fact that the 32 numbers in the Auslan source text consisted of 30 small, simple whole numbers (typically smaller than 100) and two years (2006 and 2008). Prior studies (e.g., Mazza, 2001; Pinochi, 2009) have revealed that simple numbers with fewer than four digits (e.g., 87) and dates (e.g., 2012) were interpreted remarkably more accurately than large numbers with four or more digits (e.g., 65400) and decimals (e.g., 28.3). The professional interpreters’ use of particular strategies to deal with numbers, use of the printed PowerPoint slides, and the Deaf presenter’s generally clear and slow production of the numbers may

Processing Time 145 have also contributed to the high success rate of number interpretation in the current study. Another related finding is that despite the high accuracy rate of interpreting individual Auslan numbers (89%), the accuracy rate of rendering a series of numbers in neighbouring Auslan sentences was substantially lower (65%). This result echoes Mazza’s (2001) finding and Gile’s (2009) argument that numbers not only constitute a source of difficulty for interpreters in simultaneous interpreting but also affect the cognitive processing of their surrounding information in the source text. This finding indicates that numbers may easily overload working memory and require interpreters to allocate their attentional resources appropriately to the concurrent sub-processes of simultaneous interpreting (e.g., retaining a number that was just produced by a speaker/signer, planning the target language production of the number and its related information, and understanding the incoming message). Adjacent sentences each containing a number are cognitively more demanding than a single sentence featuring a number. The results of this study revealed that substitutions (N = 34), omissions (N = 29), and generalisations (N = 8) were the most frequent types of interpretation errors regarding numbers.That is, when the omissions and generalisations are combined, they (37) exceed the substitutions (34). It is important to acknowledge that there are differences in error taxonomy between some previous studies and the current study. For example, for interpretation errors regarding numbers that were categorised as ‘omissions’ in three previous studies (Braun & Clarici, 1996; Mazza, 2001; Pinochi, 2009), I divided such errors into two distinct types: omissions and generalisations. In addition, for the lexical errors, syntactical errors, errors of phonemic perception, and errors of transposition regarding numbers in the previous studies, I coded all of them as substitutions. Despite the differences in error taxonomy, my findings corroborate the results of these previous studies that omissions (i.e., omissions and generalisations in the present study) proved to be the most frequent interpretation error of numbers. Regarding numbers, in the current study the number of omissions is similar to the number of substitutions (29 vs. 34), which contradicts Collard and Defrancq’s (2021) finding that omissions were considerably more frequent than substitutions. Sample size of both interpreters and interpretation data corpus may account for these discrepancies in the findings of different studies. More professional interpreters, numbers, and simultaneous interpretation data are needed in future studies in order to analyse the success rate of interpreting numbers, frequency of different types of number errors, and processing time for numbers in more detail. A crucial finding of the present study is that for each of 17 (53% of 32) Auslan numbers, the median processing time for accurate interpretations of the number (typically shorter than 3 seconds) was notably shorter than the median processing time for inaccurate interpretations of the number (typically longer than 3 seconds). Although the uneven samples made it impossible to carry out statistical tests to prove whether such differences reached statistical significance, the between-group differences were indeed considerable. This finding corroborates

146 Processing Time Collard and Defrancq’s (2021) finding that processing time for the correct renditions of numbers was significantly shorter than processing time for the erroneous renditions of the numbers. This finding also coincides with Timarová et al.’s (2014) research revealing a significant negative correlation between professional interpreters’ median onset processing time and their accuracy of interpreting numbers during spoken language simultaneous interpreting, which indicates that shorter onset processing time is related to higher accuracy of number interpretation. My finding indicates that professional interpreters’ use of short processing time for numbers contributes to the high accuracy rate of interpreting the numbers, and that long processing time for numbers increases the risk of rendering the numbers wrongly due to working memory overload (see also Taylor, 2002). My finding that the median processing time for inaccurate renditions of numbers being typically longer than 3 seconds supports Setton’s (1999) observation that errors often occur when interpreters’ processing time is longer than 3–4 seconds. It is also worth noting that the result pattern on number processing time and accuracy of number interpretation was complex rather than definitive. For each of three Auslan numbers, the median processing time for accurate renditions of the number was longer than the median processing time for inaccurate renditions of the number. This finding indicates that number processing time (short, medium, or long) is not the only factor that determines the accuracy of number renditions. Indeed, other factors such as miscomprehension of a number (possibly due to the Auslan source text being a two-dimensional video rather than a live presentation), momentary attention deficit, poor working memory, and ineffective coordination of the three key Efforts of simultaneous interpreting (source speech Comprehension, Working Memory, Target Speech Production) may jeopardise the interpretation of numbers in simultaneous interpreting. More importantly, the results revealed that, regarding Auslan sentences whose numbers were rendered accurately, for each of 19 (59% of 32) Auslan numbers, onset processing time for the respective Auslan sentence was significantly longer than processing time for the number. The effect sizes for these results were typically very large, indicating that the differences were salient. These findings corroborate Timarová et al.’s (2011) results that a professional spoken language interpreter’s onset processing time measured on sentence beginnings was significantly longer than his/her processing time measured on numbers in simultaneous interpreting. These results indicate that it makes a difference at which juncture in the source text processing time was measured. Given that sentence beginnings usually differ from numbers in terms of linguistic features, these results suggest that source messages that are different in terms of linguistic properties may require different types of cognitive processing during simultaneous interpreting (Timarová et al., 2011). These findings lend support to many interpreter educators’ advice for interpreting students and many professional interpreters’ frequently applied strategy (Collard & Defrancq, 2021; Napier et al., 2010; Setton & Dawrant, 2016; Taylor, 2002) that spoken and signed language interpreters should shorten their processing time as necessary in order to render

Processing Time 147 frozen items such as numbers and names accurately during simultaneous interpreting. These findings indicate that interpreting students need to be trained on adjusting their processing time during simultaneous interpreting in order to cope with numbers and other memory-demanding items. The quantitative and qualitative analyses of processing time in signed-to-spoken language simultaneous interpreting indicate that processing time is a sensitive variable that not only sheds light on the invisible and complex cognitive processes involved in simultaneous interpreting, but also provides insights into the intricate relationship among interpreting process, strategies, and interpreting performance. In addition to the quantitative results, the qualitative analysis of processing time also yielded interesting results. As illustrated by a representative example (Example 5.1) regarding end negation in the Auslan-to-English simultaneous interpretation data, exceptionally short onset processing time (producing target language output without properly analysing sufficient source language information) often resulted in miscomprehension, source language intrusion, inaccurate interpretation, literal interpretation, ungrammaticality of target language output, and frequent self-repair or self-correction.This finding is in line with previous results (Cokely, 1986, 1992; Wang, 2020) that show that interpreters’ excessively short onset processing time affects their analysis of source messages and target language reformulation, thus affecting both the simultaneous interpreting process and the quality of simultaneous interpreting performance. In addition, Example 5.2 of an interpreter’s use of exceptionally long onset processing time to cope with Auslan sentences featuring end negation illustrated that lagging excessively far behind the Deaf presenter led to unjustifiable omissions of whole sentences and unjustifiable distortion of meaning, which is in line with previous findings (Cokely, 1986, 1992; Lamberger-Felber, 2001; Wang, 2020). These findings suggest that at any moment during simultaneous interpreting, interpreters need to strike a delicate balance between waiting for sufficiently long in order to adequately understand a chunk of source language information and processing information fast enough in order to avoid overloading working memory (see also Gile, 2009;Timarová et al., 2014;Wang, 2020). Example 5.3 of successful interpretations of adjacent Auslan sentences containing numbers has illustrated that the interpreter started out with very long onset processing time to cope with syntactical asymmetry between Auslan and English, then gradually shortened her onset processing time for the subsequent Auslan sentences through the use of strategies (e.g., using pronouns rather than repeating numbers and referents, strategically omitting information, using concise target language expressions) in order to produce accurate, effective, and coherent interpretations. This finding indicates that processing time is closely associated with processing demands posed by the source text, interpreters’ own strategies, and interpreters’ source and target language competence, among other factors. The qualitative analysis of this example has also unveiled how an interpreter’s processing time varies from one sentence to another in the process of producing accurate renditions. Taken together, the results from this chapter indicate that interpreters need to adjust processing time (e.g., lengthening processing time when dealing with

148 Processing Time syntactical differences between the source language and the target language, shortening processing time when coping with numbers and names) according to the complexity of incoming chunks of information, syntactical (dis)similarity between the source language and the target language, the s­peaker’s/ signer’s delivery rate, the interpreters’ own working memory capacity, interpreting strategies, availability of target language expressions, automaticity of interpreting processes in relation to experience and expertise, among other factors (see also Gile, 2009; Setton & Dawrant, 2016;Timarová et al., 2014;Wang, 2020). While previous studies on processing time in simultaneous interpreting have focused on quantitative analysis, the present study has revealed that qualitative analysis of processing time, interpreting strategies, and accuracy yields a deeper understanding of how processing time and many other factors interact with each other and of how an interpreter’s working memory functions during simultaneous interpreting (e.g., at a specific point in time, how much information an interpreter is memorising while processing other information). Although processing time plays an important role in simultaneous interpreting, it is also important to emphasise that processing time is not the only key to guaranteeing high quality of simultaneous interpreting performance, as interpreters’ bilingual proficiency, interpreting experience, formal interpreting training, pre-assignment preparation, familiarity with the topic and the speaker/Deaf signer, ability to make use of pauses to chunk source language information into meaningful target language renditions, and other skills also contribute to effective simultaneous interpretations.

5.9 Summary This chapter has reported on findings from quantitative analysis of the 20 professional Auslan/English interpreters’ onset processing time regarding 40 Auslan sentences containing numbers or negation as well as their processing time for 32 Auslan numbers. It was found that the interpreters’ processing time varied considerably from person to person and from one measurement point to another in the Auslan-to-English simultaneous interpreting task. Overall, despite high accuracy rate (89%) of rendering individual numbers (simple numbers smaller than 100), the accuracy of interpreting a series of numbers in adjacent Auslan sentences that each contained a number proved to be much lower (65%), indicating that numbers did cause difficulties (e.g., attention allocation, working memory, coordination) for the professional interpreters. In addition, qualitative analysis of typical examples from the Auslan-toEnglish simultaneous interpretation data showed that both excessively short onset processing time and overly long onset processing time hindered simultaneous interpreting processes and affected simultaneous interpreting performance; therefore, interpreters’ onset processing time should not be exceptionally long or short. The qualitative analysis further revealed that interpreters’ onset processing time varied from one sentence to another in accordance with particular demands posed by the source language information and the interpreters’

Processing Time 149 interpreting strategies, among other factors. Given that this chapter focused on quantitative analyses of the interpreters’ processing time in regard to individual numbers (lexical level) and single sentences (sentence level), the following two chapters will extend the data analysis by conducting qualitative analysis of the interpreters’ processing time and coping strategies in rendering adjacent Auslan sentences (context level) containing numbers and end negation, respectively.

Notes 1 There were another 17 numbers in the Auslan source text, but they constituted secondary information and could be generalised or strategically omitted without causing significant loss of meaning. Therefore, they were excluded from the analysis of numbers throughout Chapters 5 and 6. 2 For further information, please refer to https://www.statisticshowto.datasciencecentral.com/unequal-sample-sizes/.

References Anderson, L. (1994). Simultaneous interpretation: Contextual and translation aspects. In S. Lambert & B. Moser-Mercer (Eds.), Bridging the Gap: Empirical Research in Simultaneous Interpretation (pp. 101–120). Amsterdam/Philadelphia: John Benjamins. Barik, H. C. (1973). Simultaneous interpretation:Temporal and quantitative data. Language and Speech, 16(3), 237–270. doi: 10.1177/002383097301600307. Braun, S., & Clarici, A. (1996). Inaccuracy for numerals in simultaneous interpretation: Neurolinguistic and neuropsychological perspectives. The Interpreters’ Newsletter, 7, 85–102. Christoffels, I. K., & De Groot, A. M. B. (2004). Components of simultaneous interpreting: Comparing interpreting with shadowing and paraphrasing. Bilingualism: Language and Cognition, 7(3), 227–240. doi: 10.1017/S1366728904001609. Cokely, D. (1986). The effects of lag time on interpreter errors. Sign Language Studies, 53(winter), 341–375. doi: 10.1353/sls.1986.0025. Cokely, D. (1992). Interpretation: A Sociolinguistic Model. Burtonsville, MD: Linstok Press. Collard, C., & Defrancq, B. (2019). Predictors of ear-voice span, a corpus-based study with special reference to sex.Perspectives,27(3),431–454.doi:10.1080/0907676X.2018.1553199. Collard, C., & Defrancq, B. (2021). Corpus study on the rendition of numbers in simultaneous interpreting, with special reference to sex. Interpreting. Defrancq, B. (2015). Corpus-based research into the presumed effects of short EVS. Interpreting, 17(1), 26–45. doi: 10.1075/intp.17.1.02def. Díaz-Galaz, S., Padilla, P., & Bajo, M. T. (2015). The role of advance preparation in simultaneous interpreting: A comparison of professional interpreters and interpreting students. Interpreting, 17(1), 1–25. doi: 10.1075/intp.17.1.01dia. Gerver, D. (1969/2002). The effects of source language presentation rate on the performance of simultaneous conference interpreters. In F. Pöchhacker & M. Shlesinger (Eds.), The Interpreting Studies Reader (pp. 52–66). London/New York: Routledge. Gile, D. (2009). Basic Concepts and Models for Interpreter and Translator Training (Rev. ed.). Amsterdam/Philadelphia: John Benjamins. Jones, R. (2002). Conference Interpreting Explained (2nd ed.). Manchester, UK: St. Jerome.

150 Processing Time Kim, H.-R. (2005). Linguistic characteristics and interpretation strategy based on EVS analysis of Korean-Chinese, Korean-Japanese interpretation. Meta, 50(4). http:// id.erudit.org/iderudit/019846ar. Korpal, P., & Stachowiak-Szymczak, K. (2020). Combined problem triggers in simultaneous interpreting: Exploring the effect of delivery rate on processing and rendering numbers. Perspectives, 28(1), 126–143. doi: 10.1080/0907676X.2019.1628285. Lamberger-Felber, H. (2001). Text-oriented research into interpreting. Examples from a case study. Hermes, Journal of Linguistics, 26, 39–64. doi: 10.7146/hjlcb.v14i26.25638. Lee, T. (2002). Ear voice span in English into Korean simultaneous interpretation. Meta, 47(4), 596–606. doi: 10.7202/008039ar. Lee, T. (2006). A comparison of simultaneous interpretation and delayed simultaneous interpretation from English into Korean. Meta, 51(2), 202–214. doi: 10.7202/013251ar. Lee, T. (2011). English into Korean simultaneous interpretation of academy awards ceremony through open captions on TV. Meta, 56(1), 145–161. doi: 10.7202/1003514ar. Mazza, C. (2001). Numbers in simultaneous interpretation. The Interpreters’ Newsletter, 11, 87–104. McKee, R., & Napier, J. (2002). Interpreting into International Sign Pidgin: An analysis. Journal of Sign Language and Linguistics, 5(1), 27–54. Napier, J., McKee, R., & Goswell, D. (2010). Sign Language Interpreting:Theory and Practice in Australia and New Zealand (2nd ed.). Sydney, Australia: The Federation Press. Nicodemus, B. (2009). Prosodic Markers and Utterance Boundaries in American Sign Language Interpretation. Washington, DC: Gallaudet University Press. Nilsson, A.-L. (2016). Interpreting from signed language into spoken language: The skills and knowledge needed to succeed. In A. Kalata-Zawłocka & B. van den Bogaerde (Eds.), To Say or Not to Say – Challenges of Interpreting from Sign Language to Spoken Language. Proceedings of the 23rd efsli Conference in Warsaw, Poland, 11th–13th September 2015 (pp. 15–48). European Forum of Sign Language Interpreters (efsli). Oléron, P., & Nanpon, H. (1965/2002). Research into simultaneous translation. In F. Pöchhacker & M. Shlesinger (Eds.), The Interpreting Studies Reader (pp. 42–50). London/ New York: Routledge. Paneth, E. (1957/2002). An investigation into conference interpreting. In F. Pöchhacker & M. Shlesinger (Eds.), The Interpreting Studies Reader (pp. 30–40). London/New York: Routledge. Pinochi, D. (2009). Simultaneous interpretation of numbers: Comparing German and English to Italian. An experimental study. The Interpreters’ Newsletter, 14, 33–57. Podhajská, K. (2008). Time lag in simultaneous interpretation from English into Czech and its dependence on text type. Folia Translatologica, 10, 87–110. Setton, R. (1999). Simultaneous Interpretation: A Cognitive-Pragmatic Analysis. Amsterdam/ Philadelphia: John Benjamins. Setton, R., & Dawrant, A. (2016). Conference Interpreting: A Complete Course. Amsterdam: John Benjamins. Taylor, M. M. (2002). Interpretation Skills: American Sign Language to English. Edmonton, Alberta: Interpreting Consolidated. Timarová, Š., Cenková, I., Meylaerts, R., Hertog, E., Szmalec, A., & Duyck, W. (2015). Simultaneous interpreting and working memory capacity. In A. Ferreira & J. W. Schwieter (Eds.), Psycholinguistic and Cognitive Inquiries into Translation and Interpreting (pp. 101–126). Philadelphia: John Benjamins.

Processing Time 151 Timarová, Š., Čeňková, I., Meylaerts, R., Hertog, E., Szmalec, A., & Duyck, W. (2014). Simultaneous interpreting and working memory executive control. Interpreting, 16(2), 139–168. doi: 10.1075/intp.16.2.01tim. Timarová, Š., Dragsted, B., & Hansen, I. G. (2011). Time lag in translation and interpreting: A methodological exploration. In C. Alvstad, A. Hild, & E. Tiselius (Eds.), Methods and Strategies of Process Research: Integrative Approaches in Translation Studies (pp. 121–146). Amsterdam: John Benjamins. Treisman, A. M. (1965). The effects of redundancy and familiarity on translating and repeating back a foreign and a native language. British Journal of Psychology, 56(4), 369– 379. doi: 10.1111/j.2044-8295.1965.tb00979.x. Wang, J. (2020). Striking a cognitive balance: Processing time in Auslan-to-English simultaneous interpreting. In D. Hunt & E. Shaw (Eds.), The Second International Symposium on Signed Language Interpretation and Translation Research: Selected Papers (pp. 108–131). Washington, DC: Gallaudet University Press.

6 Cognitive Overload and Coping Strategies Regarding Numbers

This chapter aims to draw on the Auslan-to-English simultaneous interpretation data to answer the second and third research questions of the study, which are listed as follows: 1 What causes cognitive overload in simultaneous interpreting from a signed language into a spoken language? 2 What strategies do signed language interpreters employ to cope with the challenges and cognitive overload in simultaneous interpreting from a signed language into a spoken language? The previous chapter has reported on findings from quantitative analysis of processing time and number accuracy, but the current chapter presents results from qualitative analysis of sentence-level accuracy, cognitive overload, and coping strategies in relation to interpreting adjacent Auslan sentences containing numbers. While some numbers appeared in the middle of Auslan sentences, others appeared near or at the end of Auslan sentences. It is noteworthy that some Auslan sentences under examination included multiple interpreting difficulties such as numbers, dense information, and syntactic differences between Auslan and English. Moreover, when such challenging Auslan sentences occurred one after another and formed a cluster, the source speech segment posed tremendous difficulties for interpreters. The data analysis of this chapter focused on the four primary factors that many professional Auslan/English interpreters in the present study reported to have caused cognitive overload in the Auslan-to-English simultaneous interpreting task (see Chapter 4), namely (i) numbers in the source text, (ii) the interpreters’ use of long processing time, (iii) dense information in the source text, and (iv) syntactical differences between Auslan and English. These factors are often interwoven in adjacent Auslan sentences that contain numbers near or at the end of the sentences. More importantly, this chapter reports on strategies adopted by the interpreters in order to cope with these challenges when working from Auslan into English, thus highlighting what they did well rather than what they did wrong. An innovative feature of this chapter is that it uses an adapted version of Gile’s  (2009) Effort Model of spoken language simultaneous interpreting

Cognitive Overload and Coping Strategies 153 (SLI  =  Signed Language Comprehension + Target Speech Production + Working Memory + Coordination; see Chapter 2) as a theoretical framework to guide local analysis of representative examples selected from the experimental corpus of professional Auslan/English interpreters’ Auslan-to-English simultaneous interpretations. Gile (2008, p. 59) defined local analysis as ‘analysis focused on short segments and sequences of two or three adjacent segments as opposed to general features of speeches’ and maintained that local analysis ‘may provide explanations which overall analysis [of whole speeches] fails to uncover.’ In other words, sentences were used as the units of local analysis. The prominent merit of local analysis lies in its examination of how the interpreter’s unfinished processing of the previous sentence impacts on his/her processing of the current sentence (imported cognitive load), the current cognitive load (cognitive load regarding the processing of the current sentence), and how the processing of the current sentence influences the processing of the following sentence (exported cognitive load) (Gile, 2008). Local analysis not only recognises concurrent processing (e.g., target language production of a source language sentence ending may overlap with comprehension of the beginning of the subsequent source language sentence) but also scrutinises the ‘domino effect’ of cognitive processing (i.e., how the processing of the current sentence influences the processing of the next sentence). Therefore, local analysis captures the dynamic nature, multitasking, and continuous information flow of simultaneous interpreting. Drawing on the aforementioned adapted version of Gile’s (2009) Effort Model of spoken language simultaneous interpreting, local analysis of Auslanto-English simultaneous interpretation data in this chapter not only reveals how cognitive overload occurred or was prevented in the first place, but also sheds light on the complex and interesting interrelation between the largely invisible interpreting process and the observable interpreting product.

6.1 Cognitive Overload and Sentence-level Accuracy Regarding Numbers The Deaf professional incorporated 32 Auslan numbers in 29 Auslan sentences (see Table 5.1 in Chapter 5) at various points throughout his Auslan presentation to report key findings from the international survey study on Deaf People and Human Rights that he had facilitated. The majority of the Auslan numbers were reported in the second half of his presentation and were presented in the form of bar charts or tables on his PowerPoint slides. As noted before, the Deaf professional could access the PowerPoint slides directly while delivering his presentation, but the interpreters could not see the slides on the computer screen while interpreting the Auslan video, although if they wished, they could refer to a hardcopy of the slides. If the interpreters could see the PowerPoint slides on the screen, had a team interpreter, or coordinated the interpreting process with the Deaf presenter through look-pause-nod cooperation strategies as in authentic conference interpreting situations, they may have achieved much higher accuracy regarding both the Auslan numbers and the Auslan sentences containing the numbers.

154 Cognitive Overload and Coping Strategies Given that these Auslan numbers provided precise information about the Deaf presenter’s research findings, unjustifiable omission, unjustifiable distortion, or generalisation of the numbers and their pertaining concepts would skew the meaning and result in the hearing audience being less informed or misled. At the end of his presentation, the Deaf professional drew conclusions based on his research findings, namely many of the Auslan sentences that contained numbers. Some numbers appeared in adjacent Auslan sentences, thus increasing information density and suiting the purpose of local analysis of cognitive overload. In addition, some numbers appeared near or at the end of Auslan sentences that were either rhetorical questions (pseudo-cleft sentences, e.g., Auslan sentence 97: but what about sign language interpreter services have have have? what? provide have? sixty-two country) or topic-comment structures (e.g., Auslan sentence 88: other group only oral sign ban five). English statements, however, typically have numbers and their associated concepts at the front (e.g., ‘Sixty-two countries replied that they had signed language interpreting services,’ ‘A further five countries reported an oral-only approach that bans signed language.’). Therefore, both types of Auslan sentences are syntactically asymmetrical to their idiomatic English renditions, and interpreters may need to use long onset processing time in order to provide effective English renditions of these Auslan sentences. It is worth reiterating that sentence-level accuracy throughout this book means that an interpreter has successfully conveyed the gist (i.e., central point, essential meaning, who does what to whom) of a source language sentence instead of producing a perfect rendition. 6.1.1 Number Accuracy Versus Sentence-Level Accuracy Table 6.1 displays the absolute and relative accuracy rates of the professional interpreters’ English renditions of the 32 Auslan numbers and their 29 respective Auslan sentences.1 Although the mean accuracy rate of interpreting the individual Auslan numbers was high, at 89%, the mean accuracy rate of interpreting their respective Auslan sentences was much lower, at 60%. In other words, many interpreters rendered the number of an Auslan sentence correctly but did not transfer the core meaning of the Auslan sentence into English.The accuracy rate of two Auslan sentences (73 and 85) was extremely low, at 25% and 20%, respectively, which warrants local analysis in order to shed light on possible reasons. In particular, for all but four Auslan sentences (28, 51, 53, and 69), the number accuracy rate was markedly higher than the sentence accuracy rate. The gap between number accuracy and sentence-level accuracy for Auslan sentence 62 (90% vs. 30%), Auslan sentence 85 (100% vs. 20%), and Auslan sentence 97 (100% vs. 40%) was extremely high (60%, 80%, and 60% respectively) and therefore merits local analysis to reveal potential explanations for the tremendous differences. Despite a high accuracy rate regarding individual Auslan numbers, many interpreters skewed the referents associated with the numbers (e.g., omitting or  altering the referents by saying ‘62 interpreting services’ rather than ‘62 countries’) or distorted the messages associated with the numbers (e.g., saying

Cognitive Overload and Coping Strategies 155 Table 6.1  Absolute and relative frequency of accurate interpretations of 32 Auslan numbers and their respective sentences Slide No.

7

Auslan sentences

28 30 30 No slide 51 53 54 54 55 10 62 63 64 65 11 67 69 70 71 72 73 13 77 78 81 14 85 86 87 88 17 96 97 98 99 100 101 101 Mean accuracy rate

Numbers

ninety-three sixty-eight twenty-five 2006 2008 five thirty eight seventy-seven forty-four under half fifty-four forty-four ten nineteen nineteen ten nineteen eighty-eight sixty-one eighty-eight twenty-three thirty-five thirty-one five eighty-two sixty-two forty-three thirty-one thirty-two sixty-two thirty-two 17.75

Accurate interpretations Accurate interpretations of the number of the respective sentence N

%

N

%

20 17 17 19 18 19 16 16 18 18 14 14 19 19 18 15 20 14 18 20 18 20 14 20 20 17 20 19 17 17 19 18

100% 85% 85% 95% 90% 95% 80% 80% 90% 90% 70% 70% 95% 95% 90% 75% 100% 70% 90% 100% 90% 100% 70% 100% 100% 85% 100% 95% 85% 85% 95% 90% 89%

19 13 13 19 18 12 12 11 6 11 11 13 9 18 14 12 13 5 11 16 16 4 10 10 15 12 8 11 12 11 9 9 12

95% 65% 65% 95% 90% 60% 60% 55% 30% 55% 55% 65% 45% 90% 70% 60% 65% 25% 55% 80% 80% 20% 50% 50% 75% 60% 40% 55% 60% 55% 45% 45% 60%

‘62 countries provided sign language services’ instead of ‘62 countries provided sign language interpreting services’). These number-related errors meant that the respective Auslan sentences were interpreted inaccurately, though the Auslan numbers were transferred successfully. In summary, a key finding is that while many interpreters rendered the individual Auslan numbers accurately they omitted or substituted the referents or ideas related to the numbers and therefore skewed the sentence-level meaning. In other words, numbers not only posed interpreting difficulties themselves but also affected some interpreters’ cognitive processing of adjacent elements.

156 Cognitive Overload and Coping Strategies 6.1.2 Sentence Accuracy Versus Segment Accuracy Another interesting finding is that many interpreters rendered individual Auslan sentences correctly into spoken English but omitted or misinterpreted the preceding or subsequent Auslan sentences. Consequently, the hearing audience may miss or be misinformed about a piece of relevant information and thus form an incomplete or different understanding of the whole section. Given that only 6 (30%) of the 20 professional interpreters rendered Auslan sentence 62 accurately into English, further analysis was conducted to evaluate each interpreter’s rendition of the whole segment consisting of Auslan sentences 62, 63, 64, and 65 (see Appendix B). Each of the four Auslan sentences contained a number, which was shown on a bar chart on the Deaf presenter’s PowerPoint slide 10. These Auslan sentences conveyed messages about the recognition of Deaf people, recognition of signed languages, and existence of signed language dictionaries. Table 6.2 demonstrates the absolute accuracy of each interpreter’s English renditions of the four consecutive Auslan sentences. Surprisingly, only two participants (Jane and Lauren, marked in bold in Table 6.2) interpreted these four Auslan sentences accurately into English. Another five interpreters (Liz, Emily, Wendy, Alex, and Kay) conveyed the gist of three of these four adjacent Table 6.2  Interpreters’ accuracy rate of rendering four adjacent Auslan sentences that contained numbers shown on PowerPoint slide 10 Pseudonym

62 (77) 63 (44) 64 (under half) 65 (54) Total out of 4

Liz Amber Emily Wendy Linda Jane Molly Sophia Helen Claire Alex Annie Debbie Shannon Zoe Tiffany Monica Lauren Bernie Kay Accurate renditions Inaccurate renditions

1 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 1 1 1 6 14

1 0 1 1 0 1 1 0 1 0 1 0 1 0 0 1 0 1 0 1 11 9

0 1 1 1 0 1 0 1 0 0 1 1 0 1 1 0 0 1 0 1 11 9

1 0 1 1 1 1 0 0 1 0 1 1 0 1 1 1 0 1 1 0 13 7

3 1 3 3 1 4 2 1 2 0 3 2 1 2 2 2 0 4 2 3

Note: In the top row, numbers before the brackets represented the Auslan sentences, whereas numbers in the brackets were the numbers that appeared in the respective Auslan sentences. An accurate English interpretation of an Auslan sentence was marked as 1 and an inaccurate interpretation as 0.

Cognitive Overload and Coping Strategies 157 Auslan sentences correctly into English, three of whom had difficulties processing Auslan sentence 62. The Deaf presenter then used seven adjacent Auslan sentences (67 to 73) and a table on his PowerPoint slide 11 to report on findings regarding different countries’ formal recognition of signed languages through constitution, legislation, and policy, among other forms of recognition. Table 6.3 shows each interpreter’s accuracy rate of rendering the seven consecutive Auslan sentences, six of which contained numbers and the remaining one contained fingerspelled words (abbreviated as FS in Table 6.3). Receptive fingerspelling poses a challenge for many signed language interpreters in working from a signed language into a spoken language (Nicodemus & Emmorey, 2013; Taylor, 2002). Only one interpreter (Helen, marked in bold in Table 6.3) rendered all seven Auslan sentences accurately into spoken English. Another three interpreters (Amber, Emily, and Jane) produced acceptable English renditions for six of the seven Auslan sentences. Most interpreters had difficulties in coping with Auslan sentence 67 and/or Auslan sentence 73. Interpreters’ inaccurate renditions of these introductory Table 6.3  Interpreters’ accuracy rate of rendering seven adjacent Auslan sentences that contained numbers shown on PowerPoint slide 11 Pseudonym

67 (44) 68 (FS) 69 (10) 70 (19) 71 (19) 72 (10) 73 (19) Total out of 7

Liz Amber Emily Wendy Linda Jane Molly Sophia Helen Claire Alex Annie Debbie Shannon Zoe Tiffany Monica Lauren Bernie Kay Accurate renditions Inaccurate renditions

0 1 1 0 1 0 0 1 1 0 0 1 0 1 0 1 1 0 0 0 9

1 1 1 0 1 1 1 0 1 1 1 1 0 0 0 0 0 0 1 0 11

1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 18

1 1 1 1 1 1 1 0 1 0 1 0 0 1 0 1 1 1 1 0 14

0 0 1 1 0 1 1 1 1 0 0 1 0 1 0 1 0 1 1 1 12

1 1 1 0 1 1 1 1 1 1 0 1 1 0 0 0 0 1 1 0 13

0 1 0 1 0 1 0 0 1 0 0 0 0 0 0 0 1 0 0 0 5

11

9

2

6

8

7

15

4 6 6 4 4 6 5 4 7 3 3 5 2 4 1 4 4 4 5 1

Note: In the top row, numbers before the brackets represented the Auslan sentences, whereas numbers in the brackets were the numbers that appeared in the respective Auslan sentences. An accurate English interpretation of an Auslan sentence was coded as 1 and an inaccurate interpretation as 0.

158 Cognitive Overload and Coping Strategies and summarising sentences in this segment may result in the hearing audience feeling confused about the meaning of the whole segment. Furthermore, the Deaf presenter reported on findings regarding different educational methods that were applied to educate Deaf children in various countries by referring to a bar chart on his PowerPoint slide 14.This section was extremely challenging for three reasons: (i) each of the four adjacent Auslan sentences (85, 86, 87, and 88; see Appendix B) contained a number; (ii) both Auslan sentences 85 and 86 included a critical detail (plus have other) that had to be transferred precisely to maintain the same meaning; and (iii) each of the four Auslan sentences began with a topic (e.g., the first part of Auslan sentence 85 is bilingual have nod plus have other altogether country have what) and ended with a number (e.g., the second part of Auslan sentence 85 is twentythree country have that) and therefore had to be restructured into idiomatic English. Table 6.4 demonstrates that only three interpreters (Liz, Amber, and Jane, marked in bold in Table 6.4) rendered these four Auslan sentences accurately into English. The accuracy rate with regard to the first Auslan sentence (85) in this segment was particularly low, as only four interpreters conveyed its meaning effectively in the target language. Table 6.4  Interpreters’ accuracy rate of rendering four adjacent Auslan sentences that contained numbers shown on PowerPoint slide 14 Pseudonym

85 (23)

86 (35)

87 (31)

88 (5)

Total out of 4

Liz Amber Emily Wendy Linda Jane Molly Sophia Helen Claire Alex Annie Debbie Shannon Zoe Tiffany Monica Lauren Bernie Kay Accurate renditions Inaccurate renditions

1 1 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 16

1 1 0 0 0 1 1 1 0 0 0 1 1 0 1 1 0 0 1 0 10 10

1 1 0 1 0 1 0 1 0 0 1 0 0 1 1 0 0 1 1 0 10 10

1 1 1 1 1 1 1 1 0 0 1 1 1 1 0 1 1 1 0 0 15 5

4 4 2 2 1 4 2 3 0 0 2 2 2 2 2 2 1 2 2 0

Note: In the top row, numbers before the brackets represented the Auslan sentences, whereas numbers in the brackets were the numbers that appeared in the respective Auslan sentences. An accurate English interpretation of an Auslan sentence was coded as 1 and an inaccurate interpretation as 0.

Cognitive Overload and Coping Strategies 159 In addition, the Deaf professional reported on findings regarding Deaf p­ eople’s access to government services and signed language interpreting services by incorporating numbers presented on a bar chart on his PowerPoint slide 17. This sequence of six adjacent Auslan sentences (96, 97, 98, 99, 100, and 101; see Appendix B) formed a dense and coherent segment because the numbers generally formed a specific pattern (i.e., showing a descending order) and their associated concepts (i.e., access to government services, signed language interpreting services, qualified and trained signed language interpreters, code of ethics for the interpreters, and government funding for the interpreting services) were interrelated.Table 6.5 provides information on each interpreter’s accuracy rate of rendering these six adjacent Auslan sentences into spoken English. Surprisingly, only one interpreter (Alex, marked in bold in Table 6.5) rendered all the six Auslan sentences accurately into English. Another five interpreters (Liz, Amber, Emily, Jane, and Helen) produced effective English interpretations for five of the six Auslan sentences. Taken together, the results given above reveal that it was extremely challenging for professional interpreters to render a sequence of signed language sentences containing numbers accurately into a spoken language. Inter alia, the Table 6.5  Interpreters’ accuracy rate of rendering six adjacent Auslan sentences that contained numbers shown on PowerPoint slide 17 Pseudonym

96 (86) 97 (62) 98 (43) 99 (31) 100 101 (62 Total (32) & 32) out of 6

Liz Amber Emily Wendy Linda Jane Molly Sophia Helen Claire Alex Annie Debbie Shannon Zoe Tiffany Monica Lauren Bernie Kay Accurate renditions Inaccurate renditions

0 1 1 0 1 1 1 0 1 1 1 0 0 1 0 1 0 1 0 1 12 8

1 0 1 0 0 1 1 0 0 1 1 0 0 0 1 0 0 0 1 0 8 12

1 1 1 1 1 1 1 1 1 0 1 0 0 1 0 0 0 0 0 0 11 9

1 1 1 1 0 1 0 0 1 0 1 0 1 0 0 1 0 1 1 1 12 8

1 1 1 0 0 1 0 1 1 0 1 0 1 1 1 0 0 1 0 0 11 9

1 1 0 1 0 0 1 1 1 0 1 0 0 0 1 0 0 0 0 1 9 11

5 5 5 3 2 5 4 3 5 2 6 0 2 3 3 2 0 3 2 3

Note: In the top row, numbers before the brackets represented the Auslan sentences, whereas numbers in the brackets were the numbers that appeared in the respective Auslan sentences. An accurate English interpretation of an Auslan sentence was coded as 1 and an inaccurate interpretation as 0.

160 Cognitive Overload and Coping Strategies accuracy rate of interpreting a single Auslan sentence containing a number into spoken English was substantially higher than the accuracy rate of rendering a cluster of Auslan sentences containing numbers into spoken English. Surprisingly, for adjacent Auslan sentences whose key points such as numbers were presented on a PowerPoint slide, typically only a few interpreters rendered all these Auslan sentences correctly into spoken English when they worked alone and could not see projected PowerPoint slides.

6.2 Free and Literal Interpretations for Coping with Numbers in the Middle of Sentences This section features local analysis of two interpreters’ (Jane and Lauren) acceptable English renditions of a sequence of four adjacent Auslan sentences (62, 63, 64, and 65, with all the sentences except 64 containing a number in the middle of the sentence; see Appendix B for literal English translations) to reveal their cognitive processing and distinct strategies for coping with the numbers and dense information. An innovation of this study is that the local analysis of a spatially and temporally aligned transcript using the adapted version of Gile’s (2009) Effort Model of simultaneous interpreting (see Section 2.5 in Chapter 2) as a theoretical framework provides insights into the complex interrelation of an interpreter’s processing time, processing unit, cognitive processing, strategies, and performance quality in signed-to-spoken language simultaneous interpreting. 6.2.1 Long Processing Time, a Lack of Suppression, Self-correction, and Free Interpretation As illustrated by Example 6.1, Jane used an average onset processing time of 4.6 seconds for Auslan sentences 62 to 65, made a self-correction, adopted a predominantly free interpretation approach, and strategically incorporated pauses to produce effective English interpretations of this Auslan segment. Like 12 other interpreters in this study, Jane misinterpreted recognise their deaf people in Auslan sentence 62 as ‘had recognised sign language for Deaf people’; however, she was the only interpreter who made a successful self-correction. Both the likely reasons for her initial error and the impact of her self-correction on her processing of incoming information are inferred from a local analysis of her interpreting process, as described here. The key concepts, numbers, and her selfcorrection are highlighted in bold. Interesting phenomena revealed by the detailed analysis are then discussed in relation to the relevant literature. Jane seemed to have finished her English interpretation of the previous Auslan sentence (61) when the Deaf presenter started producing Auslan sentence 62; however, she was not satisfied with her interpretation, so she paused for 3.8 seconds, then added a phrase in order to perfect it. As a consequence, her onset processing time for Auslan sentence 62 was quite long, at 7 seconds. When she began interpreting, the Deaf presenter was producing the key signs deaf people

Cognitive Overload and Coping Strategies 161

Example 6.1  Jane’s long onset processing time, lack of suppression, self-correction, and free interpretation for coping with numbers

(Object of the verb recognise), which she later distorted. She changed the question sign how many to ‘which,’ possibly because her long onset processing time weakened her short-term memory of the sign. When the Deaf presenter was producing seventy-seven in Auslan sentence 62, Jane was uttering the verb ‘recognised’ and starting to say the inaccurate object ‘sign language for Deaf people.’ In fact, many signed language interpreters and Deaf people are highly familiar with the heated debates on whether signed languages should be recognised as equal to spoken languages; in contrast, people talk less about whether Deaf people have been recognised as equal to hearing

162 Cognitive Overload and Coping Strategies people. Regarding context, 90 seconds earlier, the Deaf presenter stated that the United Nations Convention on the Rights of Persons with Disabilities (UNCRPD) defines language as consisting of both spoken language and signed language (Auslan sentence 58) and that the UNCRPD is the first document that recognises the evidence showing signed language is on a par with spoken language (Auslan sentence 59). In Auslan sentence 61, the Deaf presenter mentioned sign language-related content in the UNCRPD, including the use of sign languages, governments’ recognition of sign languages, and sign language research. These messages preceding Auslan sentence 62 are about sign language. Probably the interpreter’s own preconceptions and memory traces of the previous sign-language-related information led to her unjustifiable distortion of deaf people as ‘sign language for Deaf people.’ It is also likely that when she saw the signs deaf people for the first time, she was focusing more attention on starting her English rendition of Auslan sentence 62 than on comprehending or memorising these two signs. Another plausible explanation is that both the slightly long processing time regarding deaf people (3.3 seconds) and the concurrent processing at the particular juncture (understanding and memorising the number seventy-seven while articulating the verb ‘recognised’) left Jane with an obscure memory of the aforementioned object deaf people and insufficient processing capacity for reformulating it correctly into English. Jane rendered the numbers seventy-seven and ninety-three accurately, with processing time of 3.5 seconds and 2.5 seconds, respectively. She omitted the referent (country) of ninety-three maybe due to high cognitive load at the particular point in time (she was saying the phrase ‘out of 93’ while comprehending the signs deaf people recognise). This time she interpreted the signs deaf people as ‘Deaf people uh and sign language,’ with the filler ‘uh’ indicating some self-doubt. She put ‘Deaf people’ before ‘uh and sign language’ probably because she just saw the signs deaf people and was more confident about this concept than ‘sign language,’ which she had said when she first saw deaf people. Near the end of Jane’s English rendition of Auslan sentence 62, her grammatical error (‘was’) in ‘Deaf people uh and sign language was recognised’ was probably due to a momentary attention lapse as she was engaged in multiple cognitive sub-processes, including completing this English rendition, comprehending the essential signs sign language in Auslan sentence 63, as well as monitoring what she had just said against both what he had just signed and what he was signing at the juncture. Once she saw the Deaf presenter’s new sentence beginning with but what about sign language, she realised that Auslan sentence 62 was not about recognising sign language but about recognising Deaf people. As a result, she quickly and overtly corrected herself by saying ‘sorry, said that Deaf people were recognised.’ Overall, it appears that Jane was actively monitoring her target language output while processing the incoming source language information. Due to her self-correction, Jane lagged far (5.4 seconds) behind the Deaf presenter when she started to interpret Auslan sentence 63. The long overlap between her target language production of the previous source language sentence and her source language comprehension of the current source language

Cognitive Overload and Coping Strategies 163 sentence caused imported cognitive load (see Gile, 2008). Therefore, she strategically omitted the Deaf presenter’s redundancy in order to catch up. She swiftly rendered the two meaning units – but what about sign language what? finish they have approve confirm point sign language? (note that these two questions bear the same meaning and that approve confirm are synonyms) – as a short English phrase ‘As for sign language being recognised formally,’ in order to introduce the new topic. Here she strategically added the adverb ‘formally’ to not only highlight the novel topic of Auslan sentence 63 but also distinguish the new topic from her previous errors. Due to the aforementioned two strategic omissions, Jane was ahead and paused for 2.6 seconds to wait for more Auslan information. When Jane saw that the Deaf presenter began to produce a new number (forty-four) in Auslan sentence 63, she quickly resumed her English interpretation by saying ‘there were only forty-four countries,’ with processing time for forty-four being 1 second. She then appeared to have experienced some difficulty in reformulating the Deaf presenter’s expression of nod say government into English, as evidenced by her minor self-repair (‘who in which’) and wordiness (‘who in which the respondents said that,’ which could be simplified as ‘who said that’).This target language production difficulty resulted in her having slightly long processing time (3.1 seconds) when interpreting the recurrent number ninety-three. Jane’s free interpretation style is evidenced by her consistent use of the verb ‘recognise’ to collocate with objects such as ‘Deaf people’ and ‘sign language’ in spite of the Deaf presenter’s use of synonymous verb signs (recognise, approve, confirm) for the same concept. When Jane was finishing off her English interpretation of Auslan sentence 63, she saw under half in the following Auslan sentence (64).Therefore, she promptly started interpreting this short sentence (with onset processing time of 1.9 s­ econds) and produced a fairly literal interpretation (‘That’s under 50% of respondents’). The English gloss ‘under’ can be changed to ‘less than’ to make the English sentence more idiomatic. Her processing time for under half was 1.5 seconds. With onset processing time of 4.1 seconds for Auslan sentence 65, Jane strategically omitted the repeated question signs (have have have what what what?) for which a literal interpretation into English would be nonsensical, then introduced the new topic by saying ‘Sign language research was another area,’ during which she saw fifty-four. She waited for 1.2 seconds during which the Deaf presenter was producing a redundant phrase (finish research sign language), then started a new sentence with the subject ‘Only fifty-four countries.’ Note that Jane strategically added the referent ‘countries’ to the number ‘fifty-four’ in order to make the concept clear and complete. Her processing time for fifty-four was 2.5 seconds. For clarity, the Deaf presenter first used a depicting sign to visually describe the hand movements and actions of a person turning the pages of a book, then fingerspelled the English word ‘dictionary.’ After understanding the Deaf presenter’s redundancy, Jane made a strategic omission by saying the concept only once (‘said that there have been a sign language dictionary’).

164 Cognitive Overload and Coping Strategies Jane then paused for 1.9 seconds before completing her English rendition for Auslan sentence 65 by saying the phrase ‘published in their country.’ This phrase links back to the concept ‘sign language research’ and reflects an academic register. It is worth noting that Jane omitted the verb make (which coincided with her articulation of the number ‘fifty-four’), which probably is due to her devoting of more attention to target language production than to source language comprehension at that point in time. Her grammatical errors (‘said that there have been a sign language dictionary’), which overlapped with the Deaf presenter’s producing of the depicting sign for ‘dictionary’ and the first letter of its fingerspelling, may have resulted from her allocation of more attention to source language comprehension than to target language p­ roduction or self-monitoring. Morgan – one of the external raters – made the following comments on Jane’s English simultaneous interpretation of the whole Auslan presentation that are also true of Example 6.1: She is very good at making sure that her initial English sentences make sense and using nice little fillers to make the sentences complete and grammatical. She used her time lag [processing time] well. Excellent lexical choices. Her register was good. … She was pausing in the right places most of the part. … Her intonation was good. … She was making some very strategic decisions around. She was adding a lot to make it coherent and sometimes she would omit stuff when it would sound repetitious in English. In summary, Jane’s perfection of her initial English interpretation of Auslan sentence 61, insufficient suppression of her own biases and/or erroneous inference drawn from the context, long onset processing time, and sub-optimal timing of starting to interpret Auslan sentence 62 appeared to have led to her initial erroneous interpretation of deaf people. The Deaf presenter’s reiteration of the same concept gave Jane a second chance; and her own active self-monitoring contributed to the successful self-correction. She then made strategic omissions to deal with the Deaf presenter’s (lexical and semantic) redundancy and catch up. She also made strategic additions to ensure that her target language output is comprehensible, grammatical, and idiomatic. Overall, Jane adopted a predominantly free interpretation approach to cope with these four Auslan sentences containing numbers in the middle of the sentences, probably due to her high familiarity with the Deaf presenter and sufficiently long onset processing time.

6.2.2 Short Processing Time and Literal Interpretation This section presents local analysis of Lauren’s English interpretation of the same Auslan segment (see Example 6.2) and reveals that his predominantly literal interpretation style is primarily due to his short onset processing time. His average onset processing time for these four Auslan sentences (62 to 65) was 2.5 seconds.

Cognitive Overload and Coping Strategies 165

Example 6.2  Lauren’s literal interpretation to cope with numbers and dense information

While Jane achieved high accuracy in her English rendition through a predominantly free interpretation approach, Lauren achieved it by employing a predominantly literal interpretation approach. That is, Lauren closely adhered to the source language lexical choices and syntactic structures and used shorter (onset and number) processing time than Jane. Lauren’s approach may be his preference or strategy to cope with the dense information and numbers in this Auslan segment in order to ensure the accurate transfer of semantic content. His cognitive processing of this Auslan section is briefly described here. With onset processing time of 4.4 seconds, Lauren rendered the first meaning unit of Auslan sentence 62 (s-o well like have how many o-f those country)

166 Cognitive Overload and Coping Strategies into English as ‘So, how many of those countries,’ which matches almost all the English glosses of these Auslan signs. He then reinforced that this sentence was a question by strategically using the padding phrase ‘you may ask,’ while waiting for the object of the verb recognise (i.e., using stalling as a strategy). After seeing the awaited object deaf people, he completed the English question by saying the phrase ‘have recognised um Deaf people?’The filler ‘um’ indicates his slight hesitation and doubt about the object. He then waited for 1.9 seconds to receive further Auslan information. Following the syntactic structure of the Deaf presenter’s Auslan answer only seventy-seven … equal, Lauren rendered it as an English sentence with direct speech (‘Only seventy-seven … said, “Yes, Deaf people are recognised as equal in our country.”’). His strategic addition of the phrase ‘in our country’ suggests that he was making efforts to produce comprehensible and coherent English sentences. Lauren continued to use short onset processing time and a predominantly literal interpretation approach to process the subsequent Auslan sentences. With onset processing time of 2.1 seconds, Lauren started interpreting Auslan sentence 63 as soon as he understood its topic (but what about sign language), which he rendered as ‘But I’m wondering about sign language in particular↑.’ His literal interpretation approach is obvious from his retention of the transitive sign but as the first word of his English sentence, which can be improved by the use of the word ‘However.’ He reframed the question signs what about as ‘I’m wondering about’ in an attempt to introduce the new topic, but the first person pronoun (‘I’m’) in this phrase may single out the Deaf presenter from his research team and may also mislead the non-deaf audience to believe that the interpreter himself was wondering about the issue; hence, it is better to replace it with ‘we were.’ Drawing on the previous information (but what about sign language), Lauren successfully predicted the object (sign language) of the Auslan verbs approve confirm before the Deaf presenter produced it. Lauren again retained the rhetorical question in Auslan sentence 63 by posing the question, ‘Has sign language been accepted or confirmed as an actual language?’ Due to Lauren’s successful prediction, he strategically added the phrase ‘as an actual language’ in order to complete the current English sentence and reflect the Deaf presenter’s emphasis. He then followed the Deaf presenter’s syntactic structure and produced a full answer (‘There were only forty-four countries that said their government had approved sign language as a language, out of ninety-three countries’). In regard to grammatical error, Lauren should have added the word ‘that’ between ‘There were only forty-four countries that said’ and the clause ‘their government … ninety-three countries.’ His short processing time and fast processing enabled him to achieve high accuracy of information transfer, as he retained the referents of the numbers (‘44 countries’ and ‘93 countries’), even rendered the Deaf presenter’s comment hmmm as ‘It’s quite interesting,’ and conveyed the Deaf presenter’s nuanced meaning such as only consistently and accurately. Like Jane, Lauren not only used short onset processing time (1.53 seconds) for Auslan sentence 64 but also said the English gloss ‘under’ of the Auslan sign under.

Cognitive Overload and Coping Strategies 167 With onset processing time of 1.9 seconds, Lauren began the interpretation of Auslan sentence 65 with a transition word and a topic ‘But sign language research.’ Then he retained the rhetorical question by first asking a question ‘is sign language research happening within these countries?’ then providing an answer that contained the number.The Deaf presenter’s omission of the referent of the number fifty-four appeared to be a reason for Lauren’s minor self-­ correction (‘Only fifty-four from those sign [0.7-second pause] from those countries’). He made the correction in an attempt to produce a grammatical and comprehensible English sentence. Similar to Jane, Lauren also made a strategic omission of the Deaf presenter’s redundancy regarding dictionary. There are also minor grammar mistakes in Lauren’s English interpretations. For example, the past perfect tense in ‘had produced their own sign language dictionary’ needs to be changed to past tense ‘produced’; and the singular noun ‘dictionary’ needs to be changed to its plural form ‘dictionaries’ to match the subject ‘fifty-four … from those countries.’ Lauren interpreted all six numbers accurately, probably due to his short processing time regarding these items. His processing time for seventy-seven, ninety-three, forty-four, ninety-three, under half, and fifty-four was 1.3 seconds, 1.1 seconds, 1.3 seconds, 1.5 seconds, 1.1 seconds, and 1.3 seconds, respectively, with the mean number processing time being 1.3 seconds. Interestingly, Lauren’s processing time for these numbers is generally shorter than Jane’s (3.5 seconds, 2.5 seconds, 1 second, 3.1 seconds, 1.5 seconds, and 2.5 seconds, respectively, with her average number processing time being 2.4 seconds). Since Lauren retained three rhetorical questions in this Auslan segment (see sentences 62, 63, and 65) and used a direct quote (see sentence 62), overall his English renditions reflected an informal register, which did not optimally represent the Deaf professional’s formal register at the conference setting. Moreover, Lauren used the English glosses of most signs in the Auslan source text, thus occasionally producing ungrammatical and unidiomatic expressions in English. For example, when the Deaf presenter used the verb sign recognise to collocate with the object deaf people, Lauren followed suit and said ‘have recognised um Deaf people,’ which is idiomatic English. However, when the Deaf presenter switched to synonymous verb signs approve confirm and approve to collocate with the object sign language, Lauren still used the English glosses of these signs and said ‘confirmed’ and ‘approved,’ which do not collocate effectively with the object ‘sign language’ in English. Lauren’s Auslan-to-English simultaneous interpretations in Example 6.2 show evidence of source language intrusion (lexical and syntactic interference) that has been found in previous studies on spoken-to-signed language interpreting (see Cokely, 1986, 1992; Napier, 2016; Taylor, 2002). Alternative lexical choices such as ‘recognised’ would collocate better with ‘sign language’ and match the formal register typically used in a conference setting, as what Jane did in Example 6.1. In regard to Auslan sentence 65, Lauren’s use of the verb ‘happening’ in the English question (‘But sign language research, is sign language research happening within these countries?’) also sounds somewhat informal for a formal context.

168 Cognitive Overload and Coping Strategies Intonation also has an impact on register. Speakers typically signal the end of English statements with a falling intonation, slow pace, and a pause (Taylor, 2002). However, Lauren ended two English sentences (see his English renditions for Auslan sentences 63 and 65) with a rising intonation, which probably lowered the register of his English output and misrepresented the Deaf professional in the conference setting. Interpreters’ use of rising intonation at the end of English statements may portray a Deaf professional as a subordinate or powerless figure, potentially causing the hearing audience to have a wrong and negative perception of the Deaf professional (Morgan, 2008).The identified issues such as the persistent use of rhetorical questions, inappropriate lexical choices regarding collocation, rising intonation, and informal register in Lauren’s English interpretations appear to indicate that he may not be consciously aware of these patterns in his own target language renditions. An alternative explanation may be that Lauren was under immense cognitive pressure and did not have sufficient processing capacity for actively monitoring his target language output. In addition, during the analytic assessment of Lauren’s English interpretations of the entire Auslan monologue, external rater Jamie pointed out some of the aforementioned issues: There are many areas where he wasn’t listening to what he was saying and it wasn’t grammatical, even though he speaks beautifully. And then the little low registers. … He follows the Auslan syntax, doing topic-comment. He just has to wait, and embed it [the source language information] all into one instead of using rhetorical questions all the time. So, he brought down the register a bit more than I remembered. … He is quite literal, but I also note that he understands what the Deaf presenter is saying. For me, it’s a couple of things, including he is also someone who is not university-educated.You can tell that, the English that he uses, he is not even hearing that it’s ungrammatical a lot of the time. He doesn’t know [that what he is saying is ungrammatical].Yeah, I get a sense because it happened so often. Interestingly, Lauren explained in the retrospective interview (see Section 4.3.1.3 in Chapter 4) that he used quite short processing time in this Auslan-to-English simultaneous interpreting task because he was unfamiliar with both the Deaf presenter and the content of his Auslan presentation. Lauren also stated that he had to devote more attention to comprehending the signed discourse than searching for the perfect English words (see Lauren’s remarks in Section 4.3.1.3), which affected the register of his English renditions (see Lauren’s comments in Section 4.3.3.2). Lauren also acknowledged that his predominantly literal interpretation approach occasionally led to ungrammatical English renditions and/or interpretation breakdowns, and that his informal English register may not have accurately represented the Deaf professional in the formal context (see Lauren’s quote in Section 4.4). In addition, it appears that both Lauren’s predominantly literal interpretation style and informal register largely resulted from his very short onset processing time (approximately 2 seconds for Auslan sentences 63, 64, and 65). It seems that

Cognitive Overload and Coping Strategies 169 he generally adhered to the Deaf presenter’s lexical choices and syntactical structures (i.e., transcoding, literal interpretation) in order to cope with the numbers and dense information in the source text. Overall, this strategy enabled him to achieve high accuracy in terms of semantic transfer, but it led to some source language interference, grammatical errors, and informal register in his Auslanto-English simultaneous interpretation. It is plausible that the notable differences between Jane and Lauren in terms of the length of processing time, interpretation style, and register are related to their differences in terms of familiarity with the Deaf presenter, formal interpreting training, and general educational background. Jane’s processing time (e.g., onset processing time, number processing time) was often noticeably longer than Lauren’s; and Jane paused for a few seconds every now and then to process more source language information and plan her target language output. These findings suggest that Jane’s processing unit is probably larger than Lauren’s. Jane completed both a Diploma of Interpreting (Auslan/English) and a Postgraduate Diploma in Auslan-English Interpreting training, whereas Lauren only completed a Diploma of Interpreting (Auslan/English). At the time of data collection, Jane held a PhD degree related to English language, whereas Lauren did not receive university education. Although they both understood the Deaf presenter’s Auslan source text quite well and produced highly accurate English interpretations, their interpretations differed markedly in terms of style, grammaticality, idiomaticity, comprehensibility, register, and communicative effect.

6.3 Coping Strategies for Numbers Near or at the End of Sentences While the previous section focused on Auslan sentences that had numbers in the middle of the sentences, this section reports on interpreters’ use of strategies to cope with a sequence of Auslan sentences, each of which had a number near or at the end. As a result, the interpreters needed to cope with three types of interrelated challenges: numbers, dense information, and syntactical differences between Auslan and English. Interpreters needed to memorise the numbers and their associated referents and ideas while processing other information; they needed to attend to both details and central points to deal with dense information; they needed to restructure Auslan sentence components to cope with syntactical differences between Auslan and English in order to produce idiomatic English renditions; and they needed to express the meaning of the current challenging Auslan sentence while comprehending the following difficult Auslan sentence. What follows is local analysis of effective English interpretations of two particularly challenging segments in the signed presentation, that is, Auslan sentences 85 to 88 (see Table 6.4 for the accuracy rate of their English interpretations), and Auslan sentences 96 to 101 (see Table 6.5 for the accuracy rate of their English renditions). As noted earlier, each of the four Auslan sentences in the first segment (85 to 88) featured a topic-comment structure, which started with

170 Cognitive Overload and Coping Strategies a topic (i.e., a specific type of education method for Deaf children) and ended with a number. Four Auslan sentences in the second segment (97 to 100) were rhetorical questions (pseudo-cleft sentences), each of which started with a content question and ended with an answer that contained a number. Both the topic-­comment structure and the pseudo-cleft structure serve to bring the new information in these sentences (i.e., the comment or the answer) into focus (see Johnston & Schembri, 2007), thus emphasising the new information. Another two Auslan sentences (96 and 101) in the second segment also contained numbers, but in the middle of the sentences. I conducted local analysis of the accurate English renditions of Auslan sentences 85, 86, 87, 88, 97, 98, 99, and 100, each of which contained a number near or at the end, and found that interpreters employed strategies to cope with the numbers, syntactical differences between the source language and the target language, and dense information. Interestingly, many interpreters drew on contextual knowledge (e.g., the signs how many before or within these Auslan sentences indicating that numbers would be mentioned), redundancy, and pauses within/between the Auslan sentences to produce effective English renditions. Interestingly, interpreters’ accurate renditions featured both diversity and creativity. Since it was almost impossible to predict the specific numbers near or at the end of these Auslan sentences, interpreters typically rendered the numbers only after seeing them. Interpreters primarily employed the following four strategies to cope with the combined challenges (numbers, dense information, and syntactical differences between the source language and the target language) of adjacent Auslan sentences featuring numbers near or at the end: 1 The interpreter commenced interpreting upon or after seeing the Auslan number, and, due to the exceptionally long onset processing time, rendered the core meaning of the Auslan sentence accurately by using an idiomatic and concise English sentence. 2 With moderate onset processing time, the interpreter started interpreting after seeing the topic of the Auslan sentence, utilised appropriate strategies (e.g., repeating a certain part of the topic, adding a padding phrase [i.e., stalling], pausing for a while) to absorb more information, and finally said the number near or at the end of the English sentence. 3 With short onset processing time, the interpreter began interpreting after seeing a few initial signs of the Auslan sentence, used appropriate strategies (e.g., repeating a certain part of the topic, adding a padding phrase, pausing for a while) to receive more information, and ultimately uttered the number near or at the end of the English sentence. 4 Instead of rendering an Auslan pseudo-cleft structure (rhetorical question) as an English statement, the interpreter used moderate onset processing time, adhered to the question-and-answer structure, and finally provided an answer containing the number.

Cognitive Overload and Coping Strategies 171 This section is divided into two subsections, one on the segment that featured Auslan topic-comment structure, and the other on the segment featuring Auslan pseudo-cleft structure. 6.3.1 Topic-Comment Structure and Numbers Three interpreters (Liz, Amber, and Jane) rendered Auslan sentences 85 to 88 accurately into English. I selected only Liz and Amber’s English interpretations for local analysis, because Jane’s strategy was analogous to Amber’s. 6.3.1.1 Use of Different Strategies and Printed PowerPoint Slide Liz’s strategies varied considerably from Amber and Jane’s. Liz used the third strategy (given earlier) to cope with Auslan sentence 85, then switched to the first interpreting strategy to deal with Auslan sentences 86, 87, and 88. Moreover, Liz repeatedly checked the printed PowerPoint slides throughout this Auslan segment in order to ensure the accuracy of her English interpretations and follow the Deaf presenter’s track of thought. Amber and Jane, however, consistently adopted the second strategy to process these four Auslan sentences, without any reference to the printed PowerPoint slides. Liz’s renditions will be presented before Amber’s. Example 6.3 illustrates Liz’s strategies for coping with these four

Example 6.3  Liz’s use of the first and third strategies and the printed PowerPoint slides for coping with numbers near or at the end of Auslan sentences

172 Cognitive Overload and Coping Strategies adjacent Auslan sentences containing numbers near or at the end. Liz’s average onset processing time for Auslan sentences 85 to 88 was 4.7 seconds. Her processing time for twenty-three, thirty-five, thirty-one, and five was 5 ­seconds, 2.8 seconds, −0.1 second, and 0.6 second, respectively, with an average of 2.1 seconds. With onset processing time of 0.7 second, Liz started interpreting Auslan sentence 85 when the Deaf presenter was producing the first sign bilingual. She uttered a padding phrase (‘We wanted to look at’; i.e., stalling), then paused for 2.2 seconds to wait for the topic of this Auslan sentence. After looking at the printed PowerPoint slide, she introduced the topic by saying ‘countries that have bilingual education as well as other education.’ It is worth noting that her articulation of ‘(edu)cation as well as’ coincided with the Deaf presenter’s number twenty-three. After seeing this number, Liz looked at the printed PowerPoint slide again and said ‘other education,’ probably to verify this important detail (plus have other) and the number against the bar chart presented on the PowerPoint slide. After pausing for 2.1 seconds (during which she looked at the printed PowerPoint slide and the Deaf presenter multiple times), she uttered the number and repeated the aforementioned topic by saying ‘twenty-three countries have bilingual education as one of the other methods of e, one of several methods of education.’ As expected, Liz’s reiteration of the topic and self-repair resulted in her tailto-tail span (the temporal interval between the end of an Auslan sentence and the end of its corresponding English interpretation) for Auslan sentence 85 being 8.6 seconds. When she completed her English interpretation of Auslan sentence 85, the Deaf presenter just finished producing Auslan sentence 86. Given that Liz’s onset processing time for Auslan sentence 85 was quite short, her strategy can be deemed as the aforementioned third interpreting strategy. Liz glimpsed at the printed PowerPoint slide again when she finished interpreting Auslan sentence 85 probably to check information about Auslan sentence 86. Seeing that the Deaf presenter was about to start a new Auslan sentence (87), Liz quickly began interpreting Auslan sentence 86 by producing an idiomatic and concise English sentence fluently (‘Thirty-five have total communication as one of several approaches’), namely saying the number first and then its associated concept. Note that she omitted the referent countries for the number thirty-five maybe because of high cognitive load, time pressure, and her urge to catch up with the Deaf presenter.This succinct interpretation did shorten her processing time to some extent.With her tail-to-tail span being 4 seconds for Auslan sentence 86, Liz finished interpreting this Auslan sentence when the Deaf presenter was producing speak(-sign) in Auslan sentence 87. Liz was also repeatedly checking the printed PowerPoint slide probably to keep herself on track and check the number and its pertaining concept on the bar chart. When the Deaf presenter was producing thirty-one, the final sign of Auslan sentence 87, Liz began interpreting this sentence by saying this number first, strategically adding its referent (‘countries’), and expressing its pertaining concept (‘Thirty-one countries have total communication as the only method of

Cognitive Overload and Coping Strategies 173 communi education’). Partly due to Liz’s strategic omission of the redundant explanation of total communication (speak-sign altogether), her tail-to-tail span for Auslan sentence 87 was 2.8 seconds, which is shorter than her tail-totail span for the previous two Auslan sentences. With onset processing time of 3.8 seconds, Liz started interpreting Auslan sentence 88 when the Deaf presenter was producing five, the final sign of this sentence. After seeing the number, she produced an idiomatic English sentence that started with the number and its referent, then proceeded to the pertaining idea (‘And five countries have the, have the oral education as the only method of communication in education’). Taken together, Liz adopted the first interpreting strategy to cope with Auslan sentences 86, 87, and 88, each of which ended with a number. She lagged behind the Deaf presenter for almost a whole Auslan sentence and referred to the printed PowerPoint slide repeatedly in order to produce accurate English renditions and keep up. It is interesting to note that Liz’s frequent reference to the printed PowerPoint slide did not constitute a distraction but facilitated her interpreting process and interpreting performance. This finding is largely in line with previous results (Desmet, Vandierendonck, & Defrancq, 2018; LambergerFelber, 2001) that interpreters’ visual access to projected PowerPoint slides and their use of source speech manuscript during simultaneous interpreting improved the accuracy of interpreting performance. In retrospect, in terms of research design, it would have been better if the PowerPoint slides were projected as the background of the Auslan presentation. 6.3.1.2 Use of Strategic Additions and Linearity Example 6.4 illustrates that Amber consistently adopted the second strategy to cope with the numbers near or at the end of Auslan sentences, dense information, and syntactical differences between Auslan and English. She first expressed the number-related concepts by using indefinite phrases such as ‘Some countries,’ ‘Other countries,’ ‘And the countries that responded,’ and ‘And those’ – strategic additions – then specified the exact numbers at the end of her English renditions. This strategy enabled Amber to use moderate onset processing time, follow the Deaf presenter’s train of thought, avoid cognitive overload, and produce accurate English renditions. Her average onset processing time for Auslan sentences 85 to 88 was 3.9 seconds. Her processing time for twenty-three, thirty-five, thirty-one, and five was 4.5 seconds, 4.2 seconds, 3.5 seconds, and 4.5 seconds, respectively, with an average of 3.9 seconds. Amber was finishing off her English interpretation of the previous Auslan sentence when the Deaf presenter started Auslan sentence 85. With onset processing time of 4 seconds, she began interpreting this Auslan sentence when the Deaf presenter had just finished introducing the topic of this sentence (bilingual have nod plus have other altogether). Drawing on the context,2 she strategically added an indefinite quantifier (‘Some’) to the referent ‘countries’ before conveying the aforementioned meaning unit in English (‘have a

174 Cognitive Overload and Coping Strategies

Example 6.4  Amber’s use of the second strategy to cope with numbers near or at the end of Auslan sentences

bilingual education among other options for educations’). Her minor grammar mistakes – ‘a bilingual education’ and ‘educations’ – are probably due to a lack of effective self-monitoring during the complex multitasking of simultaneous interpreting. While she was uttering this meaning unit, the Deaf presenter was finishing off Auslan sentence 85 with the number and its referent (twentythree countries have that) and moving onto the next Auslan sentence. After she saw the first two signs (other group) of Auslan sentence 86, she uttered the short English clause (‘that was 23 countries’) to specify the number and its referent, eliminating the ambiguity surrounding her initial noun phrase (‘Some countries’). Her use of this strategic addition to cope with the Auslan topiccomment structure contributed to short tail-to-tail span for Auslan sentence 85, at 2.5 seconds. Amber employed the same strategy to deal with Auslan sentence 86 that had the same syntactic structure.With moderate onset processing time of 4.2 s­ econds, she started interpreting this Auslan sentence by rendering the three initial signs (other group say) literally as ‘Other countries say.’ As she specified the general concept group as a concrete referent ‘countries,’ this instance can be deemed a strategic addition. Uttering this English phrase not only bought her time to fully understand the topic of this sentence (have total communication plus

Cognitive Overload and Coping Strategies 175 other ­methods), but also prepared the listeners for the subsequent number. As soon as she finished analysing this Auslan meaning unit, she rendered it into English correctly as ‘total communication as well as other methods of education.’ While she was saying this phrase, the Deaf presenter was finishing this Auslan sentence with the number phrase (thirty-five countries hmmmm) and producing the first two signs of Auslan sentence 87. After analysing this Auslan number phrase, Amber used a short clause (‘that was 35 countries’) to express its meaning faithfully. Due to her strategy of beginning with the indefinite phrase ‘Other countries say,’ then specifying the number at the end of her English sentence, her tail-to-tail span for Auslan sentence 86 was again short, at 2.9 seconds. With onset processing time of 3.3 seconds, Amber started interpreting Auslan sentence 87 after seeing the Deaf presenter’s signs those only total communication. She expanded the first sign those into a long English phrase ‘And the countries that responded,’ probably to buy time and receive more Auslan information. When saying this phrase, she saw and understood the complete topic (only total communication speak-sign altogether). She paused for half a second after saying ‘responded,’ during which she saw that he was producing the handshape for three. Realising that the presenter was about to produce a number, she resumed her interpretation by reformulating the previous meaning unit as ‘saying that only total communication was used,’ making use of his pause on the number one. Like Liz, Amber also strategically omitted the Deaf presenter’s redundant explanation for total communication (speak-sign altogether). Amber completed her English rendition of Auslan sentence 87 by saying the number and strategically adding its referent (‘was 31 countries’). Her tail-to-tail span for this Auslan sentence was 2.5 seconds. Lagging behind the Deaf presenter by 4.2 seconds, Amber began interpreting Auslan sentence 88 when he was holding the final sign of this sentence – the number five. Adopting the same interpreting strategy as before, she started her English sentence with the pronoun ‘those,’ then conveyed the relevant concept ‘using an oral method only without sign language at all,’ and finished the English sentence with the specific number and its referent (‘was 5 countries’). Interestingly, she unpacked the rich meaning of his four signs only oral sign ban as the long English phrase (‘using an oral method only without sign language at all’) to maintain accuracy and convey his emphasis. As a result, her tail-to-tail span for this Auslan sentence became longer than before, at 3.5 seconds. In summary, Amber’s average onset processing time for Auslan sentences 85 to 88 (3.9 seconds) was shorter than Liz’s (4.7 seconds). Although both Liz and Amber produced highly accurate English interpretations, their strategies differed considerably. While Liz often expressed the number before both its referent and its concept in her English interpretations (e.g., ‘Thirty-one countries have total communication …’), Amber consistently introduced the concept first by using an indefinite noun phrase (e.g. ‘And the countries that …’) as the subject and put the number and its referent (e.g., ‘31 countries’) at the end of her English renditions. Typically using onset processing time of approximately 4 seconds (about half of each Auslan sentence), Amber generally followed the Deaf

176 Cognitive Overload and Coping Strategies presenter’s logical thread by introducing the topic – countries having what type(s) of education – and finishing off the English interpretations with a number and its referent – a specific number of countries. This strategy prevented Amber from lagging excessively far behind the Deaf presenter when starting off and wrapping up her English renditions, thus avoiding working memory overload and cognitive overload. 6.3.2 Pseudo-Cleft Structure and Numbers This section presents local analysis of two interpreters’ (Jane and Alex) English renditions of a sequence of six Auslan sentences (96, 97, 98, 99, 100, and 101) featuring numbers, dense information, and pseudo-cleft structures (also referred to as rhetorical questions or question-and-answer structures). Both the first and last Auslan sentences (96 and 101) contain a number in the middle of the respective sentence, whereas each of the other four Auslan sentences (97, 98, 99, and 100) is a rhetorical question that ends with a number. The numbers in these four adjacent Auslan sentences are logically related to each other and show a descending order. I selected this Auslan segment for the in-depth analysis of cognitive overload and coping strategies not only because it laid the foundation for the Deaf presenter’s final conclusion, but also because it proved to be highly challenging for the 20 professional interpreters involved in this study. As shown in Table 6.5, only one interpreter (Alex) rendered all the six adjacent Auslan sentences accurately into English; and another five interpreters (Liz, Amber, Emily, Jane, and Helen) transferred five of the six Auslan sentences correctly into English. Given the precise information regarding the Auslan numbers and the logical relationship among the numbers, interpreters needed to not only convey the messages accurately but also maintain coherence in the target speech segment for the non-signing audience to understand the Deaf professional’s key research findings. 6.3.2.1 Coping Strategies, Self-correction, and Cognitive Overload Example 6.5 illustrates Jane’s English interpretations of Auslan sentences 96 to 101, revealing her various coping strategies for the Auslan pseudo-cleft structures, possible reasons for her inaccurate rendition of Auslan sentence 101, as well as her fluctuating processing time from one juncture to another. Jane’s average onset processing time of the six Auslan sentences was 3.4 seconds. Overall, Jane followed the Deaf presenter’s thread of thought by introducing the topic first, then specifying the number near or at the end of her English sentences while maintaining the logical relationship among the numbers. She used different strategies to interpret the four Auslan sentences (97, 98, 99, and 100) correctly. She made a self-correction when interpreting Auslan sentence 98, and produced an acceptable English rendition whose meaning deviates slightly from that of the Deaf presenter. Jane, with short onset processing time, restructured Auslan rhetorical questions 97 and 98 into English statements (i.e.,

Cognitive Overload and Coping Strategies 177

Example 6.5  Jane’s coping strategies (for numbers and syntactical differences) and cognitive overload

178 Cognitive Overload and Coping Strategies using the third strategy; see Section 6.3); however, she then adhered to the question-and-answer structure (i.e., use of the fourth strategy) when processing Auslan rhetorical questions 99 and 100. Her change of strategies seems to be related to her perception of increased cognitive load, which resulted in her English interpretation for Auslan sentence 101 being ungrammatical. Her cognitive processing that led up to the cognitive overload is described here. Jane was still interpreting Auslan sentence 94 (think-about concept access relate-to interpreters) when the Deaf presenter started to produce Auslan sentence 95. With onset processing time of 5.4 seconds, she started interpreting Auslan sentence 95 when the Deaf presenter was producing the final sign of this sentence (how many), which also occurred at the beginning of this sentence and refers to ‘how many countries.’ Given that the Deaf presenter did not sign the referent (countries) of how many, Auslan sentence 95 may be vague for some interpreters. Jane misinterpreted Auslan sentence 95 as she not only omitted the important information at the beginning of this sentence (well ask how many deaf people can) but also unjustifiably added a link between the understood information in this sentence (access t-o government, which she rendered as ‘Access to government’) and a piece of information in Auslan sentence 94 (interpreters, which she transferred as ‘through interpreters’). Her misinterpretation of Auslan sentence 95 may have occurred due to her allocating much attention to producing the English rendition of Auslan sentence 94 while paying insufficient attention to comprehending the initial part of Auslan sentence 95. Nevertheless, Jane was not affected by her miscue and produced an accurate English interpretation of the subsequent Auslan sentence (96). With short onset processing time of 1.7 seconds, Jane started interpreting Auslan sentence 96 after seeing its second sign – eighty(-two) and said ‘Altogether,’ which appears to be a filler (i.e., stalling strategy). After rendering eighty-two accurately into English (with processing time of 1.6 seconds), she paused for 1.2 seconds, perhaps to think about its referent and wait for/analyse more source language information, then resumed to complete her English sentence (‘countries said that Deaf people had access to government’). With short onset processing time of 1.6 seconds, Jane started interpreting Auslan sentence 97 when the Deaf presenter was producing the sign interpreter in the first meaning unit of this sentence (but what about sign language interpreter services). She reframed the Auslan question to introduce the topic as ‘But when I asked about interpreting services.’ She waited for half a second to analyse the incoming information, then reformulated the second meaning unit (have have have? what? provide have?) succinctly as ‘and what was provided.’ When uttering these English words, she saw the Auslan number sixty-two. Hence, she transferred the Auslan number into English, conveyed its referent (countries) as ‘countries,’ then completed the English sentence by strategically adding the verb and object ‘said that they had that access’ to explicate the implied information in the source text. Her processing time for sixty-two was 1.3 seconds.

Cognitive Overload and Coping Strategies 179 With onset processing time of 2 seconds, Jane started interpreting Auslan sentence 98 when the Deaf presenter was producing the third sign of this ­sentence – interpreting. She provided a literal interpretation (‘So, 62 countries have interpreting services’) of the repetition so sixty-two interpreting services have, to convey the logical relationship between the two adjacent Auslan sentences. Note that she strategically added the referent (‘countries’) of the number sixty-two to ensure grammaticality. She then paused for 3.2 seconds to analyse the incoming information – those finish have training qualified. She rendered but how many those as ‘but out of those’ to highlight the logical link, but she omitted the question sign how many. By this time, she had seen the whole meaning unit finish have training qualified professional interpreters, which she misunderstood and misinterpreted as ‘I wanted to know how many of those interpreters have professional qualifications.’ This English interpretation is erroneous, as the Deaf presenter meant how many of those 62 countries (rather than how many of those interpreters) have trained, qualified, and professional signed language interpreters. After understanding the Deaf presenter’s answer only forty-three at the end of Auslan sentence 98, Jane rendered it in English as ‘And only 43 countries.’ Interestingly, here she strategically added the implied referent (‘countries’) of the number forty-three, indicating that she realised that the previous sign how many those meant how many countries instead of how many interpreters. Under much cognitive pressure, as indicated by her filled pause ‘uh,’ she quickly corrected her error by saying ‘had professional qualifications for sign language interpreters available.’ Note that this English rendition differs slightly from the Deaf presenter’s Auslan message in terms of meaning and emphasis. Regarding Auslan sentence 98, Jane’s interpretation indicates that only 43 countries had formal interpreting training programs and/or national certification systems available to signed language interpreters, but the Deaf presenter meant that only 43 countries had trained, qualified, and professional signed language interpreters. Her processing time for forty-three was 1.4 seconds. Due to her self-correction, Jane lagged quite far behind the Deaf presenter. However, luckily, while he was reiterating the essential idea of Auslan sentence 98 at the beginning of Auslan sentence 99 (that forty-three group those), she was making use of the redundancy to complete her English interpretation of Auslan sentence 98. With onset processing time of 4.3 seconds, Jane started interpreting Auslan sentence 99 by rendering the repetition literally and faithfully as ‘Out of those 43,’ to reflect the logical connection between the two adjacent Auslan sentences. Jane then waited for 1.2 seconds to receive more information of the Deaf presenter’s fingerspelling (c-o-d-e-o-f-e-t-h-i-c-s) in Auslan sentence 99. At the point of the filled pause ‘er’ (possibly indicating her heavy cognitive load when planning her English sentence and understanding his fingerspelling at the same time), she probably deciphered the words that he had fingerspelled (‘code of ’) and anticipated the remaining word (‘ethics’) by drawing on her linguistic and contextual knowledge. Therefore, she resumed her interpretation by

180 Cognitive Overload and Coping Strategies saying ‘how many had a code of ethics,’ then strategically added the referent ‘for their interpreters’ to clarify the meaning and enhance the coherence of her English interpretation. She provided the full answer ‘The number in response to that question was only 31,’ probably to ensure both intra-textual coherence and formal register at the conference setting. She took advantage of both the Deaf presenter’s 1.8-second hold on the number sign one and his 1.1-second pause between Auslan sentences 99 and 100 to catch up. Her processing time for thirty-one was 4.4 seconds. With onset processing time of 5.7 seconds, Jane started interpreting Auslan sentence 100 when the Deaf presenter had produced two meaning units – important part that group look-over-there look-here and government responsible pay interpreters. After making an unjustifiable minor addition regarding inter-sentential logical relationship (‘And out of those’), she introduced the new idea in this Auslan sentence by posing an English question ‘how many er had interpreters paid by their, um, interpreters paid by their government?’ Both her filled pauses (‘er’ and ‘um’) and repetition ‘interpreters paid by their’ indicate that she was close to or was experiencing cognitive saturation, very likely because she had difficulties in recalling the critical sign government, as her processing time for this sign was very long, at 9.8 seconds. Her syntactic choice to use passive voice in English (‘interpreters paid by their’) also meant that the subject (‘government’) had to be delayed, thus increasing the risk of experiencing working memory overload and cognitive overload. For unknown reasons, she recalled the sign government successfully and produced a complete question. At this point in time, he was well into producing the following Auslan sentence (101). Jane paused for 0.8 second and rendered thirty-two in Auslan sentence 100 as a short answer ‘That was only 32,’ with the number processing time being 6.1 seconds. Note that she did not add the referent of this number perhaps due to high cognitive load and the need to catch up. Despite the long processing time, Jane rendered the number thirty-two accurately, maybe due to her exceptionally good working memory capacity, semantic memory, and visuospatial memory, among other factors. Note that Jane’s English working memory capacity score was 89 out of 108 (82%), and her Auslan working memory capacity score was 102 out of 108 (94%) (for further information about the working memory span tasks, see Wang, 2013, 2016). Interestingly, the accumulation of cognitive load, long processing time, and fatigue appeared to have resulted in Jane producing an ungrammatical English interpretation for Auslan sentence 101, which summarises this whole section (Auslan sentences 94 to 100).With onset processing time of 5.2 seconds, she rendered the first Auslan meaning unit (s-o talk about sign language interpreting service have sixty-two) as two comprehensible English clauses (‘So, to have [0.8-second pause] sign language interpreters, there were 62 countries’). Note that Jane strategically added the implicit referent (‘countries’) of the number sixty-two. Her processing time for sixty-two was slightly long, at 5.4 seconds. While Jane was saying those English clauses, she saw the second meaning unit in Auslan sentence 101 (those thirty-two government responsible pay). She

Cognitive Overload and Coping Strategies 181 formulated an ungrammatical English rendition for this Auslan meaning unit by saying ‘but to have paid by the government, sign language interpreters were only 32,’ with her processing time for thirty-two being quite long, at 7.1 seconds.This grammar error may be ascribed to several factors: (i) Jane’s long processing time caused cognitive overload; (ii) she probably experienced fatigue, as she had interpreted for 15 minutes of the 17-minute Auslan video; (iii) her strategic addition of the implied referent (‘countries’) of sixty-two may have left her with inadequate attention for processing the subsequent meaning unit; and (iv) the Deaf presenter used concise Auslan (those thirty-two government responsible pay) to express his meaning, but the interpreter needed to not only add the implicit referent of the number thirty-two but also make explicit the implied object (‘sign language interpreters’) of the verb pay. In other words, Jane probably needed more processing capacity and more processing time than what was available to her at the juncture in order to deal with the final Auslan sentence of this long and challenging segment. Had the Deaf presenter had paused intentionally after signing the aforementioned first meaning unit of Auslan sentence 101 (or had paused deliberately between Auslan sentence 100 and Auslan sentence 101), Jane could have had more time to re-structure the information of the second meaning unit of Auslan sentence 101 and to produce a grammatical English rendition. In addition, had Jane had a team interpreter to assist her, she could have rendered Auslan sentence 101 accurately, grammatically, and idiomatically in English. In summary, both Jane’s onset processing time and number processing time generally were within 2 seconds when she used the third strategy (i.e., begin interpreting after understanding a few initial signs of an Auslan sentence, use appropriate strategies such as stalling or pausing to receive more Auslan information, and finally say the number near or at the end of the English sentence) to interpret Auslan sentences 97 and 98. It is her self-correction (in an attempt to ensure high accuracy) in her English interpretation of Auslan sentence 98 that led to the long tail-to-tail span for this sentence, increased her onset processing time of Auslan sentences 99, 100, and 101, and ultimately resulted in an ungrammatical English rendition of Auslan sentence 101. As her onset processing time increased as the Auslan presentation continued, she experienced heavier cognitive load and chose to adhere to the Deaf presenter’s rhetorical question structure (i.e., using the fourth strategy to cope with Auslan sentences 99 and 100), probably in order to save target language Production Effort, avoid cognitive overload, and catch up with the Deaf presenter. Example 6.5 shows that the interpreter’s comprehension of the source text, aspiration to maintain coherence in the target text, choice of particular target language syntactic structures, coping strategies for dealing with challenges such as numbers and syntactic differences between the source language and the target language, self-correction of an interpretation error, adjustment of processing time, cognitive load, use of formal register in the target language, and accuracy of interpretation are all interrelated in complex and dynamic ways during signed-to-spoken language simultaneous interpreting.This example also illustrates that, even when dealing with Auslan sentences that have the same syntactic

182 Cognitive Overload and Coping Strategies structure (e.g., rhetorical questions ending with numbers), the interpreter’s coping strategies varied from one sentence to another, depending on a range of factors such as processing time at a particular point in time, decision-making, consequence of a previous interpreting strategy, availability of target language expression, and fatigue. Moreover, Example 6.5 shows that the interpreter’s miscomprehension of a source message, maybe due to ambiguity in the source text, triggered a chain effect in simultaneous interpreting, namely misinterpretation, self-correction, lagging quite far behind the source language speaker/signer, an increase in cognitive load, use of strategies to catch up, and ungrammatical rendition of a subsequent source message. Last but not least, the in-depth analysis of this example reveals that long processing time, multitasking in simultaneous interpreting, and the accumulated cognitive load in a dense and challenging section ultimately resulted in cognitive overload and a serious interpretation miscue (ungrammatical rendition). On the positive side, the interpreter not only made strategic additions and strategic omissions to enhance the effectiveness of her Auslan-to-English simultaneous interpretations, but also adjusted her processing time to avoid cognitive overload at certain junctures during simultaneous interpreting. This local analysis and these new findings add to the existing literature on processing time, signed language interpreting, simultaneous interpreting process, strategies, cognitive overload, and simultaneous interpreting performance. 6.3.2.2 Literal Interpretation, Strategic Additions, and Unjustifiable Omissions This section presents local analysis of Alex’s English interpretations of Auslan sentences 96 to 101 (1 minute, 19 seconds), highlighting the fact that interpreters’ strategic additions sometimes lead to significant omissions of incoming information. Alex’s English renditions were selected for the detailed analysis because he was the only interpreter in this study who rendered all six Auslan sentences correctly into English (see Table 6.5). Regarding this segment, his onset processing time ranged from 2.6 seconds for Auslan sentence 97 to 6.2 seconds for Auslan sentence 98, with an average of 4.4 seconds. Despite his long processing time and elaboration at several junctures, he avoided cognitive overload. The in-depth analysis reveals that Alex’s success was primarily due to his frequent use of a literal interpretation approach (i.e., retaining the question-andanswer structure), his exceptionally good working memory capacity which was indeed useful when he was lagging far behind the Deaf presenter at particular points in time, and the Deaf presenter’s pauses and repetition of numbers at certain places. In other words, the Deaf professional in this study is a good presenter who is very experienced with working with signed language interpreters. It appears that Alex’s long processing time at several junctures, determination to achieve high accuracy and maintain coherence, strategic addition in his English interpretation of Auslan sentence 101, time pressure, and the absence of look-pause-nod cooperation strategies between the Deaf presenter and the

Cognitive Overload and Coping Strategies 183 interpreter due to the artificial testing condition collectively led to the cognitive overload that he experienced when processing Auslan sentences 102 and 103, which were just after the dense and challenging segment (Auslan sentences 96 to 101). As a result of the cognitive overload, he omitted an important detail in Auslan sentence 102 and left out the entire Auslan sentence 103. Example 6.6 illustrates Alex’s cognitive process of adopting a literal interpretation approach to successfully transfer the meaning of Auslan sentences 96 to 101 into spoken English. This example also reveals how his decision-making and interpreting behaviour gave rise to the cognitive overload and unjustifiable omissions in the end. With onset processing time of 3.7 seconds, Alex started interpreting Auslan sentence 95 after he analysed the first half of this sentence (well ask how many deaf people can access). While rendering this message as an English question (‘How many Deaf people can access’), he saw the second half of this Auslan sentence (t-o government how many?). He expanded government into ‘government services’ (an example of strategic addition) in order to make the Deaf presenter’s meaning clear to the hearing audience. Note that the Deaf presenter omitted the referent (countries) when he produced the question sign how many and thus caused ambiguity. As a result, Alex misunderstood how many as ‘How many Deaf people.’ A correct English interpretation can be ‘How many countries responded that Deaf people can access government services?’ With onset processing time of 3 seconds, Alex began rendering the first meaning unit of Auslan sentence 96 (respond eighty-two country) into idiomatic English (‘And we received responses from 82 countries’). Compared with the first person pronoun ‘I’ used by both Lauren (in Example 6.2) and Jane (in Example 6.5), Alex’s use of the second person pronoun ‘we’ is more effective because it refers to the research team rather than just the Deaf presenter. While Alex was saying that clause, he saw the second meaning unit (say yes deaf people can access). Although his English rendition of the second meaning unit (‘said, “Yes,” they had access to government’) was accurate, it was ungrammatical (‘said’ should be changed to ‘that said’ or ‘who said’), contained a direct speech answer (‘Yes’) which sounded informal, and included a potentially ambiguous personal pronoun ‘they’ which actually referred to the explicit subject deaf people in this Auslan sentence. Some listeners may mistake the pronoun ‘they’ for the nearby noun phrase ‘eighty-two countries.’ Alex omitted the Deaf presenter’s comment (wow). With onset processing time of 2.6 seconds, Alex embarked on interpreting Auslan sentence 97 after understanding the message (but what about sign language interpreter services). Again, he transferred the Auslan message as an English question (‘What about the provision of sign language interpreting services?’). Alex then strategically omitted the Deaf presenter’s Auslan questions (have have have? what? provide have?) and paused for 1.5 seconds to wait for the answer. After seeing the number sixty-two, he resumed interpreting by saying it in English (‘Sixty-two’). He then rendered its referent country as

184 Cognitive Overload and Coping Strategies

Example 6.6  Alex’s use of literal interpretation to cope with numbers, rhetorical questions, and dense information

Cognitive Overload and Coping Strategies 185 ‘countries’ and strategically added a concise answer (‘responded,“Yes.”’). Linking this answer to the English interpretations of the previous two Auslan sentences, the hearing audience could understand that this affirmative referred to the message that 62 countries responded that they had signed language interpreting services. Alex decided to delay his interpretation when he saw the Deaf presenter’s reiteration (so sixty-two interpreting services have) in Auslan sentence 98. With onset processing time of 6.2 seconds, he started interpreting this Auslan sentence by saying ‘Of those 62,’ then paused for 1.7 seconds before strategically adding the referent ‘countries’ of the number sixty-two in order to provide a complete message. The pause was probably because Alex was devoting much attention to understanding the Auslan information training qualified while thinking about the missing referent of sixty-two. Note that Alex strategically omitted the Deaf presenter’s redundant information in Auslan sentence 98 (interpreting services have) probably to save time and attention. Once Alex understood the important message but how many … training qualified professional interpreters in Auslan sentence 98, he again rendered it as an English question (‘did they have formal interpreter training programs producing [1.9-second pause] qualified professional interpreters?’). His expansion (another type of strategic addition) of training into ‘formal interpreter training programs’ made the Deaf presenter’s meaning immediately comprehensible to the hearing audience. The 1.9-second pause between ‘producing’ and ‘qualified’ is also quite interesting, because Alex was comprehending and memorising the number forty-three while trying to finish his English question. The aforementioned strategic addition and the two pauses increased Alex’s processing time, with the tail-to-tail span for the key concept professional interpreters (the time interval between the ending time of the Auslan sign interpreters and the ending time of its English interpretation ‘interpreters’) being 7 seconds. In Auslan sentence 99, the Deaf presenter’s repetition of the number fortythree triggered or refreshed Alex’s memory of the same number mentioned at the end of Auslan sentence 98. Consequently, Alex quickly provided a succinct answer (‘Forty-three of them did’) to complete his English interpretation for Auslan sentence 98. With onset processing time of 4.3 seconds, Alex began interpreting Auslan sentence 99 by conveying the reiterated number forty-three and strategically adding its referent (‘Of those 43 countries’) while reading the initial letters of the Deaf presenter’s fingerspelling. Although Alex again retained the Auslan question structure (‘did those interpreters have a code of ethics?’), he strategically added the subject ‘those interpreters,’ which not only linked to the previous concept (‘professional interpreters’) but also made it easy for the naïve hearing audience to understand the implicit referent of ‘code of ethics.’ While Alex was articulating ‘code of ethics,’ he was understanding and memorising a new number (thirty-one). The Deaf presenter paused for 1.8 seconds by holding on the final sign of Auslan sentence 99 (one), during which Alex also paused for 1.1 seconds.

186 Cognitive Overload and Coping Strategies Alex  then provided a full answer to that question (‘Thirty-one countries did have a code of ethics that interpreters abided by’). This expanded answer contains multiple strategic additions; first, the referent (‘countries’) for the number thirty-one, second, the aforementioned idea associated with the number (‘did have a code of ethics’), and third, the attributive clause ‘that interpreters abided by.’ The excellent lexical choice (‘abided by’) reflected formal register in English, which suited the conference setting. It is this full answer that delayed Alex’s English interpretation of the subsequent Auslan sentence (100) and increased the cognitive load. With onset processing time of 6 seconds, Alex started to interpret Auslan sentence 100. Given that the initial part (important part that group lookover-there look-here) was of lesser importance than the subsequent message (government responsible pay interpreters have?), this may have mitigated the negative impact of Alex’s long onset processing time. He rendered the understood message as an English question (‘Was the government responsible for the payment of interpreting services and interpreter salaries?’). It is worthy of attention that Alex elaborated on the short Auslan phrase pay interpreters in Auslan sentence 100 (by saying ‘the payment of interpreting services and interpreter salaries’) in order to make the meaning accessible to the hearing audience who were unfamiliar with Deafness-related issues such as sign language interpreting services. He lagged quite far behind the Deaf presenter, with the onset processing time for the subject government (the interval between the starting point of this sign and the starting point of its English equivalent) being 4.6 seconds and the tail-to-tail span for pay interpreters (the interval between the ending point of interpreters and the ending point of ‘salaries’) being 7 seconds. Alex then quickly provided the concise answer ‘Thirty-two responded, “Yes.”’ He gained some time to catch up with the Deaf presenter by both strategically omitting the redundant information (how many country what?) and making using of the presenter’s 1.6-second pause following thirty-two. In Auslan sentence 101, the Deaf presenter summed up the aforementioned sentences by saying that 62 countries reported having sign language interpreting services but only 32 of them had governments that were responsible for paying for the services. He expressed the latter message succinctly by using five signs (those thirty-two government responsible pay). Here, both the Deaf presenter’s concise Auslan and Alex’s strategic additions for enhancing intratextual coherence and maximising audience design led to Alex lagging quite far behind the Deaf presenter. With onset processing time of 4.3 seconds, Alex began interpreting Auslan sentence 101 after he understood the complete meaning unit (s-o talk about sign language interpreting service have sixty-two). Due to the sufficient processing time, Alex rendered the meaning into idiomatic English (‘So, there are 62 countries that have sign language interpreting services’). While producing that English clause, Alex saw the Deaf presenter’s subsequent signs thirty-two government. Alex then strategically added the

Cognitive Overload and Coping Strategies 187 transition word ‘yet’ to make the implied logical link explicit. He made two strategic additions in his utterance ‘only 32 of those 62 respondents have governments that are paying for interpreting services,’ one being the reiteration of the previously mentioned concept (‘62 respondents’) and the other being the implicit object (‘interpreting services’) for the verb pay. While both strategic additions made the target language renditions coherent and comprehensible to the non-signing audience, they made Alex lag quite far behind the Deaf presenter, with the tail-to-tail span for Auslan sentence 101 being 5.3 seconds. At an authentic conference, if the Deaf presenter saw that Alex was still interpreting, the Deaf presenter would probably have paused between Auslan sentence 101 and Auslan sentence 102 to give Alex sufficient time to finish his English interpretation of Auslan sentence 101. When Alex completed interpreting Auslan sentence 101, the Deaf presenter already finished producing the first two meaning units of Auslan sentence 102 (mean that group who responsible and could b-e deaf people have-to pay interpreters). With onset processing time of 5.6 seconds, Alex started interpreting Auslan sentence 102 by quickly rendering those two understood meaning units as two new questions (‘So, the basic question is who is paying for interpreting services?’ and ‘Deaf people?’).The Deaf presenter’s elongation of producing the sign for or may have given Alex some time to plan that target language output. Alex spoke fast and used a concise expression (‘Deaf people?’) in an attempt to catch up. While Alex was articulating those two English questions, he saw the third Auslan meaning unit of Auslan sentence 102 (interpreters heart volunteer interpret). Interestingly, here, the Deaf presenter used two signs from International Sign (heart and volunteer) for the concept ‘volunteer.’ Alex provided a faithful and idiomatic English interpretation of this Auslan message and conveyed the meaning of both synonymous signs (‘Are interpreters doing it [1.7-second pause um] voluntarily, out of the goodness of their heart?’). While Alex was producing this highly accurate and immediately comprehensible English rendition, the Deaf presenter was signing two new Auslan messages (mean link no qualification// 103. that very vague vague vague), both of which Alex omitted. Alex paused for 1.7 seconds before saying ‘voluntarily,’ possibly because he was considering how to express the iconic sign heart in idiomatic English, or because he was trying to understand the less familiar sign volunteer, or because he was having difficulties with coordinating source language comprehension and target language production. When Alex finished interpreting Auslan sentence 102, he refocused his attention and started to render Auslan sentence 104 and did not appear to have been affected by his unjustifiable omissions of the aforementioned two Auslan messages. Without interviewing Alex on these two unjustifiable omissions, it is unclear if he was conscious or unconscious of them. Interestingly, Alex’s processing time for the eight Auslan numbers in this long segment varied considerably. His processing time for eighty-two in Auslan ­sentence 96 was 3.7 seconds; that for sixty-two in Auslan sentence 97 was

188 Cognitive Overload and Coping Strategies 1 second; and that for sixty-two in Auslan sentence 98 was 6.1 seconds. His processing time for forty-three in Auslan sentence 98 was 5.1 seconds; that for forty-three in Auslan sentence 99 was 4.1 seconds; that for thirty-one in Auslan sentence 99 was 2.2 seconds; and that for thirty-two in Auslan sentence 100 was 4.4 seconds. His processing time for sixty-two in Auslan sentence 101 was 1.7 seconds; and his processing time for thirty-two in the same sentence was 3.2 seconds. Compared with Lauren’s consistently short processing time regarding numbers (approximately 1.5 seconds) in Example 6.2, Alex’s processing time regarding numbers in Example 6.6 ranged from 1 second to 6.1 seconds. These results indicate that processing time not only varies from one measurement point to another but also differs from one interpreter to another, without a clear-cut or consistent pattern. Possible explanations for these findings include the dynamic nature of simultaneous interpreting process, as well as the fact that interpreters vary from one another in terms of processing capacity, working memory capacity, use of strategies, bilingual proficiency, familiarity with the Deaf professional, and familiarity with this particular Auslan presentation. These findings reflect both the uniqueness and diversity of professional interpreters and indicate both the uniqueness and diversity of their cognitive processing in simultaneous interpreting. In summary, when Auslan sentences each featuring a pseudo-cleft structure and containing a number near or at the end occurred in a row (e.g., Auslan sentences 95 to 101), very few professional Auslan/English interpreters in the current study rendered the whole Auslan segment accurately into English. Due to Jane’s misunderstanding of an Auslan sentence in the middle of this segment, she produced an erroneous English interpretation and then corrected it. As a result of the selfcorrection, she lagged quite far behind the Deaf presenter, encountered cognitive overload, and produced an ungrammatical English sentence at the end of this segment. Prior to the self-correction, Jane generally produced free and idiomatic English interpretations of Auslan pseudo-cleft structures. After the self-correction, however, she decided to adhere to the Deaf presenter’s pseudo-cleft structures and produced literal interpretations, in order to save attentional resources and catch up with the Deaf presenter. By contrast, Alex interpreted the whole Auslan segment correctly into English, by adhering to the Deaf presenter’s q­ uestion-and-answer structures and producing literal interpretations. Nevertheless, Alex made unjustifiable omissions of some messages in the subsequent Auslan sentences, partially because his strategic additions caused him to lag far behind the Deaf presenter and ultimately experience cognitive overload.

6.4 Summary This chapter has presented findings regarding signed language interpreters’ cognitive overload and their coping strategies for numbers in signed-to-spoken language simultaneous interpreting. Firstly, the chapter has reported quantitative results regarding the 20 professional Auslan/English interpreters’ accuracy rate of

Cognitive Overload and Coping Strategies 189 interpreting 32 Auslan numbers and their accuracy rate of interpreting 29 Auslan sentences that contained those numbers (see Section 6.1). Results showed that the interpreters’ mean accuracy rate of rendering the individual Auslan numbers (89%) was considerably higher than their mean accuracy rate of interpreting the respective Auslan sentences (60%). In other words, many interpreters rendered Auslan numbers accurately into English, but they did not transfer the core meaning of the respective Auslan sentences accurately into English. In addition, when Auslan sentences each containing a number occurred in a row, the interpreters’ accuracy rate of interpreting the individual Auslan sentences was markedly higher than their accuracy rate of interpreting the whole Auslan segment. Typically, only a few interpreters rendered the whole Auslan segment, in which each Auslan sentence included a number, correctly into English. While some Auslan numbers were located in the middle of Auslan sentences, other Auslan numbers appeared near or at the end of Auslan sentences that featured either topic-comment structures or pseudo-cleft structures (i.e., rhetorical questions or question-and-answer structures). Furthermore, this chapter has presented findings from local (qualitative) analysis of five professional interpreters’ overall accurate English renditions of some adjacent Auslan sentences which featured multiple salient challenges such as numbers, dense information, and syntactical differences between Auslan and English. The in-depth local analysis has revealed the interpreters’ various coping strategies to deal with those difficulties, potential reasons for imperfect renditions, impact of long or short processing time on interpreting performance, and complex interrelations among cognitive processing, strategy use, and simultaneous interpreting performance. The local analysis revealed that although different professional interpreters produced effective renditions of the same source language segment, their strategies often varied from person to person and changed from sentence to sentence, highlighting diversity and creativity among the interpreters. While one interpreter (Jane in Example 6.1) generally adopted a predominantly free interpretation approach to cope with some adjacent Auslan sentences each containing a number in the middle, another interpreter (Lauren in Example 6.2) consistently employed a predominantly literal interpretation approach and short onset processing time to deal with those Auslan sentences. Despite these differences, both interpreters produced accurate English renditions of those Auslan sentences. An interesting finding is that interpreters’ prior knowledge and biases sometimes affected the accuracy of their Auslan-to-English simultaneous interpretations; therefore, they needed to suppress their biases and concentrate on understanding source language messages. A key finding is that, when numbers occurred in adjacent Auslan topiccomment structures or a series of pseudo-cleft sentences, the professional Auslan/English interpreters primarily employed four types of strategies to cope with multiple challenges in those Auslan sentences such as numbers, dense information, and syntactical differences between Auslan and English. These strategies include: (i) using quite long onset processing time and producing idiomatic

190 Cognitive Overload and Coping Strategies English renditions, (ii) using medium onset processing time coupled with other suitable strategies such as repeating a particular part of the topic, stalling, and waiting, (iii) using short onset processing time together with other strategies such as repeating a certain part of the topic, stalling, and waiting, as well as (iv) adopting a predominantly literal interpretation approach to retain the questionand-answer structure. Some interpreters’ overuse of the predominantly literal interpretation approach did not match the Deaf professional’s academic register at the formal conference setting and led to ungrammatical English renditions. A constellation of factors (such as long processing time; strategic additions to ensure accuracy, grammaticality, coherence, and comprehensibility of target language renditions; insufficient time to interpret; no opportunity to cooperate with the Deaf presenter during the Auslan presentation [e.g., through eye contact, hand gesture] to control the pace of the Auslan presentation; no team interpreter; and no visual access to projected PowerPoint slides) occasionally culminated in professional Auslan/English interpreters’ cognitive overload and unjustifiable omissions. Interpreters’ self-corrections and verbose target language expressions sometimes also led to cognitive overload and interpretation errors. It appears that the Deaf presenter’s omissions of the referents (e.g., country) of some Auslan signs (e.g., how many, sixty-two) caused ambiguity in the source text, which led to some interpreters’ miscomprehension of the source messages. Interpreters’ misunderstanding of the source language information sometimes led to a chain effect of misinterpretation, self-correction, and cognitive overload. The local analysis has proved to be an effective method for providing insight into the generally hidden and mysterious cognitive process of simultaneous interpreting and for establishing the complex links between interpreting process and interpreting product. I will discuss the findings of this chapter in Chapter 7.

Notes 1 Note that three Auslan sentences in Table 6.1 (30, 54, and 101) included more than one number. 2 In his previous sentence, the Deaf presenter mentioned that he wanted to explore from 88 countries how many were using the different education methods – bilingual education, total communication, and oral education – to educate Deaf children in Deaf schools.

References Cokely, D. (1986). The effects of lag time on interpreter errors. Sign Language Studies, 53(Winter), 341–375. doi: 10.1353/sls.1986.0025. Cokely, D. (1992). Interpretation: A Sociolinguistic Model. Burtonsville, MD: Linstok Press. Desmet, B., Vandierendonck, M., & Defrancq, B. (2018). Simultaneous interpretation of numbers and the impact of technological support. In C. Fantinuoli (Ed.), Interpreting and Technology (pp. 13–27). Berlin: Language Science Press. Gile, D. (2008). Local cognitive load in simultaneous interpreting and its implications for empirical research. Forum, 6(2), 59–77. doi: 10.1075/forum.6.2.04gil.

Cognitive Overload and Coping Strategies 191 Gile, D. (2009). Basic Concepts and Models for Interpreter and Translator Training (Rev. ed.). Amsterdam/Philadelphia: John Benjamins. Johnston, T., & Schembri, A. (2007). Australian Sign Language (Auslan): An Introduction to Sign Language Linguistics. Cambridge, UK: Cambridge University Press. Lamberger-Felber, H. (2001). Text-oriented research into interpreting. Examples from a case study. Hermes, Journal of Linguistics, 26, 39–64. doi: 10.7146/hjlcb.v14i26.25638. Morgan, E. F. (2008). Interpreters, conversational style, and gender at work. In P. C. Hauser, K. L. Finch, & A. B. Hauser (Eds.), Deaf Professionals and Designated Interpreters: A New Paradigm (pp. 66–80). Washington, DC: Gallaudet University Press. Napier, J. (2016). Linguistic Coping Strategies in Sign Language Interpreting.  Washington, DC: Gallaudet University. Nicodemus, B., & Emmorey, K. (2013). Direction asymmetries in spoken and signed language interpreting. Bilingualism: Language and Cognition, 16(3), 624–636. doi: 10.1017/S1366728912000521. Taylor, M. M. (2002). Interpretation Skills: American Sign Language to English. Edmonton, AB: Interpreting Consolidated. Wang, J. (2013). Bilingual working memory capacity of professional Auslan/English interpreters. Interpreting, 15(2), 139–167. doi: 10.1075/intp.15.2.01wan. Wang, J. (2016). The relationship between working memory capacity and simultaneous interpreting performance: A mixed methods study on professional Auslan/English interpreters. Interpreting, 18(1), 1–33. doi: 10.1075/intp.18.1.01wan.

7 Cognitive Overload and Coping Strategies Regarding End Negation

This chapter continues to address the second and third research questions of the study but focuses on the 20 professional Auslan/English interpreters’ English renditions of a sequence of Auslan sentences containing end negation rather than numbers. These adjacent Auslan sentences posed two challenges to the interpreters: (i) syntactical differences between Auslan and English, and (ii) the need for the interpreters to use sufficiently long onset processing time in order to understand the meaning of the negated Auslan sentences and change the order of sentence components in order to produce grammatical and idiomatic English renditions. First, this chapter reports on quantitative results regarding the interpreters’ accuracy rate of rendering the adjacent Auslan sentences containing end negation into English. Second, the chapter uses representative examples from the Auslan-to-English simultaneous interpretation corpus to illustrate strategies employed by the interpreters in order to cope with the adjacent Auslan sentences ending with negation. Third, this chapter discusses findings from both Chapter 6 and this chapter in relation to previous relevant literature. Finally, the chapter summarises the key points.

7.1 Cognitive Overload and Sentence-Level Accuracy Regarding End Negation This section applies local analysis to throw light on interpreters’ cognitive process and coping strategies in relation to a sequence of seven adjacent Auslan sentences (from 105 to 111; see Appendix B). Five of these Auslan sentences (105, 107, 108, 110, and 111) were pseudo-cleft structures (i.e., rhetorical questions or question-and-answer structures) that featured negation near or at the end of the sentences; and pseudo-cleft structures are often used for emphasising an important point (Johnston & Schembri, 2007). Interpreters needed to use sufficiently long onset processing time and restructure the components of these Auslan sentences in order to produce idiomatic and comprehensible English renditions. While the Deaf presenter completed Auslan sentence 107 with a noticeable headshake, he ended Auslan sentences 108, 110, and 111 with an obvious negating sign no. These Auslan sentences (whose key points were presented on PowerPoint slide 18) concluded the Deaf professional’s Auslan presentation by

Cognitive Overload and Coping Strategies 193 summarising his key findings based on the numbers mentioned in Section 6.1 of Chapter 6. As a result, it is pivotal for interpreters to convey the essential meaning of each of these Auslan sentences accurately so that the hearing ­audience would understand the Deaf presenter’s gist of the whole Auslan presentation. This segment of seven adjacent Auslan sentences proved to be highly challenging for the professional Auslan/English interpreters involved in this study, as only three (Amber, Emily, and Molly; see Table 7.1) transferred the core meaning of all these sentences into spoken English. It stands to reason that Auslan sentence 107 (many many many have sign language interpreting services? not really) is the most challenging sentence of this summary section (or maybe the entire Auslan presentation), as only Amber, Emily, and Molly rendered the

Table 7.1  Interpreters’ accuracy rate of rendering seven adjacent Auslan sentences that contained end negation Pseudonym

105 106 107 108 109 110 (equal 111 Total (many (l-a-c-k (many (know (access … human rights (equal out … no) o-f) … not … no) oppressed) … no) … no) of 7 really)

Liz Amber Emily Wendy Linda Jane Molly Sophia Helen Claire Alex Annie Debbie Shannon Zoe Tiffany Monica Lauren Bernie Kay Accurate renditions Inaccurate renditions

1 1 1 0 1 1 1 1 1 0 0 0 0 1 0 1 1 1 1 1 14

0 1 1 0 1 1 1 1 1 0 0 0 0 1 1 0 0 1 0 1 11

0 1 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 3

0 1 1 1 1 1 1 1 1 0 0 1 0 0 0 1 1 0 1 1 13

0 1 1 1 0 1 1 1 1 1 0 0 1 1 1 1 0 0 1 0 13

0 1 1 1 0 0 1 1 1 1 1 1 1 1 1 0 1 0 0 0 13

0 1 1 1 0 0 1 1 1 0 0 0 0 1 1 0 0 0 1 0 8

6

9

17

7

7

7

12

1 7 7 4 3 4 7 6 6 2 1 2 2 5 4 3 3 2 4 3

Note: In the top row, numbers in front of the brackets represented the Auslan sentences, and Auslan signs in the brackets were the concepts and end negation that appeared in the respective sentences. An accurate English interpretation of an Auslan sentence was marked as 1 and an inaccurate one as 0.

194 Cognitive Overload and Coping Strategies sentence accurately into English. Given that the sentence-final headshake negated the sentence-initial sign, the only chance for interpreters to understand and render Auslan sentence 107 correctly was to wait till the end of the sentence before starting to interpret. In other words, interpreters’ onset processing time for Auslan sentence 107 needed to be a minimum of the length of the entire sentence (5.4 seconds). External rater Jamie commented that interpreters need to use long processing time and watch the Deaf presenter’s face apart from his hands in order to interpret the subtle end negation in Auslan sentence 107 correctly: If interpreters’ time lag [processing time] is too short, they would have said it [in affirmative] before they realise that a negation was happening.You can see that there is a question already on his face. He has already got doubt on this face as he was signing it. So, that’s just a little flag that says, ‘Don’t affirm this yet, keep it ambiguous.’ So, it is a subtle thing, but it’s a very important thing for interpreters to watch. It’s a really good case study why end negations trip interpreters up all the time. … You may say, ‘Although many countries have sign language interpreting services, they weren’t that solid.’

7.2 Coping Strategies for End Negation Local analysis of interpreters’ accurate English renditions of the adjacent Auslan sentences featuring end negation revealed that interpreters employed the following four strategies to cope with this particular syntactical difference between the source language and the target language: 1 The interpreter waited for nearly the entire Auslan sentence (i.e., onset processing time being typically more than 4 seconds) and started interpreting only when or after seeing the end negation in order to produce an accurate and idiomatic English rendition. 2 The interpreter used moderate or long onset processing time, started to interpret the initial meaning unit of the Auslan sentence before seeing the end negation, absorbed further information (e.g., the end negation) while rendering the initial meaning unit into English, or paused to receive the further information, then interpreted the end negation appropriately in order to produce a complete target language sentence (e.g., putting the English equivalent of the end negation near or at the end of the target language sentence). 3 The interpreter adhered to the pseudo-cleft structure (rhetorical question) of the Auslan sentence, started to pose an English question before seeing the end negation, then provided a negative answer to the question, thus making the English interpretation quite literal. 4 The interpreter started to interpret the Auslan sentence before seeing the end negation, drew on his/her linguistic knowledge (e.g., the Deaf presenter’s squinting of his eyes before producing the end negation), contextual

Cognitive Overload and Coping Strategies 195 knowledge, and/or sufficient pre-task preparation, and predicted the end negation successfully, thus rendering the end negation correctly before actually seeing it. I conducted local analysis of Amber and Emily’s English interpretations of the seven Auslan sentences (from 105 to 111) because they employed similar strategies to deal with some sentences but different strategies to cope with the others. The primary aim of the detailed analysis was to investigate the underlying interpreting process and coping strategies that resulted in their effective English renditions. Analysis of Emily’s English renditions will be presented before that of Amber’s. In addition, Kay’s English rendition of Auslan sentence 105 and Zoe’s English interpretations of Auslan sentences 109 to 111 will also be analysed because Kay achieved higher accuracy than both Emily and Amber on Auslan sentence 105, and Zoe predicted the end negation in Auslan sentences 109 to 111 successfully. All these correct English interpretations demonstrate the professional signed language interpreters’ creativity, diversity, and accuracy when working from a signed language into a spoken language at a formal setting. 7.2.1 Long Onset Processing Time and Free Interpretation The Deaf presenter arrived at the conclusion in Auslan sentence 105 by drawing on three of the aforementioned findings of the international survey study: 88 out of 93 countries reported that they have recognised Deaf children’s right to access education (Auslan sentence 77); 82 countries responded that their Deaf people could access government services (Auslan sentence 96); and 77 from the 93 respondents stated that their Deaf people were recognised as equal to hearing people (Auslan sentence 62). This section presents local analysis of Emily’s English renditions of Auslan sentences 105 to 112 (Example 7.1). With onset processing time of 8.7 seconds, Emily began interpreting Auslan sentence 105 after understanding the Deaf presenter’s essential meaning of the sentence segment s-o altogether research talk like many many block (prevent) deaf people no few, which contained the negation (many many … no) and its synonym few country. She rendered the initial part s-o altogether research talk like as ‘So, in the research, overall, we found that’ to introduce the topic, namely a summary of the key findings of the research. After waiting for 2 seconds to receive more information in Auslan sentence 105 (for deaf people access), Emily resumed her English interpretation and reframed the Deaf presenter’s end negation (many many block … no) and synonymous phrase (few country those put-down) as a positive and equivalent statement ‘most countries that responded, [0.9-second pause] um there were, there was access to …’). This syntactic restructuring in the target language cost Emily a substantial amount of processing capacity, as reflected by her withinsentential pause and minor self-correction regarding grammar. Interestingly, with regard to access t-o education government, she not only changed the order of education and government (‘there was access to government, there

196 Cognitive Overload and Coping Strategies

Example 7.1  Emily’s long onset processing time and free interpretation for coping with end negation

Cognitive Overload and Coping Strategies 197 was access to education’) but also repeated ‘there was access to’ probably to convey the Deaf presenter’s emphasis through the aforementioned use of negation and synonym in this Auslan sentence. Emily’s restructuring of sentence components led to her onset processing time for education being 5.2 seconds. After pausing for 1 second and seeing that the Deaf presenter moved onto the subsequent Auslan sentence (signing b-u-t in Auslan sentence 106), Emily resumed interpreting by summarising the overall meaning of Auslan sentence 105 (‘there was a level of accessibility’). A closer look reveals that this English rendition deviates from the meaning of the Deaf presenter’s final message (perception equal same same other people). Emily’s summarisation and (unjustifiable but minor) omission can be ascribed to her momentary attention deficit; that is, she was focusing most of her attention on target language production (‘there was access to government, there was access to education’) and did not pay sufficient attention to source language comprehension. An alternative explanation is that Emily may have realised that she had missed the Deaf presenter’s meaning unit at the end of Auslan sentence 105 but the Deaf presenter moved onto a new Auslan sentence; therefore, she decided to recapture the essential meaning of Auslan sentence 105 and focus her attention on understanding the new Auslan information. It is also possible that the interpreter’s aforementioned self-correction to achieve grammaticality, reordering of messages, changing the negation to the positive statement, and long processing time regarding education may have increased the cognitive pressure and collectively culminated in her momentary attentional deficit. Fatigue may be another reason for her minor omission, as by this juncture she had done simultaneous interpreting for 16 minutes. Despite the omission, Emily was well-paced in finishing off her English rendition of Auslan sentence 105, thus demonstrating good skills of smoothing over her English interpretation. Given her exceptionally long onset processing time, it is reasonable to say that she adopted the aforementioned first strategy to cope with this Auslan sentence containing end negation. With onset processing time of 5.4 seconds, Emily began interpreting Auslan sentence 106 after understanding the signs b-u-t but have what l-a-c-k o-f recognise and rendered them as ‘But there was [0.4-second pause] a lack [0.7-second pause] of recognition.’While saying this, she saw the remaining part of the complete message (approve sign language); hence, she then completed her English interpretation by saying ‘for [0.8-second pause] sign language,’ strategically omitting the redundant information approve, a synonym for recognise. She then paused for 1.5 seconds and rendered the second meaning unit also l-a-c-k o-f bilingual education as ‘a lack of bilingual education.’ The four noticeable micropauses within Emily’s English rendition of Auslan sentence 106 indicate that she was not only consciously waiting for more source language information in order to complete her ideas but also carefully planning her target language sentence to ensure both accuracy and grammaticality. In regard to Auslan sentence 107, Emily did not start interpreting until she saw and understood the Deaf presenter’s headshake at the end of the sentence.

198 Cognitive Overload and Coping Strategies As a result, her long onset processing time – 5.4 seconds – enabled her to understand the meaning of this Auslan sentence correctly. She conveyed the meaning faithfully and concisely by saying ‘a lack of [0.5-second pause] sufficient sign language interpreting services.’ Her continuous use of the phrase ‘a lack of ’ enhances coherence within her English renditions; and her strategic addition ‘sufficient’ reflects the Deaf presenter’s emphasis through his use of the rhetorical question. As expected, Emily’s long onset processing time regarding Auslan sentence 107 contributed to the long onset processing time (5.8 seconds) of the following Auslan sentence. Emily started interpreting Auslan sentence 108 only after seeing the negating sign at the end of the sentence. The long onset processing time enabled her to understand the full meaning of this Auslan sentence and produce an accurate and concise English rendition (‘and awareness about the Deaf community [0.6-second pause] in those countries’), which makes sense only if the target audience associates it with the phrase ‘a lack of ’ in the preceding English rendition. The analysis of both the interpreting process and the interpreting output thus far shows that Emily was following the Deaf presenter’s train of thought and was aware of conveying the logical relationship among his messages. Emily’s use of succinct English sentences shorted her processing time to some extent so that she could still follow the Deaf presenter. With onset processing time of 3.8 seconds, Emily rendered the topic of Auslan sentence 109 (means many those deaf people their lives can) into an English question (‘So, what does that mean for Deaf people living in those countries?’). While uttering these words, she saw the complete meaning through the signs (access t-o services really oppressed) and expressed the meaning as a clear answer (‘That potentially, there is some oppression there and some inequality’). Her precise lexical choice ‘oppression’ for oppressed, strategic addition ‘and some inequality,’ and intonation conveyed the Deaf presenter’s meaning, emphasis, and communicative effect successfully. The Deaf presenter’s pause (holding the sentence-final sign oppressed for 1.5 seconds) assisted with the simultaneous interpreting process to some degree, again reflecting that he was experienced with working effectively with signed language interpreters. Emily paused after completing her English interpretation of Auslan sentence 109. With onset processing time of 6.4 seconds, Emily started interpreting Auslan sentence 110 after deciphering the subject and verb (so those deaf people true enjoy enjoy enjoy equal human). So, she transferred the topic as ‘And it certainly tells us that Deaf people,’ using ‘And it certainly tells us that’ as a filler or linguistic padding to employ the stalling strategy in order to wait for more source language information. While she was saying those words, she saw the ending of Auslan sentence 110 (rights? no); so, she incorporated the meaning of the end negation by saying ‘don’t enjoy equal.’ She then paused for 1.6 seconds, possibly to plan her target language sentence or to recall the object of ‘enjoy,’ and finally completed her English sentence with the critical information (‘human rights [0.4-second pause] as their peers’). It is worth noting that Emily strategically added the referent (‘as their peers’) to explicitly convey the Deaf

Cognitive Overload and Coping Strategies 199 presenter’s meaning of comparing Deaf people with hearing people in terms of human rights. What is extraordinary about Emily’s English interpretation of this Auslan segment is her processing of Auslan sentence 111, which she seemed to have omitted but actually integrated after rendering the subsequent Auslan sentence into English.With considerably long onset processing time of 8.8 seconds, she started interpreting Auslan sentence 111 after seeing its sentence-final negating sign no and introduced its topic (so summarise that say) by saying ‘So, the report does indicate that.’ Note that her emphasis through the use of ‘does’ exemplifies dynamic equivalence of conveying the Deaf presenter’s signs so summarise to highlight that this Auslan sentence summarises the final conclusion of the whole research/presentation.While she was saying that, the Deaf presenter was producing the initial three signs of Auslan sentence 112 still more work; so, she rendered those signs as ‘there is still more work to be done.’ While saying that, she saw the signs t-o lobby lobby lobby, which she interpreted as ‘more lobbying that’s necessary.’ After pausing for 1.4 seconds, Emily did not continue to interpret Auslan sentence 113 but inserted the core meaning of Auslan sentence 111 (deaf people equal like other round world have have have? no) by reframing it as the purpose of doing more work and lobbying – ‘to enable Deaf people to achieve equal human rights,’ thus striking a positive note. A literal interpretation of the Deaf presenter’s Auslan sentence 111 may be ‘In summary, the research reveals that Deaf people are not equal to others around the world.’ Emily’s rendition of the need to do more work and lobbying in order to ensure Deaf people have equal human rights emphasises the goal state but at the same time indicates that the goal needs to be achieved through concrete actions because the current situation is not optimal (i.e., Deaf people have not yet attained equal human rights compared with hearing people). Thus, her free interpretation can be deemed as achieving dynamic equivalence. Example 7.1 also illustrates the importance for researchers and assessors to consider the context when evaluating the accuracy of interpretation. If I solely looked at ‘So, the report does indicate that’ to assess the accuracy of interpretation regarding Auslan sentence 111, I would have considered it as an unjustifiable omission of the essential meaning of this Auslan sentence; however, looking beyond this main clause, I found that Emily actually combined the core meaning of Auslan sentences 111 and 112 into one long English sentence and achieved a good communicative effect. In summary, given that Emily began interpreting Auslan sentences 105, 107, 108, and 111 only after seeing the end negation of each sentence, it is reasonable to think that she employed the first strategy (see the list of strategies at the beginning of Section 7.2) to deal with these pseudo-cleft sentences that ended with negation. However, she started interpreting Auslan sentence 110 with long onset processing time (6.4 seconds) before seeing the end negation, and thus this can be considered as using the second strategy to cope with the end negation in this sentence. Example 7.1 therefore illustrates that the interpreter

200 Cognitive Overload and Coping Strategies switched from one strategy to another to cope with adjacent Auslan rhetorical questions that contained end negation. 7.2.2 Switch from Free Interpretation to Literal Interpretation Amber also used a variety of strategies to cope with this sequence of Auslan rhetorical questions that ended with negation (Example 7.2). With quite long onset processing time of 9.7 seconds, Amber only embarked on interpreting Auslan sentence 105 after seeing both the end negation (many many … no) and the synonymous phrase (few country), and reflected the Deaf presenter’s emphasis by both using the phrase ‘very few countries’ and stressing the verb ‘deny.’ She produced a free interpretation of block (prevent) deaf people as ‘deny Deaf people their human rights [0.6-second pause] solely because of their Deafness [0.5-second pause],’ which re-captured the central theme of the Auslan presentation. Note that she made a strategic addition ‘solely because of their Deafness’ to highlight the issue and clarify the meaning, but the strategic addition increased her processing time and cognitive load.The 0.6-second pause before the strategic addition indicates that Amber was deliberating whether to add the phrase to make the meaning clear, or that she was planning this phrase in her head. The 0.5-second pause after the strategic addition may indicate that she was trying to link what she had just said with what she had just seen (access t-o education government) to maintain intra-sentential coherence, or that she was trying to recall the visual message that she had just seen, or that she was planning her target language expressions. She rendered that Auslan meaning unit accurately as ‘in terms of access to education, [0.4-second pause] government.’ Regarding Amber’s English interpretation of Auslan sentence 105, her 2.6-second pause after saying ‘government’ may indicate that she was experiencing cognitive overload, which may be due to her long processing time, coping with both the syntactical differences and dense information, as well as her multitasking. While she was saying ‘in terms of … government’ and pausing, the Deaf presenter was signing his final meaning unit of Auslan sentence 105 perception equal same same other people. Given the concurrent processing, it is highly likely that she allocated the majority of her attention to target language production of the previous meaning units and did not have sufficient attention for comprehending this final Auslan meaning unit, which caused the unjustifiable minor omission of this signed message. Being aware of the omission, Amber decided to generalise the meaning by strategically adding ‘and services,’ which explicated the implicit meaning of government and echoed the overall meaning of Auslan sentence 105. Due to the omission, her processing time for Auslan sentence 106 was medium, at 3.8 seconds. Similar to Emily, Amber started interpreting only after seeing the negation at the end of Auslan sentences 107 and 108 and rephrased the negation as ‘a lack of,’ which echoed her interpretation of Auslan sentence 106 and contributed to inter-sentential coherence. Note that Amber conveyed the meaning of sign

Cognitive Overload and Coping Strategies 201

Example 7.2  Amber switched from free interpretation to literal interpretation to cope with adjacent Auslan sentences with end negation

language interpreting services concisely as ‘interpreting services,’ which was slightly vague but still comprehensible and acceptable in the context. Given Amber’s long onset processing time regarding Auslan sentences 105, 107, and 108 (each of these sentences featured both the rhetorical question structure and end negation), it is reasonable to say that she adopted the first strategy to cope with this type of syntactical difference between Auslan and English.

202 Cognitive Overload and Coping Strategies As to Auslan sentence 109, Amber understood the initial meaning unit means many those deaf people their and introduced the subject as ‘So, many Deaf people,’ then paused for 2 seconds to absorb more Auslan information, during which she saw the critical message lives can access t-o services really oppressed. After seeing the key sign oppressed and understanding this meaning unit, she resumed interpreting and rendered this message as ‘ah do not have access’ to reflect that the signs can access were negated by the sentence-final sign oppressed. Note that she made an unjustifiable minor omission of t-o services. She then paused for 1.4 seconds and strategically added ‘or equal standing to others in their community’ to expand on the meaning of the sign oppressed and to reflect the Deaf presenter’s emphasis. Interestingly, Amber then switched to a different strategy – maintaining the question-and-answer structure – in order to cope with Auslan sentence 110 that also featured the pseudo-cleft structure and end negation.With onset processing time of 5 seconds, she started interpreting this Auslan sentence after understanding the subject and verb phrase so those deaf people true enjoy enjoy enjoy and conveyed the meaning as ‘So, can we say that Deaf people enjoy.’ While she was uttering those English words, the Deaf presenter was signing the remaining of the question equal human rights?. Amber then added the object ‘the equivalent human rights to their hearing counterparts?’. It is interesting to note that she strategically added the implied referent of equal – ‘to their hearing counterparts’ – in order to convey the complete idea and produce an immediately comprehensible English sentence. After seeing the negating sign no, she rendered it as the answer to the aforementioned question ‘No, I don’t think so.’This literal interpretation approach enabled Amber to follow the Deaf presenter closely. With onset processing time of 2.8 seconds, Amber rendered the understood initial small meaning units of Auslan sentence 111 so summarise that say as ‘So, in conclusion, we are saying.’ She paused for 1.8 seconds to plan the target language expression for the signs that she had just seen deaf people equal like other, then introduced the subject ‘Deaf people around the world.’ At this point, she was still waiting for more information about the sign equal. As soon as she saw the sign no, she expressed the awaited meaning by saying ‘do not have equal access to human rights [1-second pause] as others.’ The 1-second pause before the strategic addition ‘as others’ indicates that at the particular juncture Amber might be monitoring her English sentence for grammaticality and coherence and wondering how to achieve these goals, or she might be devoting the majority of her attention to making sense of the critical sign lobby in the subsequent Auslan sentence (112). Given that Amber used medium onset processing time and started interpreting Auslan sentence 111 before seeing its end negation, it is reasonable to state that she used the second strategy to cope with this Auslan sentence that also had the rhetorical question structure. Despite their similar level of familiarity with the Deaf presenter, an obvious difference between Emily and Amber lies in their cognitive processing of Auslan

Cognitive Overload and Coping Strategies 203 sentence 111 in that Emily incorporated its meaning quite late but Amber relayed its meaning early on, mainly because Emily used onset processing time of 8.8 seconds and Amber used onset processing time of 2.8 seconds to process the same Auslan sentence. Emily’s exceptionally long onset processing time required her to memorise a considerable amount of Auslan information and increased the possibility of experiencing cognitive overload and omitting important information; however, she retained the understood meaning, processed the subsequent Auslan message, then inserted that meaning to her English interpretation. Probably Emily’s good working memory capacity and effective English interpretation can be ascribed to her meaning-based information processing and an excellent grasp of the Deaf presenter’s key messages of the whole signed presentation. Examples 7.1 and 7.2 illustrate that an individual interpreter used a range of strategies to cope with the same type of syntactical structure (e.g., rhetorical question that contained end negation) in a continuous segment. These examples also demonstrate noticeable differences between the two versions of accurate interpretation of the same source language segment, thus highlighting the fact that interpreters employed diverse strategies to achieve high accuracy and effective communication. Taken together, the two examples show both creativity and diversity within a piece of interpretation and between different versions of interpretation.There are multiple approaches for achieving accuracy of interpretation; and simultaneous interpreting involves a creative and complex cognitive process. 7.2.3 An Exception: Use of Short Processing Time and Pause to Cope with End Negation Given that both Emily and Amber omitted the final message of Auslan sentence 105, I conducted local analysis of the most accurate English rendition of this difficult sentence – Kay’s interpretation (Example 7.3) – in order to investigate how she achieved exceptionally high accuracy. Kay was finishing off her English interpretation of the previous Auslan sentence while the Deaf presenter was producing the first two signs of Auslan sentence 105 (s-o altogether). Unlike both Emily and Amber, whose onset processing time for this Auslan sentence was approximately 9 seconds, Kay used onset processing time of 2.5 seconds that enabled her to produce a literal interpretation (‘So, altogether’) of s-o altogether. Then, making use of the Deaf presenter’s 1.5-second pause between talk like and many, Kay conveyed the meaning of research talk by saying ‘what the research was telling us was.’ After that, she paused for 3.8 seconds to comprehend incoming meaning units that contained the end negation (many block [prevent] deaf people no) and its synonym (few countries), which she correctly interpreted as ‘there were few countries who, uh, very few who disallowed Deaf people.’ It is obvious that her 3.8-second pause enabled her to concentrate on understanding the negation many block (prevent) deaf people no and its synonym few countries and planning the target language

204 Cognitive Overload and Coping Strategies

Example 7.3  Kay’s use of short onset processing time and pause to achieve high accuracy of interpreting Auslan sentence 105

rendition. Her self-correction ‘few countries who, uh, very few who’ suggests that she was making efforts to convey the Deaf presenter’s emphasis through the aforementioned end negation (many … no) and synonym (few countries). Kay’s repetition of the word ‘disallowed’ in ‘who disallowed Deaf people access to government, disallowed education’ also conveyed the Deaf presenter’s emphasis through his use of the synonymous signs block (prevent) and putdown. Like Emily, Kay interpreted government before education, perhaps to reduce the pressure on her working memory. Unlike Emily and Amber, Kay rendered the final Auslan phrase perception equal same same other people accurately as ‘who did not see Deaf people as equal to those who are not deaf,’ with ‘who did not see’ logically linked to the previous concept ‘few countries who, uh, very few who.’ The Deaf presenter’s use of constructed action (roleshift) in this Auslan phrase required interpreters to add negation (‘not perceiving Deaf people as equal to hearing people’) in order to make this message meaningful in the context; hence, this Auslan meaning unit required that interpreters should allocate much processing capacity to target language production. It is likely that Kay’s short onset processing time (2.5 seconds) for Auslan sentence 105, short processing time regarding perception (i.e., the time interval between the beginning of perception and the beginning of ‘see,’ 3.3 seconds), and fast processing contributed to her accurate English rendition of the final Auslan phrase perception equal same same other people. This example

Cognitive Overload and Coping Strategies 205 indicates that interpreters’ short onset processing time, fast processing, pause, and making use of the speaker/signer’s pause may enable them to produce highly accurate interpretations. 7.2.4 Prediction While the aforementioned examples have illustrated professional interpreters’ use of the first three strategies to cope with Auslan rhetorical questions containing end negation, Example 7.4 illustrates how Zoe used anticipation/prediction (the fourth strategy) to deal with two adjacent Auslan sentences (110 and 111) that ended with negation. I chose Zoe for the local analysis because she was the only interpreter in this study who successfully predicted the end negation of these two challenging Auslan sentences. After understanding the logical link (so) and the subject of Auslan sentence 110 (so those deaf people), Zoe rendered these messages into English as ‘So, these Deaf people.’ She then waited for 1 second to understand more Auslan information (enjoy equal) and plan her target language production. When she resumed interpreting, she said ‘were not able to enjoy,’ indicating that she predicted the end negation no successfully (i.e., using the fourth strategy). It is highly probable that Zoe anticipated the end negation because she noticed the Deaf presenter’s squinting when he was signing true enjoy enjoy enjoy, as well as because she made use of her advance preparation and contextual knowledge (Auslan sentences 105 to 109) to comprehend Auslan sentence 110. When the Deaf presenter was producing the end negator no of Auslan sentence 110 and the first two signs (so summarise) of Auslan sentence 111, Zoe was articulating the object of the aforementioned subject and verb ‘these Deaf people were not able to enjoy’ – ‘the same level of human rights’ – and strategically adding the implicit referent of the comparison – ‘as anybody else within their community’ – in order to complete her English sentence. It is worth noting that the referent ‘as anybody else within their community’ was a strategic addition that completed and clarified the Deaf presenter’s Auslan meaning unit regarding the comparison. Due to Zoe’s correct prediction, her tail-to-tail span in regard to Auslan sentence 110 was short, at 2.1 seconds. With onset processing time of 2.1 seconds, Zoe started interpreting Auslan sentence 111 when the Deaf presenter was producing the third sign (say) of this sentence. She rendered the logical link and subject (so summarise say deaf people) in a literal manner as ‘So, in summary, what it states is that, Deaf people,’ then paused for 0.9 second. While she was saying those words and pausing, the Deaf presenter was signing say deaf people equal like other. Interestingly, Zoe then resumed interpreting by uttering ‘have not’ while the Deaf presenter was signing round world, indicating that she again predicted the end negation successfully. She then paused for 0.9 second, during which the Deaf presenter was signing have have have. Zoe was saying ‘become’ while the Deaf presenter was signing the negator no at the end of Auslan sentence 111.

206 Cognitive Overload and Coping Strategies

Example 7.4  Zoe’s successful prediction of end negation in two adjacent Auslan sentences

After a 0.9-second pause, Zoe completed her English rendition of Auslan sentence 111 by conveying the meaning of the aforementioned signs equal like other round world as ‘as equal as others in the wider community.’ Probably Zoe’s successful prediction of the end negation of Auslan sentence 111 was attributable to her paying much attention to the Deaf presenter’s non-manual features (e.g., facial expression, squinting) when he was signing this Auslan sentence, her sufficient pre-task preparation (high familiarity with the conclusion of the whole Auslan presentation), and her use of the contextual knowledge (i.e., there were four other Auslan sentences with end negation before Auslan sentence 111).

7.3 Discussion of Results Chapter 6 and this chapter have analysed the 20 professional Auslan/English interpreters’ accurate English renditions of adjacent Auslan sentences containing numbers and end negation, respectively, with a primary focus on local analysis of ten representative examples from the experimental corpus of Auslan-to-English simultaneous interpretation data. Results from quantitative analysis revealed that the interpreters’ average accuracy rate of interpreting individual Auslan numbers (89%) was considerably higher than their average accuracy rate of interpreting Auslan sentences containing the numbers (60%). Results also showed that many interpreters rendered the Auslan numbers accurately into spoken English, but omitted or distorted their associated referents and/or ideas, thus interpreting the respective Auslan sentences inaccurately. These findings corroborate previous

Cognitive Overload and Coping Strategies 207 findings (Mackintosh, 1983; Mazza, 2001) that numbers not only pose interpreting difficulties themselves but also affect interpreters’ processing of the surrounding information. These findings indicate a need to train both interpreting students and professional interpreters on how to deal with numbers in simultaneous interpreting. In addition, results revealed that the professional interpreters’ accuracy rate of interpreting a single Auslan sentence containing a number was substantially higher than their accuracy rate of rendering some consecutive Auslan sentences containing numbers into spoken English. Furthermore, the interpreters’ accuracy rate of interpreting a single Auslan sentence featuring end negation was markedly higher than their accuracy rate of interpreting a sequence of Auslan sentences featuring end negation into spoken English.These new findings bring an added perspective to the existing literature on problem triggers in spoken and signed language simultaneous interpreting. These findings reaffirm many interpreters’ retrospective comments on the various challenges in the Auslan-toEnglish simultaneous interpreting task (see Chapter 4), including interpreting numbers, syntactical differences between Auslan and English, dense information in the source text, and maintaining coherence in the target language speech. These findings substantiate Gile’s (2009) as well as Seeber and Kerzel’s (2012) arguments that syntactical differences between the source language and the target language (i) require interpreters to re-order sentence components in order to produce grammatical and idiomatic renditions, (ii) increase cognitive load, and (iii) pose challenges for interpreters during simultaneous interpreting. Plausible explanations for these findings include: (i) problem triggers appearing in a row increased both information density and processing capacity requirements, and (ii) the artificial testing conditions deprived the interpreters of useful resources (e.g., team interpreter, visual aids, cooperation with the Deaf presenter) to cope with those challenges. Aside from the results of the quantitative analysis, the local (qualitative) analysis of selected interpreters’ accurate English renditions of adjacent Auslan sentences containing numbers or end negation has revealed numerous interesting findings. An intriguing finding is that an interpreter’s (Jane in Example 6.1) perfection of a previous English rendition increased her tail-to-tail span regarding the previous Auslan sentence, delayed her English interpretation of the present Auslan sentence, and partly led to an incorrect English rendition of the present Auslan sentence, which she later rectified thanks to the Deaf presenter’s repetition of the same concept and her own self-monitoring. This finding lends support to Gile’s (2009) Effort Model of simultaneous interpreting which explains that if an interpreter allocates excessive attention to producing a perfect target language sentence, then there may be insufficient attention available for source language comprehension of the incoming information, and interpreting performance may deteriorate. This finding also provides empirical support to Gile’s (2008) claim about imported load, current load, and exported load. In addition, this finding corroborates previous findings (Lamberger-Felber, 2001;

208 Cognitive Overload and Coping Strategies Lee, 2002, 2011) that long onset processing time often results in major omissions and other interpretation errors. Jane’s lack of suppression of her biases and erroneous inference drawn from the context also contributed to the occurrence of the aforementioned interpretation error. This finding substantiates Gernsbacher and Shlesinger’s (1997) claim that interpreters need to suppress their own prejudices and inaccurately drawn inferences from the context during interpreting. The fact that interpreters can immediately hear their own interpretations during signed-tospoken language simultaneous interpreting facilitates their error detection and self-monitoring in this language direction (Napier, Rohan, & Slatyer, 2005; Nicodemus & Emmorey, 2013; Wang & Napier, 2015). I also found that some interpreters who used sufficiently long onset processing time (e.g., Jane in Example 6.1, Emily in Example 7.1) adopted a predominantly free interpretation approach, conveyed the core messages accurately, maintained intra-textual coherence in their target language speeches, and used a formal register, whereas other interpreters who used relatively short onset processing time (e.g., Lauren in Example 6.2) conveyed key information and details accurately but applied a predominantly literal interpretation approach, had source language interference, made more grammatical errors, and used an informal register, which was inappropriate for the formal conference setting. These findings echo previous results (Cokely, 1986; McKee & Napier, 2002; Seeber & Kerzel, 2012; Wang, 2020) that interpreters need to use sufficiently long processing time during simultaneous interpreting in order to ensure thorough understanding of source language messages, cope with syntactical differences between the source language and the target language, search for semantic equivalents in the target language, and produce idiomatic renditions.These findings also reaffirm previous findings (Cokely, 1986; Defrancq, 2015; see also Chapter 5, this volume) that short or very short processing time leads to source language interference. Another related finding is that interpreters who started with relatively short onset processing time then used stalling or pausing where necessary in order to receive more incoming information (e.g., Jane in the first half of Example 6.5, Kay in Example 7.3) also managed the simultaneous interpreting process successfully and provided accurate and effective signed-to-spoken language interpretations. This new finding adds to the existential literature about short processing time, fast processing, chunking, and pause in simultaneous interpreting.This finding suggests that interpreters can use a combination of long and short onset processing time during simultaneous interpreting and adjust their processing time by employing appropriate strategies when necessary (e.g., when dealing with numbers, syntactical differences, unfamiliar signs, dense information, self-­correction) in order to strike a cognitive balance between concurrent sub-processes. When coping with Auslan sentences that contained numbers or negation at the end, some interpreters used sufficiently long onset processing time (e.g., waiting till the end of an Auslan sentence; see Liz in Example 6.3 and Emily in Example 7.1) in order to understand the meaning of the whole Auslan sentence accurately, whereas other interpreters used medium onset processing time together with particular strategies such as making strategic additions (see Amber

Cognitive Overload and Coping Strategies 209 in Example 6.4) and making predictions (see Zoe in Example 7.4). These findings indicate professional interpreters’ creativity and strengths in diversity in simultaneous interpreting. In other words, depending on the nature of source language material, interpreters’ working memory capacity, their interpreting skills, their strengths, and their weaknesses, they can choose different lengths of onset processing time coupled with appropriate strategies to produce accurate and effective simultaneous interpretations. Another crucial finding is that the interpreters primarily employed the ­following four strategies to cope with adjacent Auslan sentences (e.g., ­topic-comment structures and pseudo-cleft sentences) each containing a number near or at the end (i.e., a syntactical difference between Auslan and English): (i) using long onset processing time and producing idiomatic and succinct renditions, (ii) using medium onset processing time in combination with suitable strategies such as ­stalling and waiting, (iii) using short onset processing time together with other strategies such as stalling and pausing, and (iv) retaining the question-and-answer structure of the Auslan pseudo-cleft sentences. As expected, interpreters’ strategies differed from person to person; and the same interpreter’s strategies varied from one Auslan sentence to another even though the Auslan sentences had similar syntactic features. I also found that interpreters’ processing time interacted with their strategies, use of the printed PowerPoint slides, cognitive load, and the nature of source language materials in complex ways (see Examples 6.3, 6.4, 6.5, 6.6, 7.1, and 7.2). Despite excessively long onset processing time at a particular juncture (9.1 seconds), an interpreter’s (see Liz in Example 6.3) frequent reference to the printed PowerPoint slides and production of concise English renditions enabled her to catch up with the Deaf presenter, maintain cognitive balance, and produce accurate English interpretations of a sequence of Auslan sentences containing numbers near or at the end. Another interpreter’s (see Amber in Example 6.4) use of strategic additions to deal with a series of Auslan sentences containing numbers near or at the end enabled her to use medium onset processing time, follow the Deaf presenter’s logic thread, and produce effective English interpretations. In addition, as illustrated by Example 6.5, an interpreter’s (Jane) long pause (3.2 seconds) in the middle of a target language sentence and miscomprehension of a source language message resulted in an incorrect interpretation; and her self-correction of the erroneous interpretation increased her onset processing time for the subsequent Auslan sentence, which triggered her to switch from a free interpretation approach to a literal interpretation approach in order to save cognitive processing capacity, keep up with the Deaf presenter, and maintain an acceptable level of accuracy of interpretation. Taken together, these findings indicate that examining processing time, cognitive load, use of strategies, and interpreting quality together rather than in isolation can provide insights into cognitive processes, reasons for accurate and inaccurate renditions, and the link between interpreting process and interpreting quality. Like processing time, interpreters’ strategies varied from person to person, from sentence to sentence, and from juncture to juncture, depending on a wide range of factors. As illustrated by Example 6.5, when a source language

210 Cognitive Overload and Coping Strategies comprehension problem led to an inaccurate interpretation, the interpreter (Jane) corrected the mistake and lagged behind the Deaf presenter, then switched from a free interpretation approach to a literal interpretation approach in order to strike a good cognitive balance, provide accurate renditions, and keep up with the Deaf presenter.When interpreting a long source language segment featuring numbers, dense information, and syntactical differences between the source language and the target language, professional interpreters did experience cognitive saturation and produced interpretation errors such as ungrammaticality (see Example 6.5) and unjustifiable omissions (see Example 6.6). When the professional interpreters recovered from the cognitive overload or interpretation breakdown, they often refocused their attention and continued to process the incoming source language information. Furthermore, the local analysis in this chapter revealed that the professional interpreters employed the following strategies in order to deal with Auslan sentences featuring end negation (i.e., a syntactical difference between Auslan and English): (i) waiting for the whole Auslan sentence to finish before starting to interpret; (ii) using moderate or long onset processing time, comprehending more incoming source messages while reformulating the understood sentence-initial information in the target language or pausing if necessary to understand more incoming information, then rendering the end negation in the target language; (iii) adhering to the question-and-answer structure of the Auslan sentence and producing a predominantly literal interpretation; and (iv) predicting the end negation successfully. My findings regarding the strategies employed by the professional interpreters to deal with syntactical differences between Auslan and English (e.g., Auslan sentences containing numbers near or at the end, Auslan sentences ending with negation) verify Seeber’s (2011) postulation that interpreters adopt waiting, stalling, chunking, and anticipation to cope with syntactical asymmetry between the source language and the target language during simultaneous interpreting. Interestingly, I found that the timing of concurrent processing of crucial information is somewhat random but vital for the success of simultaneous interpretation, as exemplified by the fact that Jane’s (Example 6.1) start of interpreting Auslan sentence 62 coincided with the Deaf presenter’s production of the key point in the sentence (deaf people), and this overlapping and less-thanoptimal timing gave rise to Jane’s attention allocation problems, an erroneous interpretation of the key point, and a subsequent self-correction.This new finding contributes to a better understanding of the link between concurrent processing and simultaneous interpreting performance. This finding corroborates previous results (Gile, 2009; Liu, Schallert, & Carroll, 2004) that interpreters’ attention allocation ability plays an important role in producing highly accurate simultaneous interpretations. Moreover, I found the following novel results regarding timing, strategies, and cognitive balance: 1 Interpreters sometimes made strategic additions (e.g., expansion of ideas, insertion of referents of numbers; see Examples 6.1, 6.6, and 7.2) and selfcorrections (see Examples 6.1 and 6.5) in order to ensure the clarity of

Cognitive Overload and Coping Strategies 211

2

3

4

5

messages, grammaticality of target language renditions, intra-textual coherence, achievement of communicative effect (e.g., emphasis), and accuracy of interpretation, thus lagging far behind the Deaf presenter.This finding is consistent with previous results that spoken and signed language interpreters make strategic additions (Bartłomiejczyk, 2006; Braun, 2017; Major & Napier, 2012; Wadensjö, 1998; Wang & Fang, 2019) in order to ensure the effectiveness of their interpretations and maximise audience design. Some interpreters made use of the Deaf presenter’s inter-sentential pauses (e.g., elongation/hold on the sentence-final sign such as a number or negation, pause between sentences, see Examples 6.4, 6.5, and 7.1) in order to finish off their English interpretations and catch up with the Deaf presenter. This finding is in line with prior results (Barik, 1973; Goldman-Eisler, 1972/2002) that interpreters take advantage of the speaker’s pauses to reduce the amount of time during which they must engage in simultaneous processing of information (multitasking). Interpreters often made strategic omissions of redundant or secondary information in the source text in order to save time and processing capacity, catch up with the Deaf presenter, and produce effective renditions (see Examples 6.1 and 6.2 regarding ‘dictionary,’ Examples 6.3 and 6.4 regarding total communication speak-sign altogether, Examples 7.1 and 7.2 regarding many many … no few country). These findings are consistent with previous results (Bartłomiejczyk, 2006; Gile, 2009; Major & Napier, 2012; Napier, 2004, 2015, 2016; Shlesinger, 2000; Wadensjö, 1998; Wang & Fang, 2019) that signed and spoken language interpreters make strategic omissions to reduce cognitive load and enhance the effectiveness of their interpretations. Interpreters sometimes summarised information (see Example 7.1), omitted secondary information or details (see Examples 7.1 and 7.2 regarding perception equal same same other people), used concise and creative target language expressions and syntactic structures (see Examples 7.1 and 7.2 regarding the interpreters’ repeated use of ‘a lack of ’), switched from a free interpretation approach to a literal interpretation approach (see Example 7.2), applied a predominantly literal interpretation approach (see Examples 6.2 and 6.6), and re-ordered source language information units (see Example 7.1 regarding access t-o education government in Auslan sentence 105 and Auslan sentence 111) in order to deal with dense information and other interpreting difficulties, save attentional resources, buy time, catch up with the Deaf presenter, and still produce comprehensible and acceptable (but maybe not perfect) renditions. These findings corroborate previous results (Bartłomiejczyk, 2006; Gile, 2009; Napier, 2016; Wang, 2016) that spoken and signed language interpreters use strategies to cope with challenges in simultaneous interpreting, deal with problem triggers, and achieve high quality of simultaneous interpreting performance. When processing Auslan sentences ending with negation (a syntactical difference between Auslan and English), a few interpreters (e.g., Zoe in

212 Cognitive Overload and Coping Strategies Example 7.4) predicted the end negation successfully, probably based on their understanding of the context, their linguistic and extralinguistic knowledge, interpreting experience, and good advance preparation, thus avoiding lagging excessively far behind the Deaf presenter and producing accurate simultaneous interpretations. This finding corroborates previous empirical results (Bartłomiejczyk, 2006; Gile, 2008, 2009; Jorg, 1997; Liontou, 2012) that interpreters make use of both linguistic cues (e.g., collocation in spoken languages, non-manual features in signed languages) and contextual cues (e.g., background information, topics, text structure, previous sentences) in the source text to make linguistic and extralinguistic prediction, and that interpreters use successful prediction as a strategy to overcome challenges in simultaneous interpreting posed by syntactical differences regarding particular language combinations. 6 As illustrated by Example 6.6, interpreters’ strategic additions and efforts to produce highly accurate and comprehensible renditions increased their processing time and caused them to make unjustifiable omissions of subsequent information. This finding corroborates previous results (De Meulder, Napier, & Stone, 2018; Napier, 2007; Napier, Carmichael, & Wiltshire, 2008) that signed language interpreters need to maintain eye contact and cooperate with the Deaf presenter in order to pace the signed presentation together, avoid cognitive overload, and ensure the accuracy of signed-tospoken language simultaneous interpretation. Taken together, results from the local analysis of the ten representative examples of accurate English interpretations of highly challenging Auslan segments indicate that each professional interpreter operated within his/her ‘safe’ processing time range and employed appropriate strategies in order to manage multitasking, cope with time pressure, strike a delicate cognitive balance, and ensure accuracy of signed-to-spoken language simultaneous interpretations. Results of the local analysis also indicate that each professional interpreter’s ‘safe’ processing time range appeared to depend on internal and external factors such as his/her working memory capacity, interpreting experience, familiarity with the presenter, familiarity with the source text content, his/her use of strategies, syntactical differences between the source language and the target language, and complexity of the source language materials. The interpreter’s onset processing time, processing time for particular items (such as numbers, names, a list of items, and technical terms), and tail-to-tail span need to fall within that ‘safe’ range in order to avoid working memory overload and cognitive overload. Here, it is important to acknowledge that the particular length of processing time (e.g., 5 seconds) may be manageable for one interpreter but not for another interpreter; and it may be manageable for the same interpreter at one juncture but problematic at a different juncture, depending on many factors such as the individual’s working memory capacity, timing of concurrent processing, and the amount of available processing capacity. When the interpreter exceeds his/her ‘safe’ processing time range, he/she may need to

Cognitive Overload and Coping Strategies 213 adopt appropriate strategies or rely on external help (e.g., a team interpreter’s prompt, the presenter’s repetition or pause) in order to deal with interpreting challenges, recover from interpretation breakdowns, catch up with the presenter, and maintain or regain a cognitive balance. Results also indicate that interpreters may lag far behind the speaker/signer/ presenter due to reasons such as waiting for sufficient incoming information to ensure good comprehension, making strategic additions, having used unconcise target language expressions or particular syntactic structures, experiencing difficulties with target language reformulation, and correcting an erroneous interpretation. When interpreters sense that they are lagging quite far behind and may experience cognitive saturation, they try to catch up and still produce acceptable renditions by employing a broad range of strategies (e.g., making use of the speaker/signer/­ presenter’s pauses, strategically omitting redundant and secondary information, generalising, summarising, speaking faster) in order to strike a cognitive balance. Apart from the aforementioned findings in relation to cognitive overload and coping strategies, an interesting phenomenon observed from the local analysis of the ten examples of accurate English interpretations of highly challenging Auslan segments is that many of those English renditions contain imperfections such as minor grammar mistakes (see Example 6.1), long pauses between English sentences, pauses within the English sentences, hesitations, rising intonation at the end of English sentences that were statements (see Example 6.2), and self-­corrections (see Example 6.5).Those features indicate that target speech production under the condition of simultaneous interpreting is effortful and non-automatic (Gile, 2009). Some of those characteristics (e.g., filled pauses in compound words) suggest that interpreters are under heavy cognitive load at particular junctures during simultaneous interpreting (Defrancq & Plevoets, 2018). Those characteristics of simultaneous interpretations may also be taken as features of interpretese (i.e., the revealing indicators that someone is listening to an interpretation rather than a natural speech) (Shlesinger, 2008). In terms of accuracy, although the Auslan-to-English simultaneous interpretations in the ten examples are overall accurate and acceptable, some contain unjustifiable but minor omissions (see Examples 7.1 and 7.2) or minor distortions (see Examples 6.5 and 6.6 regarding Auslan sentence 98). This finding not only supports Gile’s (2009, p. 157) observation that ‘errors and omissions are found … in the work of seasoned professional interpreters’ but also lends some support to Setton and Dawrant’s (2016, p. 305) claim that ‘All simultaneous interpreting is, inevitably and almost by definition, a form of approximation at the first pass.’ Taken together, these findings indicate that simultaneous interpreting is a highly challenging cognitive task involving multitasking, and that simultaneous interpreting performance is typically imperfect despite professional interpreters’ best efforts.

7.4 Comments on the Local Analysis Local analysis throughout this book refers to the researcher’s detailed analysis of selected interpreters’ effective renditions of adjacent source language sentences.

214 Cognitive Overload and Coping Strategies The aforementioned local analysis of representative examples from the Auslan-to-English simultaneous interpretation corpus has illustrated failure sequences (Gile, 2009, pp. 171–173) in signed-to-spoken language simultaneous interpreting.1 On the one hand, the examples regarding syntactical differences between Auslan and English – Auslan sentences featuring numbers or negation at the end that require interpreters to re-order sentence elements and thus may overload working memory – illustrate problem triggers that impose considerably high cognitive pressure on interpreters at particular junctures (local cognitive load; see Gile, 2008, p. 59). On the other hand, the examples regarding numbers in the middle of Auslan sentences illustrate problem triggers (e.g., two-digit numbers, short spoken words) that do not require much processing capacity but disappear from interpreters’ working memory or focus of attention very quickly. Therefore, interpreters need to allocate their attention appropriately to multitasking at any point during simultaneous interpreting. Taken together, in simultaneous interpreting, interpreters’ inadequate total available processing capacity, insufficient processing capacity for a particular Effort (source language comprehension, short-term memory, target language production, or coordination), and their mismanagement of sufficient total available processing capacity at any juncture may lead to problems in cognitive processing and deterioration of the quality of their interpreting performance (e.g., inaccurate semantic content, inferior fluency, unnatural intonation). An obvious limitation of the local analysis lies in its speculative explanation of what is happening in a specific case (as shown by the frequent hedges ‘perhaps,’ ‘maybe,’ etc.). To compensate for the limitation, researchers can ask interpreters to engage in a post-task retrospection of the simultaneous interpreting process and can incorporate the interpreters’ retrospective quotes to strengthen the researchers’ arguments based on the local analysis. In addition, given that the local analysis is markedly time-consuming, it is recommended that it be used for analysis of particularly challenging source language segments or for analysis of highly skilled interpreters’ simultaneous interpretations in order to shed interesting light on simultaneous interpreting process and product. Given that the local analysis in Chapter 6 and in this chapter has provided many insights into how different interpreters coped with multiple challenges in extremely difficult source language segments, I recommend that spoken language interpreters and signed language interpreters use local analysis in conjunction with other methods (e.g., retrospective think aloud protocol) to investigate the intricate relationship among cognitive processing, strategies, and simultaneous interpreting performance.

7.5 Summary This chapter has reported quantitative results of the 20 professional Auslan/ English interpreters’ accuracy rate of interpreting adjacent Auslan sentences containing end negation. Furthermore, the chapter has focused on reporting

Cognitive Overload and Coping Strategies 215 findings from local (qualitative) analysis of accurate English renditions of adjacent Auslan sentences containing end negation, which also featured dense information and syntactical differences between the source language and the target language. The local analysis has shed interesting light on the interpreters’ cognitive processing, cognitive overload, coping strategies for those challenges, and simultaneous interpreting performance. Moreover, this chapter has discussed the findings in Chapter 6 and at the beginning of this chapter in relation to the relevant literature. Finally, the chapter has discussed the strengths and limitations of the local analysis, pointing out that it is useful for analysing spoken and signed language simultaneous interpretation data.

Note 1 I would like to thank Professor Daniel Gile for generously sharing his insights on his Effort Model of simultaneous interpreting and on my local analysis.

References Barik, H. C. (1973). Simultaneous interpretation:Temporal and quantitative data. Language and Speech, 16(3), 237–270. doi: 10.1177/002383097301600307. Bartłomiejczyk, M. (2006). Strategies of simultaneous interpreting and directionality. Interpreting, 8(2), 149–174. doi: 10.1075/intp.8.2.03bar. Braun, S. (2017). What a micro-analytical investigation of additions and expansions in remote interpreting can tell us about interpreter’s participation in a shared virtual space. Journal of Pragmatics. Special Issue ‘Participation in Interpreter-Mediated Interaction’ 107, 165–177. doi: 10.1016/j.pragma.2016.09.011. Cokely, D. (1986). The effects of lag time on interpreter errors. Sign Language Studies, 53(Winter), 341–375. doi: 10.1353/sls.1986.0025. De Meulder, M., Napier, J., & Stone, C. (2018). Designated or preferred? A deaf academic and two signed language interpreters working together for a PhD defense: A case study of best practice. International Journal of Interpreter Education, 10(2), 5–26. Defrancq, B. (2015). Corpus-based research into the presumed effects of short EVS. Interpreting, 17(1), 26–45. doi: 10.1075/intp.17.1.02def. Defrancq, B., & Plevoets, K. (2018). Over-uh-load, filled pauses in compounds as a signal of cognitive load. In M. Russo, C. Bendazzoli, & B. Defrancq (Eds.), Making Way in Corpus-Based Interpreting Studies (pp. 43–64). Singapore: Springer. Gernsbacher, M. A., & Shlesinger, M. (1997). The proposed role of suppression in simultaneous interpretation. Interpreting, 2(1–2), 119–140. doi: 10.1075/intp.2.1-2.05ger. Gile, D. (2008). Local cognitive load in simultaneous interpreting and its implications for empirical research. Forum, 6(2), 59–77. doi: 10.1075/forum.6.2.04gil. Gile, D. (2009). Basic Concepts and Models for Interpreter and Translator Training (Rev. ed.). Amsterdam/Philadelphia: John Benjamins. Goldman-Eisler, F. (1972/2002). Segmentation of input in simultaneous translation. In F. Pöchhacker & M. Shlesinger (Eds.), The Interpreting Studies Reader (pp. 69–76). London/ New York: Routledge. Johnston, T., & Schembri, A. (2007). Australian Sign Language (Auslan): An Introduction to Sign Language Linguistics. Cambridge, UK: Cambridge University Press.

216 Cognitive Overload and Coping Strategies Jorg, U. (1997). Bridging the gap: Verb anticipation in German-English simultaneous interpreting. In M. Snell-Hornby, Z. Jettmarova, & K. Kaindl (Eds.), Translation as Intercultural Communication. Selected Papers from the EST Congress. Prague. 1995 (pp. 217– 228). Amsterdam/Philadelphia: John Benjamins. Lamberger-Felber, H. (2001). Text-oriented research into interpreting. Examples from a case study. Hermes, Journal of Linguistics, 26, 39–64. doi: 10.7146/hjlcb.v14i26.25638. Lee, T. (2002). Ear voice span in English into Korean simultaneous interpretation. Meta, 47(4), 596–606. doi: 10.7202/008039ar. Lee, T. (2011). English into Korean simultaneous interpretation of Academy Awards ceremony through open captions on TV. Meta, 56(1), 145–161. doi: 10.7202/1003514ar Liontou, K. (2012). Anticipation in German-to-Greek SI: A Corpus-Based Approach. (PhD dissertation). University of Vienna. Liu, M., Schallert, D. L., & Carroll, P. J. (2004). Working memory and expertise in simultaneous interpreting. Interpreting, 6(1), 19–42. doi: 10.1075/intp.6.1.04liu. Mackintosh, J. (1983). Relay interpretation: An exploratory study (Unpublished MA thesis). Birkbeck College, University of London, London. Major, G., & Napier, J. (2012). Interpreting and knowledge mediation in the healthcare setting: What do we really mean by ‘Accuracy’? Linguistica Antverpiensia, 11, 207–225. Mazza, C. (2001). Numbers in simultaneous interpretation. The Interpreters’ Newsletter, 11, 87–104. McKee, R., & Napier, J. (2002). Interpreting into International Sign Pidgin: An analysis. Journal of Sign Language and Linguistics, 5(1), 27–54. Napier, J. (2004). Interpreting omissions: A new perspective. Interpreting, 6(2), 117–142. doi: 10.1075/intp.6.2.02nap. Napier, J. (2007). Cooperation in interpreter-mediated monologic talk. Discourse & Communication, 1(4), 407–432. doi: 10.1177/1750481307082206. Napier, J. (2015). Omissions. In F. Pöchhacker (Ed.), Routledge Encyclopedia of Interpreting Studies (pp. 289–290). New York: Routledge. Napier, J. (2016). Linguistic Coping Strategies in Sign Language Interpreting.  Washington, DC: Gallaudet University. Napier, J., Carmichael, A., & Wiltshire, A. (2008). Look-pause-nod: A linguistic case study of a deaf professional and interpreters working together. In P. C. Hauser, K. L. Finch, & A. B. Hauser (Eds.), Deaf Professionals and Designated Interpreters: A New Paradigm (pp. 22–42). Washington, DC: Gallaudet University Press. Napier, J., Rohan, M., & Slatyer, H. (2005). Perceptions of bilingual competence and preferred language direction in Auslan/English interpreters. Journal of Applied Linguistics, 2(2), 185–218. doi: 10.1558/japl.2005.2.2.185. Nicodemus, B., & Emmorey, K. (2013). Direction asymmetries in spoken and signed language interpreting. Bilingualism: Language and Cognition, 16(3), 624–636. doi: 10.1017/S1366728912000521. Seeber, K. G. (2011). Cognitive load in simultaneous interpreting: Existing theories— new models. Interpreting, 13(2), 176–204. doi: 10.1075/intp.13.2.02see. Seeber, K. G., & Kerzel, D. (2012). Cognitive load in simultaneous interpreting: Model meets data. International Journal of Bilingualism, 16(2), 228–242. doi: 10.1177/1367006911402982. Setton, R., & Dawrant, A. (2016). Conference Interpreting: A Complete Course. Amsterdam: John Benjamins. Shlesinger, M. (2000). Strategic allocation of memory and other attentional resources in simultaneous interpreting (Unpublished doctoral dissertation). Bar-Ilan University, Ramat Gan, Israel.

Cognitive Overload and Coping Strategies 217 Shlesinger, M. (2008). Towards a definition of Interpretese: An intermodal, corpus-based study. In G. Hansen, A. Chesterman, & H. Gerzymisch-Arbogast (Eds.), Efforts and Models in Interpreting and Translation Research: A Tribute to Daniel Gile (pp. 237–253). Amsterdam/Philadelphia: John Benjamins. Wadensjö, C. (1998). Interpreting as Interaction (1st ed.). London/New York: Longman. Wang, J. (2016). The relationship between working memory capacity and simultaneous interpreting performance: A mixed methods study on professional Auslan/English interpreters. Interpreting, 18(1), 1–33. doi: 10.1075/intp.18.1.01wan. Wang, J. (2020). Striking a cognitive balance: Processing time in Auslan-to-English simultaneous interpreting. In D. Hunt & E. Shaw (Eds.), The Second International Symposium on Signed Language Interpretation and Translation Research: Selected Papers (pp. 108–131). Washington, DC: Gallaudet University Press. Wang, J., & Fang, J. (2019). Accuracy in telephone interpreting and on-site interpreting: A comparative study. Interpreting, 21(1), 36–61. doi: 10.1075/intp.00019.wan. Wang, J., & Napier, J. (2015). Directionality in signed language interpreting. Meta, 60(3), 518–541. doi: 10.7202/1036141ar.

8 Summary and Recommendations

This empirical research has explored 20 professional Auslan/English interpreters’ spoken English renditions of a Deaf professional’s Auslan lecture at a mock conference setting. I analysed the interpreters’ Auslan-to-English simultaneous interpreting performance as assessed by three external raters, and scrutinised the participants’ post-task interview data to identify their self-perceived challenges in the simultaneous interpreting task and their self-reported coping strategies (Chapter 4). Additionally, I measured the interpreters’ processing time regarding 40 Auslan sentences featuring numbers and negation (Chapter 5), which were among the salient challenges reported by the interpreters. Moreover, I compared the interpreters’ accuracy rate of rendering the following Auslan messages that differed in terms of level: numbers, sentences containing the numbers, and segments of adjacent sentences containing the numbers (Chapter 6). Furthermore, I conducted local analysis of selected interpreters’ accurate English renditions of adjacent Auslan sentences containing numbers in an attempt to investigate the intricate interrelation among interpreters’ processing time, coping strategies for the challenges (e.g., numbers, dense information, and syntactical differences between the source language and the target language), and accuracy of simultaneous interpretation (Chapter 6). I then replicated both the accuracy rate analysis and the local analysis on interpreters’ English renditions of Auslan sentences containing end negation (Chapter 7). This chapter summarises the findings of Chapters 4, 5, 6, and 7, culminating in a proposal of a list of essential knowledge and skills that are necessary for interpreters to interpret accurately from a signed language into a spoken language at formal settings. It then discusses the implications of these findings for Deaf professionals, signed language interpreters who work at formal settings, and spoken and signed language interpreters in general. This chapter also suggests recommendations for further research. Sections 8.1, 8.2, and 8.3 summarise the answers to the three research questions.

8.1 Answers to the First Research Question The first research question was, ‘What are the challenges in simultaneous interpreting from a signed language (Auslan) into a spoken language (English) at a formal setting?’

Summary and Recommendations 219 A key finding is that, according to the grades assigned by the three external raters (two interpreter educators and one practitioner), the ten native signers scored similarly to the ten non-native signers in terms of Auslan-to-English simultaneous interpreting performance. Both groups were similar not only in terms of the overall performance (total score) but also in terms of each of the four sub-scores regarding accuracy, target text features, delivery features, and processing skills. These findings indicate that age of signed language acquisition did not significantly influence the professional interpreters’ signed-to-spoken language simultaneous interpreting performance.These findings send an encouraging message to signed language interpreting students and practitioners who are non-native signers. Results also revealed that there was a level of information loss in each interpreter’s English interpretation of the Auslan presentation.This finding reinforces existing claims (Gile, 2009; Napier, 2016) that simultaneous interpreting is a fundamentally challenging and complex cognitive task, and that simultaneous interpreters may not convey all the source messages accurately, idiomatically, fluently, and perfectly in the target language despite their best efforts, partly due to multitasking and time pressure in simultaneous interpreting. This finding suggests that perfectionism may not be realistic for interpreting students and professional interpreters engaging in simultaneous interpreting. Yet, interpreters can try their best to accurately convey as much essential information in the source text as possible and make strategic additions and strategic omissions as necessary to ensure the accuracy and effectiveness of their simultaneous interpretations. Furthermore, analysis of the interpreters’ qualitative interview data revealed the following self-perceived challenges in working from a signed language into a spoken language in the formal discourse environment: 1 signed language comprehension problems (e.g., unfamiliarity with the Deaf presenter, unfamiliarity with his particular signing style, unfamiliar signs, sign variation, unfamiliarity with the subject matter, and ambiguity in the source text); 2 cognitive overload (e.g., in relation to numbers, long processing time, syntactical differences between Auslan and English, dense information in the source text); 3 target speech production difficulties (e.g., trying to produce comprehensible English sentences, using a formal register at the conference setting, and maintaining coherence in the target speech); and 4 artificial testing conditions (e.g., a two-dimensional video rather than a live presentation, no chance to meet the Deaf presenter beforehand, no opportunity to cooperate with the Deaf presenter during interpreting, no team interpreter, no PowerPoint slides on show, and no real audience). These findings indicate that it is vital for signed language interpreters to (i) familiarise themselves with Deaf presenters and their presentation topics before signed-to-spoken language simultaneous interpreting assignments (e.g., through

220 Summary and Recommendations briefing, joint preparation, or previous experience of working together), and (ii) cooperate with the Deaf presenters through the use of look-pause-nod strategies (Napier, Carmichael, and Wiltshire, 2008) during the interpreting assignments in order to produce highly accurate and effective spoken language interpretations (De Meulder, Napier, & Stone, 2018; Napier et al., 2008).The results also suggest that signed language interpreters need to interact with a wide range of Deaf signers to enhance their own receptive and productive skills in a signed language in order to cope with considerable sign variations in the Deaf community. The findings also indicate that it is necessary for signed language interpreters to actively monitor their signed-to-spoken language simultaneous interpretations in order to ensure accuracy, comprehensibility, grammaticality, coherence, fluency, and appropriate register for the particular discourse environment. Signed language interpreters not only need to improve their signed language proficiency but also need to enhance their flexibility and versatility in using subjectspecific terminology and formal register in their spoken language. In addition, the findings suggest that various resources in real-life conferences – such as preparation materials, briefing, a live three-dimensional signed presentation, eye contact between the Deaf presenter and signed language interpreters, team interpreter, projected PowerPoint slides, being present, and having a real audience – are indispensable for interpreters to achieve highly accurate and effective simultaneous interpretations when working from a signed language into a spoken language.Taken together, these findings suggest that Deaf presenters and signed language interpreters need to work closely together as a team both before and during signed-to-spoken language simultaneous interpreting assignments to ensure the high quality of signed-to-spoken language simultaneous interpreting performance.

8.2 Answers to the Second Research Question The second research question was, ‘What causes cognitive overload in simultaneous interpreting from a signed language into a spoken language?’ Analyses of both the interpreters’ qualitative interview data and their Auslanto-English simultaneous interpretation data provided insights into potential reasons for cognitive overload in signed-to-spoken language simultaneous interpreting. The interpreters reported that rendering numbers, using long processing time, syntactical differences between Auslan and English, and dense information in the source text led to cognitive overload. These factors often co-occurred in highly challenging Auslan sentences or Auslan segments rather than appeared separately. Analysis of the interpreters’ processing time regarding Auslan sentences containing numbers and negation revealed that their processing time not only varied considerably from one person to another but also fluctuated throughout the Auslan source text (Chapter 5). Analysis of the Auslan-to-English simultaneous interpretation data revealed that numbers, especially numbers in clusters that occurred in adjacent Auslan sentences, resulted in momentary attentional deficit, cognitive overload, as well

Summary and Recommendations 221 as some interpretation errors. An interesting finding is that, regarding each of 17 (53% of 32) Auslan numbers, the median processing time for accurate renditions of the number was markedly shorter than the median processing time for inaccurate interpretations of the number. This finding indicates that interpreters who used short processing time to cope with numbers during simultaneous interpreting had greater possibility to render them accurately than interpreters who used long processing time for the numbers. Another revealing finding is that, regarding Auslan sentences whose numbers were interpreted accurately, for each of 19 (59% of 32) Auslan numbers, interpreters’ onset processing time for the respective Auslan sentence (i.e., time interval between the beginning of a source language sentence and the beginning of its corresponding target language sentence) was significantly longer than the interpreters’ number processing time (i.e., time interval between the beginning of an Auslan number and the beginning of its English equivalent). This finding indicates that many interpreters shortened their processing time in order to render the numbers accurately. As expected, representative examples from the Auslan-to-English simultaneous interpretation data illustrated that some interpreters’ overly short onset processing time prevented them from grasping the meaning of source messages. As a result, excessively short onset processing time (see Example 5.1) frequently led to interpretation errors such as source language intrusion, unidiomatic target language expressions, and frequent self-repairs. Example 5.2 illustrated that interpreters’ use of excessively long onset processing time caused serious problems such as unjustifiable omissions of entire Auslan sentences and important details, interpretation breakdowns, overly generalised or summarised renditions, as well as unjustifiable distortions. Interestingly, local analysis of Example 5.3 showed that the interpreter’s processing time changed in length as she switched from one coping strategy to another to deal with adjacent Auslan sentences containing challenges such as numbers and end negation. This example also illustrated that the interpreter’s use of pronouns and concise target language expressions enabled her to catch up with the Deaf presenter and ensure coherence in the target speech segment. These findings indicate that interpreters’ processing time, strategy use, selection of target language expressions, and accuracy of simultaneous interpretation interact with each other in a dynamic manner. Taken together, my findings regarding interpreters’ processing time in Auslanto-English simultaneous interpreting evidence that interpreters need to monitor and adjust the length of their processing time according to the nature and complexity of the current and incoming information, syntactic similarity or difference between the source language and the target language, the speaker/signer’s speed of delivery, the interpreters’ own working memory capacity, strategies, availability of target language expressions, and processing speed, among other factors (see also Gile, 2009; Setton & Dawrant, 2016; Timarová et al., 2014; Wang, 2020). Furthermore, I found that interpreters’ accuracy rate of rendering individual Auslan numbers was substantially higher than their accuracy rate of interpreting

222 Summary and Recommendations individual Auslan sentences containing the numbers, which was considerably higher than their accuracy rate of rendering a sequence of adjacent Auslan sentences containing the numbers. In other words, accuracy regarding numbers was much higher than accuracy regarding individual sentences, which was in turn noticeably higher than accuracy regarding a series of sentences. These findings further evidence that numbers and dense information in the source text (i.e., numbers occurring in adjacent sentences) potentially led to cognitive overload and interpretation errors. Adjacent Auslan sentences featuring numbers near or at the end of the sentences posed an additional difficulty for the interpreters, namely syntactical differences between Auslan and English, which required the interpreters to use sufficiently long onset processing time or employ other strategies (e.g., adding a proxy or subject to start an English sentence before specifying the number). Another key finding is that few interpreters accurately rendered a series of adjacent Auslan sentences featuring end negation. These Auslan sentences typically required interpreters to use long onset processing time in order to understand the meaning of the whole sentence. This finding indicates that syntactical differences between Auslan and English, quite long onset processing time, as well as dense information in the source text potentially led to cognitive overload and inaccurate interpretations. In summary, both thematic analysis of the professional Auslan/English interpreters’ qualitative interview data and local analysis of their Auslan-to-English simultaneous interpretation data revealed that interpreting numbers, using long processing time, syntactical differences between the source language and the target language, as well as dense information in the source text resulted in cognitive overload for the professional interpreters during simultaneous interpreting from a signed language into a spoken language.

8.3 Answers to the Third Research Question The third research question was, ‘What strategies do signed language interpreters employ to cope with the challenges and cognitive overload in simultaneous interpreting from a signed language into a spoken language?’ Both thematic analysis of the interpreters’ qualitative interview data and local analysis of their Auslan-to-English simultaneous interpretation data shed interesting light on strategies they adopted to deal with the difficulties and cognitive overload when working from a signed language into a spoken language. The interpreters reported using the following strategies during the Auslan-to-English simultaneous interpreting task: (i) adjusting processing time (e.g., waiting), (ii) stalling, (iii) producing strategic additions, (iv) making strategic omissions, (v) rephrasing or paraphrasing, (vi) summarising, (vii) compressing information, (viii) generalising, (ix) switching between a free interpretation style and a literal interpretation style, as well as (x) using information stored in long-term memory. In addition, based on an adapted version of Gile’s (2009) Effort Model of spoken language simultaneous interpreting (see Section 2.5 in Chapter 2), my local analysis of selected interpreters’ accurate English renditions of adjacent

Summary and Recommendations 223 Auslan sentences containing numbers in the middle of the sentences revealed that the interpreters used diverse strategies to cope with numbers and dense information. For example, although both Jane and Lauren (see Section 6.2 in Chapter 6) produced accurate English renditions of a segment of Auslan sentences featuring numbers in the middle of the sentences, Jane predominantly used long onset processing time and a free interpretation style, whereas Lauren typically used short onset processing time and a literal interpretation style. Moreover, in-depth analysis of selected interpreters’ correct English renditions of adjacent Auslan sentences containing numbers near or at the end of the sentences revealed that the interpreters adopted the following strategies to deal with numbers, syntactical differences between the source language and the target language, and dense information: 1 The interpreter waited until he/she saw the number and, because of the long onset processing time, produced an accurate and succinct English rendition of the Auslan sentence. 2 The interpreter used medium onset processing time, started interpreting after seeing the topic of the Auslan sentence, employed appropriate strategies (such as repeating a particular part of the topic, inserting a padding phrase [stalling], pausing) to receive more incoming information, and finally rendered the Auslan number near or at the end of the English sentence. 3 The interpreter used short onset processing time, began interpreting after seeing a few signs at the beginning of the Auslan sentence, used appropriate strategies (e.g., repeating a particular part of the topic, stalling, pausing) to absorb more incoming information, and ultimately said the number near or at the end of the English sentence. 4 The interpreter used moderate onset processing time, adhered to the original pseudo-cleft (question-and-answer) structure, and produced an answer that contained the number. Furthermore, local analysis of some interpreters’ accurate English renditions of adjacent Auslan sentences each containing end negation revealed that the interpreters adopted the following strategies to deal with syntactical differences between Auslan and English as well as dense information in the source text: 1 The interpreter waited for almost the entire Auslan sentence, began interpreting only when or after seeing the end negation, and, thanks to the long onset processing time, produced an accurate and idiomatic English rendition. 2 The interpreter used moderate or long onset processing time, started to convey the initial message of the Auslan sentence before seeing the end negation, received more incoming information (e.g., the end negation) while rendering the initial meaning unit into English, then interpreted the end negation appropriately to complete the English sentence. 3 The interpreter followed the Auslan pseudo-cleft (rhetorical question) structure, started to pose an English question before seeing the end negation, then gave a negative answer to the question, thus producing a literal interpretation.

224 Summary and Recommendations 4 The interpreter began interpreting before seeing the end negation, drew on his/her linguistic knowledge, contextual knowledge, and/or thorough pretask preparation, predicted the end negation successfully, hence producing an accurate English rendition of the end negation before seeing it. Unsurprisingly, different interpreters employed different strategies to deal with the same challenge, be it a number or end negation. When dealing with source language sentences of the same syntactic structure, one interpreter’s strategies varied from one sentence to another. These inter-subject and intra-subject differences regarding the use of strategies during simultaneous interpreting may be due to the interpreters’ available processing capacity, processing time, availability of target language expressions, and other factors. Intriguingly, local analysis of Example 6.1 also revealed that the interpreter’s miscomprehension of an ambiguous signed message and lack of suppression of personal biases led to an interpretation error and subsequent self-correction, which, surprisingly, did not cause cognitive overload because the interpreter made strategic omissions to catch up with the Deaf presenter. Another interesting finding is that in Example 6.6, the interpreter’s strategic additions to produce highly accurate renditions of source messages resulted in his unjustifiable omission of important incoming information. These findings indicate that professional interpreters make strategic omissions and strategic additions in simultaneous interpreting from a signed language into a spoken language, and that accuracy of simultaneous interpretation is a slippery and complex concept that merits further research. Taken together, the findings suggest strength in diversity among the professional signed language interpreters in relation to strategy use in simultaneous interpreting from a signed language into a spoken language.

8.4 Key Knowledge and Skills for Success This investigation of both the cognitive process and performance quality of signedto-spoken language simultaneous interpreting at a formal conference setting suggests that the essential skills necessary for interpreters to work successfully in this language direction and discourse environment include, but are not limited to, the following: 1 signed language comprehension skills (e.g., ability to deal with sign variation), 2 joint preparation with the Deaf presenter (e.g., asking questions during a briefing session, interpreting the whole signed presentation during a rehearsal with the Deaf presenter, developing subject-specific terminology, perusing relevant documents), 3 adjusting processing time during simultaneous interpreting, 4 coping with syntactical differences between the source language and the target language, 5 conveying visual-spatial signed messages in an auditory linear spoken language (i.e., dealing with bimodality),

Summary and Recommendations 225 6 using target language appropriately and flexibly to ensure comprehensibility, coherence, grammaticality, idiomaticity, fluency, and formal register, 7 making strategic additions and strategic omissions to enhance the accuracy and effectiveness of simultaneous interpretations, 8 employing interpreting strategies, 9 switching between a free interpretation style and a literal interpretation style if necessary, 10 cooperating with the Deaf presenter and the team interpreter during simultaneous interpreting, and 11 maximising audience design. Results of this study indicate that signed language interpreters’ familiarity with Deaf presenters and their subject matters (e.g., Emily, Amber, and Jane) is crucial for them to produce highly accurate and effective spoken language interpretations. In addition, the results suggest that, for trained and skilled signed language interpreters who are unfamiliar with the Deaf presenters (e.g., Zoe), if they receive sufficient briefing and rehearse interpreting the Deaf presenters’ whole signed presentations during the briefing sessions, they can also produce highly accurate and effective spoken language interpretations. The trust between the Deaf presenters and the signed language interpreters, the Deaf presenters and the interpreters’ use of visual cues (e.g., look-pause-nod strategies) to cooperate with each other during the simultaneous interpreting assignments, teamwork between the two interpreters, the interpreters’ conscious awareness of the non-signing audience’s needs, and the interpreters’ abilities to use formal register in their native spoken language are essential for ensuring the success of the interpreted event. Both the Deaf presenters and the signed language interpreters need to realise that they are jointly responsible for ensuring both the success of the signed presentations and the quality of the signed-tospoken language simultaneous interpretations.

8.5 Strengths and Limitations of This Study I first describe strengths in the design of this research. One strength regarding authenticity is that the Deaf professional actually gave a signed presentation (perhaps using International Sign or a signed language other than Auslan) at a real international conference by using the same PowerPoint slides. Furthermore, the 20 participants in the current study are all nationally accredited professional signed language interpreters. Their signed-to-spoken language simultaneous interpretation data demonstrated diversity and creativity in their interpreting process, interpreting performance, and use of strategies.The use of a large sample size increased the chance of identifying patterns that may generally represent the real-life Auslan-to-English simultaneous interpreting work of the entire population of professional signed language interpreters in Australia. In addition, my in-depth analyses of multiple datasets – the external raters’ analytic assessment of the participants’ Auslan-to-English simultaneous

226 Summary and Recommendations interpreting performance, the raters’ oral and written comments during the analytic assessment, the participants’ retrospective interview data, as well as the participants’ Auslan-to-English simultaneous interpretation corpus in both ELAN and Excel – enabled me to triangulate these resources in order to comprehensively understand both the participants’ simultaneous interpreting process and their interpreting performance. As with any research, however, this study had inevitable limitations in research design. Despite my efforts to simulate an authentic conference, this mock conference differed from real-life conference interpreting assignments in several ways. For example, although using the Auslan video as the stimulus material had the advantage that all participants interpreted the same content delivered by the same Deaf presenter at a consistent speed under the same condition, it had a severe disadvantage that the interpreters were watching a two-dimensional video rather than a real Deaf person signing the content by using a three-dimensional language. Thus, the interpreters in this study did not have an opportunity to meet the Deaf presenter beforehand to ask for clarification about the Auslan presentation, rehearse interpreting the whole Auslan presentation, establish look-pausenod cooperation strategies, or build a good rapport with the Deaf presenter. Based on this experimental study, it is recommended that future studies that use a two-dimensional signed video as the stimulus material for signed-to-­ spoken ­language simultaneous interpreting should consider allowing interpreters to watch the signed video twice, the first time for familiarising with the signed content (similar to having a briefing session with a Deaf presenter in real-life interpreting assignments) and the second time for rendering the signed messages into a spoken language. Another limitation is that the interpreters could not collaborate with the Deaf presenter during the Auslan-to-English simultaneous interpreting task by employing look-pause-nod cooperation strategies (Napier et al., 2008). In other words, in this experiment the Deaf presenter could not maintain eye contact with the professional Auslan/English interpreters during the Auslan-to-English simultaneous interpreting task to check if the interpreters understood his signing, to pause for some seconds for the interpreters to complete their English renditions and catch up with him, or to repeat a particular Auslan sign or segment that the interpreters did not understand. The Deaf presenter produced the formal Auslan monologue while sitting on a chair, which constrained his signing space to some extent; standing up to give a signed presentation would have contributed to a larger signing space and reinforced his formal register. The interpreters in this study could not use a gesture (e.g., signing hold) to indicate to the Deaf presenter to slow down or pause, or nod to the Deaf presenter to signal that he could proceed to the next Auslan segment. However, several interpreters in the current study asked me to pause the Auslan video and go back to rewatch a particular Auslan segment that they did not understand, which is similar to asking a Deaf presenter for repetition in real-life interpreting assignments. Neither did the interpreters have a team interpreter or a real conference audience. The PowerPoint slides were not shown behind or beside the Deaf presenter. As discussed in Chapter 4, these limitations regarding the

Summary and Recommendations 227 artificial testing environment represent poor and contrived working conditions for the signed language interpreters who participated in this study and possibly have affected their Auslan-to-English simultaneous interpreting performance to some extent. These methodological issues need to be avoided in future research on signed language interpreting. In addition, despite the external raters’ analytic assessment and the research assistant’s insightful views on the Auslan-to-English simultaneous interpretation data, my analyses of the data are only as accurate as my abilities allowed. Moreover, given the sample size and considerable variability in the 20 professional Auslan/ English interpreters’ demographic characteristics, findings of this study may not be generalised to the broader community of professional signed language interpreters. Further research should include larger samples of professional signed language interpreters who are relatively homogenous in demographic data (e.g., educational background, level of interpreter education, years of interpreting experience). Another issue relates to my use of the term sentence for dividing the interpreters’ Auslan-to-English simultaneous interpretations into manageable units for data analysis. There is a lack of consensus on what is a sentence in signed discourse, so, for signed language linguists, my use of the term sentence may not have precisely marked the boundaries between signed messages. Still another issue is my use of the term sentence-level accuracy for assessing the a­ cceptability/accuracy of one English interpretation of an Auslan sentence in terms of conveying the essential message (i.e., who does what to whom). That is, my use of the term sentence-level accuracy does not indicate perfection or 100% accuracy as used in the majority of the existing interpreting studies literature and national interpreting accreditation tests, but suggests a minimum pass versus fail. My use of the term accuracy in this sense may inspire other researchers to consider the complex concept accuracy as a spectrum rather than a definitive notion.

8.6 Implications of the Findings There has been little research on professional interpreters’ signed-to-spoken language simultaneous interpretations at formal settings. This section discusses the implications of the findings of the present study for Deaf professionals, signed language interpreters working at formal settings, and spoken and signed language interpreters in general. 8.6.1 Implications for Deaf Professionals Findings from this study have the following implications for Deaf professionals whose signed lectures are rendered by signed language interpreters into spoken languages at formal settings: 1 Deaf presenters may choose highly competent and well-trained signed language interpreters with whom they are highly familiar and regularly work. Alternatively, Deaf presenters may choose qualified and well-trained signed

228 Summary and Recommendations

2

3

4

5 6

7 8

language interpreters with whom they are unfamiliar and did not work regularly; in this case, the Deaf presenters need to provide useful preparation materials (e.g., PowerPoint slides, glossary, fingerspelled words, technical terms both in signed language and spoken language, relevant signed presentations) to the interpreters, give the interpreters sufficient briefing and an opportunity to rehearse interpreting the entire signed presentation, and establish look-pause-nod cooperation strategies (De Meulder et al., 2018; Napier et al., 2008) that can be employed during the signed-to-spoken language simultaneous interpreting assignments. To facilitate signed language interpreters’ comprehension of Deaf presenters’ signed messages and sign variation, the Deaf presenters need to brief the interpreters about multiple sign choices for a particular concept before the signed presentation, or use the same sign for the same concept consistently throughout the signed presentation. When Deaf presenters introduce key concepts such as technical terms, they need to sign slowly, produce the sign first (together with mouth pattern), then fingerspell the English equivalent for that sign, then repeat the sign, in order to ensure the clarity of meaning and give the interpreters sufficient time to process and remember the new sign. When there are no established signs for particular words or concepts, Deaf presenters may use synonymous signs that bear a similar meaning as those particular words or concepts and use clear mouth patterns to express the meaning clearly. Deaf presenters need to use PowerPoint slides that show the structure of their signed presentation, key points, and numbers, in order to guide the interpreters throughout the signed presentation. Deaf presenters need to produce numbers, fingerspelling, a list of items, and rhetorical questions slowly and clearly, pause after producing these lexical items and/or signed sentences, and pause between signed sentences or sections in order to give the interpreters sufficient time to process the signed information, cope with syntactical differences between a signed language and a spoken language, and produce idiomatic and accurate spoken language renditions. Due to numbers, dense information, and syntactical differences between a signed language and a spoken language, signed language interpreters may need sufficiently long onset processing time in order to re-order source sentence components, switch attention from one task to another, and produce accurate, idiomatic, and comprehensible target language renditions for the non-signing audience. Deaf presenters need to produce key points (e.g., arguments, conclusions) slowly and clearly, pause after each of the important ideas, and repeat the essential information where appropriate. Deaf presenters need to maintain eye contact with signed language interpreters during signed-to-spoken language simultaneous interpreting assignments in order to check if the interpreters understand the signed messages and have sufficient time to finish interpreting a signed segment. If the

Summary and Recommendations 229 interpreters experience difficulties in understanding specific signs, the Deaf presenters need to repeat those signs or rephrase the concepts so that the interpreters can understand the meaning of the source messages. If the interpreters need more time to finish interpreting a signed sentence, the Deaf presenters need to pause and wait for the interpreters to complete their spoken language rendition of that signed sentence. 9 Deaf presenters and signed language interpreters need to work as a team before and during signed-to-spoken language simultaneous interpreting assignments, in order to ensure that their signed presentation is successful and that the interpreters’ signed-to-spoken language simultaneous interpretations are accurate and effective for the non-deaf audience and context. The Deaf professional in the current study is a good presenter who has extensive experience of working with signed language interpreters. For example, he produced signs clearly at a good pace (e.g., producing numbers clearly on both hands to avoid misunderstanding, pausing after numbers, pausing between sentences, using discourse markers to indicate a new section) and incorporated many signed language features (e.g., use of space, depicting signs, role-shift, pseudo-cleft structures) (for further details on Deaf professionals’ strategies for working effectively with interpreters, see Haug et al., 2017).

8.6.2 Implications for Signed Language Interpreters Working at Formal Settings Results of this research contribute new knowledge that can benefit signed-tospoken language simultaneous interpreting at formal settings. Based on my findings, I recommend the following best practices to signed language interpreters who work from a signed language into a spoken language at formal settings: 1 Interpreters need to prepare for the signed-to-spoken language simultaneous interpreting assignments thoroughly by becoming familiar with the Deaf presenters, their background and professional expertise, presentation topics, and technical terms. 2 Interpreters need to ask the Deaf presenters for a briefing to become familiar with their signing style, signing speed, and sign choices. Rehearse interpreting the entire signed presentation into a spoken language. 3 During the signed-to-spoken language simultaneous interpreting assignments, interpreters need to cooperate with the Deaf presenters by maintaining eye contact and using hand gestures where necessary in order to ensure mutual understanding and good pace of the signed presentations. 4 When working from a signed language into a spoken language in simultaneous mode, interpreters need to have clear visual access to the projected PowerPoint slides and cooperate with both the Deaf presenters and the team interpreters.

230 Summary and Recommendations 5 Interpreters need to not only watch the Deaf presenters’ hands for manual signs but also pay close attention to their mouthing (for a sign with multiple possible equivalents in English, the Deaf presenters may mouth their preferred English technical term) and non-manual features (e.g., facial expression, eyebrow, head tilt, body lean) in order to understand the signed presentations correctly and thoroughly. 8.6.3 Implications for Spoken and Signed Language Interpreters in General Findings of this research also have implications for spoken and signed language interpreting in general. According to my observations at real-life conferences, many hearing presenters focus primarily on delivering their speeches within the allocated time; therefore, they seldom pay attention to spoken language interpreters who work in simultaneous mode and may lag far behind the presenters due to syntactical differences between the source language and the target language. In light of the partnership and cooperation between Deaf presenters and signed language interpreters, it is recommended that hearing speakers, when they can see their spoken language interpreters working in the booth, maintain regular eye contact with the interpreters to ensure that the interpreters are comfortable with their pace of delivering the source speech and are following up. If this practice is adopted, spoken language conference interpreters may feel less like conduits or machines. Moreover, the methodologies and findings of this research have implications for spoken and signed language interpreting training. For example, the assessment rubric for evaluating participants’ Auslan-to-English simultaneous interpreting performance can be adapted for assessing spoken and signed language simultaneous interpreting performance of different language combinations.The method of annotating salient features of simultaneous interpreting performance using ELAN can be incorporated into spoken and signed language interpreting training programs in order to provide detailed individual feedback to interpreting students. Furthermore, the methodology of measuring different types of processing time in simultaneous interpreting can be used by educators and researchers in spoken and signed language simultaneous interpreting training programs to explain individual differences in cognitive processing and the dynamic nature of simultaneous interpreting. In addition, interpreter researchers, educators, and trainees can carry out local analysis of cognitive overload and failure sequences in spoken and signed language simultaneous interpreting in order to explain the domino effect of problem triggers in the source text, the consequences of interpreters’ inappropriate allocation of attention, the size of their processing unit at various junctures, processing time patterns, the impact of the length of processing time on interpretation style, and interpreting strategies. Students can use ELAN and local analysis to reflect on their processing time and use of strategies in simultaneous interpreting. The signed-to-spoken language simultaneous interpreting model that I adapted from Gile’s (2009) Effort Model of spoken language simultaneous

Summary and Recommendations 231 interpreting can be applied to both the training and research of signed-to-spoken language simultaneous interpreting of other language combinations such as ASL-to-English simultaneous interpreting and BSL-to-English simultaneous ­ interpreting. As illustrated by the examples in Chapters 6 and 7, this adapted model has strong explanatory power in accounting for the underlying reasons for interpretation errors in signed-to-spoken language simultaneous interpreting.

8.7 Suggestions for Further Research A number of issues have arisen from this research project that necessitate future research. One key area for further investigation is integrating multiple research methods in order to explore factors resulting in cognitive overload in spoken and signed language simultaneous interpreting. For example, it would be interesting to combine various methods – (i) psycho-physiological measures such as pupillometry to measure interpreters’ cognitive load during simultaneous interpreting, (ii) retrospective interview of interpreters on perceived triggers of cognitive overload and on perceived coping strategies, (iii) measuring processing time at various points to shed light on temporal aspects of simultaneous interpreting, and (iv) local analysis of inaccurate and accurate renditions of extremely challenging source language segments – in order to triangulate the various data sources, shed light on possible factors that lead to cognitive overload during simultaneous interpreting, and provide insights into interpreters’ use of strategies to avoid cognitive overload. In addition, essential further research would entail detailed investigations of professional, trained, and highly skilled signed language interpreters’ authentic signed-to-spoken language simultaneous interpretation data at formal settings such as seminars, university lectures, doctoral dissertation defence, national conferences, and international conferences. For example, a descriptive study on features of formal register in signed lectures produced by a wide range of Deaf professionals would be highly beneficial, as it can help interpreting students recognise signed language cues that require them to use a formal register when rendering formal signed presentations into a spoken language. Empirical studies of highly skilled signed language conference interpreters’ preparation methods, interpreting strategies, decision-making process, and teamwork techniques regarding working from a spoken language into a signed language and vice versa can provide useful knowledge to signed language interpreting training programs. It would be useful to examine the features of professional signed language interpreters’ signed-to-spoken language interpretations that appropriately represent a Deaf professional’s presentation style and personality. Analysis of the interpreters’ intonation and lexical choices may shed light on this issue. In addition, an investigation of professional interpreters’ signed-to-spoken language interpretations for Deaf professionals who are proficient in multiple signed languages (e.g., the Deaf presenter in the current study) can throw light on how the contact among the various signed languages influences the interpreters’ signed language comprehension and their signed-to-spoken language interpreting performance.

232 Summary and Recommendations Another topic that warrants further attention is directionality in signed language interpreting. It would be particularly interesting to explore challenges in simultaneous interpreting from a spoken language into a signed language and vice versa. Given the paucity of empirical research into directionality effects on signed language interpreters’ processing time and strategies, further research in these areas involving naturally occurring signed language interpretation data would be of immense interest to signed language interpreting researchers, educators, and students. Although not a comprehensive list, these suggestions for further research are given with the goals of bridging gaps in signed language interpreting studies and enriching people’s understanding of the linguistic, sociolinguistic, communicative, and cognitive aspects of signed language interpreting. Future research needs to consider the interrelation among a speaker/signer, interpreters, and users in a discourse environment, as interpreting is not only a linguistic and cognitive activity, but also a communicative activity with a purpose in a context.

8.8 Final Thoughts This book ends where it began. In the opening anecdote, a team of highly skilled Deaf and hearing signed language interpreters produced ASL and spoken English simultaneous interpretations of Dr Robert Adam’s BSL presentation, which drew our attention to numerous gaps between Deaf signed language interpreters and hearing signed language interpreters in terms of training and practice. This book also ends by highlighting Mr Colin Allen’s key message of his Auslan presentation which was interpreted by the 20 professional Auslan/English interpreters into spoken English at a mock conference in the current study. Colin’s key message is that Deaf people around the world are not yet able to truly enjoy even basic human rights (Haualand & Allen, 2009). On both formal occasions, it was very important for the professional signed language interpreters to accurately and effectively convey the Deaf professionals’ powerful messages to non-signing audience members who could not understand the signed presentations. The essential skills, cognitive processing, decision-making process, and strategies that the professional interpreters employed to transfer the signed messages accurately and effectively into a spoken language continue to fascinate me and merit further investigation. I close this book with a hope that my exploratory study of the quality, cognitive overload, and strategies of simultaneous interpreting from a signed language into a spoken language will inspire further research into spoken and signed language simultaneous interpreting.

References De Meulder, M., Napier, J., & Stone, C. (2018). Designated or preferred? A deaf academic and two signed language interpreters working together for a PhD defense: A case study of best practice. International Journal of Interpreter Education, 10(2), 5–26.

Summary and Recommendations 233 Gile, D. (2009). Basic Concepts and Models for Interpreter and Translator Training (Rev. ed.). Amsterdam/Philadelphia: John Benjamins. Haualand, H., & Allen, C. (2009). Deaf People and Human Rights. Retrieved from https:// www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=2ahUKEw jMsc7KtOTnAhWCYisKHWfiBGAQFjABegQIBRAB&url=https%3A%2F%2F www.rasit.org%2Ffiles%2FDeaf-People-and-Human-Rights-Report.pdf&usg=AOv Vaw1JAQTTPOPRAQV7ZVcQ6X_E. Accessed on February 22, 2020. Haug,T., Bontempo, K., Napier, J., Nicodemus, B., van den Bogaerde, B., &Vermeerbergen, M. (2017). Deaf leaders’ strategies for working with signed language interpreters: An examination across seven countries. Across Languages and Cultures, 18(1), 107–131. doi: 10.1556/084.2017.18.1.5. Napier, J. (2016). Linguistic Coping Strategies in Sign Language Interpreting.  Washington, DC: Gallaudet University. Napier, J., Carmichael, A., & Wiltshire, A. (2008). Look-pause-nod: A linguistic case study of a deaf professional and interpreters working together. In P. C. Hauser, K. L. Finch, & A. B. Hauser (Eds.), Deaf Professionals and Designated Interpreters: A New Paradigm (pp. 22–42). Washington, DC: Gallaudet University Press. Setton, R., & Dawrant, A. (2016). Conference Interpreting: A Complete Course. Amsterdam: John Benjamins. Timarová, Š., Čeňková, I., Meylaerts, R., Hertog, E., Szmalec, A., & Duyck, W. (2014). Simultaneous interpreting and working memory executive control. Interpreting, 16(2), 139–168. doi: 10.1075/intp.16.2.01tim. Wang, J. (2020). Striking a cognitive balance: Processing time in Auslan-to-English simultaneous interpreting. In D. Hunt & E. Shaw (Eds.), The Second International Symposium on Signed Language Interpretation and Translation Research: Selected Papers (pp. 108–131). Washington, DC: Gallaudet University Press.

Appendices

Appendix A Auslan Transcription Conventions Specific instances deaf

Explanation and example

An English word (gloss) in small capital letters represents an Auslan sign. This English word is the mostly commonly associated and/or nearest translation of the sign. how-many When more than one English word is needed to gloss an Auslan sign, the English words are separated by a hyphen. d-a-t-a When an English word is fingerspelled, the letters in the word are separated by a hyphen. write* An asterisk after a sign indicates that the sign is emphasised by the signer. approve (1) approve (2) Numbers indicate that two different signs are used for the same gloss, one immediately after another. point-left/point-right The signer has pointed to a location to reference a specific place or established entity or entities. For example, country point-left -left/-right Directions following a sign indicate placement of the sign by either the left or right side of the signer’s body. For example, category-left human rights-right .. A noticeable pause, equivalent to a comma in written English. q The facial expressions and head movements used to indicate a question is being asked (often rhetorical). hd A head nod indicates affirmation. neg A headshake indicates negation. // the end of an Auslan sentence text↑ Marked rising intonation

Appendices

235

Appendix B Auslan Source Text and a Literal English Translation Deaf People And Human Rights (The first three minutes of the Auslan video as a warm-up.) hello my name c-o-l-i-n a-l-l-e-n my sign name what? colin// Hello my name is Colin Allen and this is my name sign –demonstrates. this me will talk about what? from me recent work with the world f-e-d-e-r-a-t-i-o-n o-f deaf wfd// I’ll be talking to you about my recent work with the World Federation of the Deaf,WFD. me work involve that for about three year project this project what focus on t-o erm collect collect about deaf people their human rights// I worked on a three-year project that focused on collecting data about Deaf people and their human rights status. so altogether give us picture how many what deaf people situation their living// This information will give us a clear picture of the Deaf population, and the situation for them in their everyday lives. that research funding support by who? how start what? that deaf organisation where? in point sweden// So, how did this research start and who funded it? Well, a Deaf organisation in Sweden. that sweden a-s-s-o-c-i-a-t-i-o-n o-f deaf point want t-o find number situation deaf people their// The Swedish Association of the Deaf wanted to find out the size of Deaf populations and what their lives were like in different countries. why? point sweden have involve different different project past relate with those developing countries// The reason for this is that the Swedes had for some time been involved in various projects in a number of developing countries. want have clear big picture// So that’s why they wanted a clear view of the ‘big picture’ regarding Deaf people. s-o they negotiate possible yes get funding but need have side-by-side partner another organisation// They managed to negotiate potential funding for the project, but this was reliant on finding a partner organisation. s-o they contact world-federation-of-the-deaf// So, they contacted the WFD, wfd board itself excite at last have good research because last time research in 1991// whose board were thrilled that there was an opportunity to get some good research as the last such study was conducted in 1991.

236 Appendices

nothing-since-then need update information about deaf people their lives// Since that year, no comprehensive studies have been completed, so information about Deaf people and their lifestyles needed updating. s-o both sweden deaf a-s-s-o-c plus wfd both point apply funding t-o organisation based in sweden successful funding// So the Swedish Association of the Deaf and the WFD applied to a funding source based in Sweden and were successful in their application. get funding how what they decide what invite me a-s project coordinator t-o research area seven r-e-g-i-o-n-s// As to how the funding was used, they decided to invite me to act as project coordinator and as such, I oversaw research in seven different regions. seven different r-e-g-i-o-n-s where? will explain later// I will detail where these regions are later, how? will give you think about-it// but in relation to how the project was conducted, I will let you think about that for a moment. Deaf People And Human Rights (The remaining 17 minutes of the Auslan video for Auslan-to-English simultaneous interpreting.) No slide. The deaf professional’s spontaneous introduction on his research project 1 2 3 4 5 6

how that work?// So how it worked was that I responsible t-o write* questions based on different different categories altogether number of categories1// I was responsible for compiling and drafting a list of questions across a number of categories. that discussion with have special group committee look analyse approve questions then that english text// I then discussed the questions and categories with a committee who gave the final approval of a written English version of the questions, then I take sign international-sign// which I then translated into International Sign. altogether ready available in dvd plus paper together distribute t-o all deaf organisations who member o-f world-federation-of-the-deaf // A DVD and a hard copy of the written English questions were distributed to all the Deaf organisations that are members of the World Federation of the Deaf (WFD). other too some country nothing national a-s-s-o-c deaf link worldfederation-of-the-deaf still post maybe have group deaf people in that country// In those countries where there is no national association of the Deaf affiliated to WFD, the package was still sent to any Deaf groups that were in existence.

Appendices

237

7 well send-out after that receive receive receive receive their their their different different different r-e-g-i-o-n// Then the responses to the survey came in from the various regions. 8 develop write research consolidate report for their specific r-e-g-i-o-n// Reports were generated by interpreting the research data about the situation for each specific region, 9 that mean I finish wrote seven different r-e-g-i-o-n-s report report report// meaning I wrote seven different reports, one for each region. 10 after that report-submit that how what? end make summary from seven report t-o true picture o-f what what so that document called what deaf people their human rights// After those were done, to finish the project off, a summary was generated from the seven reports that gave an accurate overall picture of what was going on. So, this document was called ‘Deaf people and their human rights.’ 11 category deaf people circumstances human rights merge good bad what?// It brought together the two concepts of Deaf people and human rights to give a picture that had both good and bad examples. 12 that what I will talk about// That’s what I will talk about. Slides 4 and 5 13 where seven region region region region where?// So, where are the seven regions? 14 first one where? in point r-u-s-s-i-a but all-over-area east (Note: The first example of the Deaf professional’s slip of the hand. He forgot to sign europe here, which was shown on his PowerPoint slide.) middle asia group all-over// The first is ‘Russia,’ but it also includes East(ern Europe) and Middle Asia. 15 after that where? asia-pacific area australia include too// After that, the next one is ‘Asia Pacific,’ and Australia is included in that region. 16 after that finish where? south-america (1) where south-america (2) area// Next is ‘South America,’ 17 after that finished next (Note: The second example of the Deaf professional’s slip of the hand. He forgot to sign mexico here, which was shown on his PowerPoint slide.) middle america c-a-r-i-b-b-e-a-n-s area group// followed by ‘(Mexico,) Central America and the Caribbean.’ 18. after that travel-far where east south africa// After that it’s across the ocean to ‘Eastern and Southern Africa.’ 19. that finish then move-to west central africa// Next is ‘Western and Central Africa’

238 Appendices

20 then where? last one arab a-r-a-b region// and the last region is called the ‘Arab Region.’ 21 s-o that mean I start t-o go go go different different region region region visit country meet special group for region discuss issues after receive finish research// That meant I had to travel to each region to meet specially selected groups to discuss their issues after survey results for their region came back. 22 group report I wrote d-r-a-f-t show ok-yeah adapt that seven report// These discussions took what had been draft reports and tweaked them accordingly to produce the seven reports. Slide 6 23 last* altogether seven reports I done true very very one big topic write me write? no// After bringing together the seven reports that I had written, there was one final thing left to do. It wasn’t me who wrote the final combined report. 24 have another person t-o like outside from that project me work involve last three years// We brought in a fresh set of eyes, an independent person from outside the project, as I had been working in the project over the past three years. 25 s-o out new eye wink quotation-markers point very-last bringtogether seven report analyse analyse// So, we had a person with a fresh perspective to compile the seven regional reports in order to form the final report. 26 person name what? h-i-l-d-e h-a-u-a-l-a-n-d her sign name what? library (hairclip) from where? Norway// Her name was Hilde Haualand. This is her name sign – demonstrates – and she is from Norway. 27 I more like lead give her d-a-t-a she analyse analyse workout workout workout ok?// So, I supplied her with all our data and she then independently audited the information. Slide 7 28 so altogether how many responses we receive survey (research) responses receive how many? ninety-three// So, altogether we received in total 93 responses to our survey. 29 but ninety-three those full link w-f-d? no// However, not all of these were from organisations affiliated to the WFD. 30 they group sixty-eight member w-f-d other group those twenty-five not member w-f-d// Sixty-eight responses were from members of the WFD, 25 were from non-member groups.

Appendices

239

31 s-o that means those twenty-five country nothing national a-s-so-c o-f deaf// So those 25 countries do not have a National Association of the Deaf. 32 that how process o-f member t-o w-f-d must have national organisation// That’s the process of gaining membership of WFD – having a national organisation. 33 e-g same australia well what deaf australia s-o full member world-federation-of-the-deaf// For example, in Australia, we have Deaf Australia, which is a full member of the WFD. 34 really exciting for large number ninety-three// We were very excited about the number of 93 responses. Slide 8 35 okay had t-o clear goal o-b-j-e-c-t-i-v-e-s o-f research// Okay, we had to clearly express the objectives of this research project. 36 well hilde me discuss make decide put-on d-i-a-g-r-a-m that have four area area area area// So, Hilde and I discussed this and made the decision that we would present the data in a diagram with four overlapping circles. 37 believe that really same minimum basic really their deaf human rights// We believe each part of the diagram sheds light on the fundamentals of the human rights situation of Deaf people at that time. 38 first area clear what? point sign language// The first circle in the diagram illustrates rights regarding signed languages, 39 point sign language hand belong deaf community language have right t-o access language acquire area// the languages of the Deaf Community and Deaf people’s rights to access and acquire those languages. 40 have sign language next area what? bilingual education// The second circle reveals the situation regarding access to a bilingual education. 41 that mean what? have point your home country sign language plus written country language parallel move-forward-together// Meaning, that you have an education in your country’s home signed language as well as the written form of your national language in parallel. 42 mean child-grow-up have sign include education// For a Deaf child growing up, access to their signed language is essential, as is signing in their education. 43 have sign same time other area what? sign language interpreter training// So, what is represented in the third circle? Signed language interpreter training is third.

240 Appendices

44 training what? high level training t-o receive high training about different area area area area area area become qualified q-u-a-l-i-fi-e-d qualified receive sign language interpreter// More specifically, the high standard that represents comprehensive training across a wide range of settings.This should lead to a qualification as a qualified sign language interpreter. 45 mean that have sign language over-there education have sign language over-there access access access a-c-c-e-s-s-i-b-i-l-i-t-i-e-s different places places places have sign language (Note: The third example of the Deaf professional’s slip of the hand. He forgot to sign interpreter here, which can be inferred from the context.) available there there there// So, starting with rights relating to signed languages themselves then following this through the education sphere, we then get to accessing the many different places that provide signed language (interpreters), 46 so fourth area access a-c-e a-c-c-e-s-s-i-b-i-l-i-t-y// so the fourth circle focuses on that very subject – accessibility. 47 so four area area area area circle foundation-support x 2 deaf standon-top can walk! can walk no I mean can move-on// Those four key areas of human rights represent the potential foundation that Deaf people could stand on and then walk on. But I do not mean literally walk on – I mean it in a sense that they can build on that and move forward. 48 i-f one (foundation post) fall-down person-stuck hit-wall e-g foundation-fall-down what? sign language finish crumble// And if one of those four pillars of that foundation were to collapse, then Deaf people’s progress would be blocked. For example, if the pillar was ‘signed languages,’ then everything would crumble. 49 well human rights mean I have right to acquire my language their language ok?// Access to our signed language is the most fundamental human right for myself and all other Deaf people. Spontaneous Introduction to the Un Convention on the Rights of Persons with Disabilities 50 so survey (research) based on concepts how concepts (four) those from where? that u-n convention c-o-n-v-e-n-t-i-o-n on rights people person person with their disability// The research is based on four key concepts from the UN Convention on the Rights of Persons with Disabilities. 51 u-n convention passed finish point u-n before 2006// The UN convention was passed at the UN in 2006. 52 that document very important for many many different different reason// That document is very important for many different reasons.

Appendices

241

53 australia government finish sign r-a-t-i-f-i-e-d in 2008// The Australian government ratified the convention in 2008. 54 one important part o-f that (hesitation) agreement convention that have five different article article article article article relate what? sign language out of thirty article article article// In one important part of the convention there are five articles that specifically relate to signed languages out of a total of 30 articles. 55 but have eight times mention mention mention sign language put-in put-in put-in// However, signed languages are mentioned eight times. 56 well u-n convention really first document up-till-now declare sign language real part of human rights// This convention is the first in history to declare signed language access as a human right. 57 why? well in first article number two have explain like definition what point mean for different headings// Article Two, which deals with definitions within the convention, 58 one point talk over language point language say their mean what spoken language same equal a-s sign language// covers the definition of ‘language’ and it cites spoken languages as being on a par with signed languages. 59 that really quotation-markers shrug-shoulders shake-head no quotation-markers shake-head evidence t-o show sign language spoken equal// This document is the first to recognise the evidence that shows signed languages are equal to spoken languages. Slide 9 60 so have one article talk about what? important give-to deaf person person person imagine access t-o different information information information// There is one article that talks about the importance of giving Deaf people the means to use and access different information. 61 second well also well look observe have can accept use their sign well government finish approve recognise their sign language like promote their research sign language different different that area area link-up o-f convention// At another point the convention mentions research showing the benefits of Deaf people having their signed language accepted and accessible, of governments recognising their national signed languages, and of research into signed languages being encouraged in different regions. These things are mentioned in the convention itself.

242 Appendices

Slide 10 62 s-o well like have how many o-f those country finish recognise their deaf people out-there? only seventy-seven respond from ninety-three country say yes deaf people recognise equal// So, how many of those countries that responded to our survey replied that they recognised Deaf people in their country? Only 77 out of the 93 countries replied that Deaf people were recognised as equals to hearing people. 63 but what about sign language what? finish they have approve confirm point sign language? only forty-four country nod say government finish approve sign language o-u-t o-f ninety-three country hmmm// But, what about the formal recognition of the national signed language itself? Only 44 countries replied that their governments have formally recognised their national signed language, out of 93 countries. 64 that surprise under half altogether// That equates to less than half of the total number of respondents. 65 what about sign language research? have have have what what what? only have fifty-four finish research sign language make that book lick-finger-flick-page flick-page flick-page d-i-c-t-i-o-n-a-r-y// And, how many countries had signed language research happening? Only 54 of the respondents reported that research had been undertaken and that dictionaries of their signed languages had been produced. Slide 11 66 67

important talk over recognise sign language why?// The recognition of signed languages is a very important topic to discuss.Why is this so? doesn’t mean all countries those forty-four have same level approval// Of the 44 countries that have recognised their signed language, not all have done so to the same extent. 68 well each country government have different different different definition definition e-g constitution c-o-n-s-t-i-t-u-t-i-o-n first second legislation l-e-g-i-s-l-a-t-i-o-n third policy p-o-l-i-c-y other categories// Each of those countries’ governments has to work within different definitions and types of recognition, such as within their constitution, within legislation, within policy, or within other types of documents. 69 s-o those forty-four country say yes my government finish recognise my sign only ten of those forty-four yes where? constitution level// Of those 44 respondents that answered yes to their government recognising their signed language, only 10 of those 44 had done so at the constitution level, 70 other group nineteen finish approve through legislation// whereas 19 had recognition within legislation,

Appendices

243

71 other nineteen country those finish approve their sign language through what? policy// with another 19 countries having done so within policy. 72 so that means only ten country round world actually recognise sign language ten// So that means that only ten countries around the world have actually, formally recognised their signed languages. 73 nineteen point well legislation also high level o-f approve// However, the 19 countries that recognised signed languages within their legislation can be seen as also having ‘ratified’ the language at a high level. He skipped Slide 12. Slide 13 74 75 76 77 78 79 80 81

now talk over access t-o education// Now I’d like to talk about the subject of access to education. remember me talk over one circle relate to bilingual education// If you remember one of the circles was related to bilingual education, s-o want to know how many country finish recognise deaf children access t-o education// so we wanted to know how many countries had recognised Deaf children’s access needs in educational terms. group those country respond say eighty-eight country finish say yes they recognise their children right t-o access their education good eighty-eight wow// From the respondents, 88 countries answered that Deaf children’s rights to access an education had been recognised, which seems an impressive number. what about legislation level what? well relate t-o deaf education only sixty-one country say nod have have legislation relate t-o deaf education// But, what existed at a legislative level regarding Deaf education? Only 61 countries replied that they had legislation pertaining to Deaf education. another one how many deaf schools?// Another question we asked in our survey was about how many Deaf schools they had. why? deaf schools very important their child life grow-up deaf// The reason for asking about Deaf schools is that these places often form an important part in Deaf children’s lives. them how many? eighty-eight country have deaf schools// Of the respondents, 88 replied that they had Deaf schools in their country.

Slide 14 82 now want move talk about method education// Now I’d like to focus on educational approaches.

244 Appendices

83 why? know what? education have different different different oral (Note: The fourth example of the Deaf professional’s slip of the hand. He meant to sign bilingual education here, which can be figured out from the following content.) total communication t-o-t-a-l-c-o-m-m-u-ni-c-a-t-i-o-n total communication speak-sign-same-time oral only oral only mean what? sign not-allowed speak-speak// As is well known, there are many approaches to Deaf education, ranging from oral approaches (bilingual education), to ‘Total Communication’ which incorporates English features into signing, to an ‘oral-only’ approach which bans the use of signed language and relies solely on spoken language and lip reading. 84 those three want analyse those eighty-eight country how-many how-many how-many what? interesting interesting// We wanted to know from those 88 countries how many schools were using these different approaches, and the results were very interesting. 85 bilingual have nod plus have other altogether country have what twenty-three country have that// There were 23 countries that reported having a bilingual education option alongside other approaches. 86 other group say have total communication plus other methods education thirty-five country// Another group of countries, numbering 35, described having a total communication approach with other methods of education available. 87 those only total communication speak-sign altogether thirty-one// Thirty-one respondents replied that their country had a total communication approach only, which incorporates speaking and signing at the same time. 88 other group only oral sign ban five// A further five countries reported an oral-only approach that bans signed language. 89 means bilingual education high level not// This meant that bilingual education methods were not in high use. 90 more still philosophy apply total communication hands oral link link well// and that Total Communication was still the prevalent philosophy regarding Deaf education. He skipped Slide 15. Slide 16 91 talk over access t-o different government// When looking at access to different governments, 92 one article that link convention that say those sign language (Note: The fifth example of the Deaf professional’s slip of the hand. He forgot to sign interpreter here, which was written on his PowerPoint slide.) must have qualified professional level t-o provide access for out-there deaf people//

Appendices

245

there is one article from the UN Convention that says signed language (interpreters) must be qualified at a professional level to give Deaf people proper access.   93 s-o that important topic t-o analyse how-many relate-to sign language interpreters// So, another important area for us to explore was numbers in relation to signed language interpreters.

Slide 17   94 think-about concept access relate-to interpreters// Bringing together the two concepts of access and interpreters,   95 well ask how many deaf people can access t-o government how many?// We asked in our survey about how many Deaf people could access government. 96 respond eighty-two country say yes deaf people can access t-o government wow// Eighty-two countries replied positively that Deaf people could access government – an impressive response.   97 but what about sign language interpreter services have have have? what? provide have? sixty-two country// But what about signed language interpreting services? Sixty-two countries replied that they had signed language interpreting services.   98 so sixty-two interpreting services have but how many those finish have training qualified professional interpreters qualified have how many? only forty-three hmmmm// Of those 62, how many had interpreters qualified to a professional level? Only 43 responded positively.   99 that forty-three group those how many have c-o-d-e-o-f-e-t-h-i-cs how many? only thirty-one// And of those 43, how many had a code of ethics in their country? Only 31. 100 important part that group look-over-there look-here government responsible pay interpreters have? how many country what? thirty-two// An important part of the research is whether governments take responsibility for paying for interpreting services. Of those surveyed, only 32 countries replied that they did. 101 s-o talk about sign language interpreting service have sixty-two those thirty-two government responsible pay// So, then you have to consider that 62 countries replied that they had signed language interpreting services, but only 32 responded that their government paid for interpreting services. 102 mean that group who responsible could b-e deaf people have-to pay interpreters or interpreters heart volunteer interpret mean link no qualification// This meant that interpreting services were possibly being paid for by Deaf people, or were being given voluntarily by interpreters – meaning without qualifications.

246 Appendices

103 that very vague vague vague// So the picture in this regard is very mixed. 104 interesting show basic concept how many sign language interpreter that why? link t-o two area area about interpreter training and access t-o different services// It’s interesting to note that the basic concept of just how many signed language interpreters a country has is linked to the two circles of interpreter training and access to different services. Slide 18 105 s-o altogether research talk like many many block (prevent) deaf people no few country those put-down for deaf people access t-o education government or perception equal same same other people// So, in summary, what does the research tell us? Firstly, we found that only in a few countries were Deaf people marginalised in their societies, and that their situation was not good in terms of access to education and government and also in terms of how well Deaf people were seen as equal to other people. 106 b-u-t but have what l-a-c-k o-f recognise approve sign language also l-a-c-k o-f bilingual education// But, there was a distinct lack of countries where the national signed language had been formally recognised. Also, there was a lack of responses regarding bilingual education. 107 many many many have sign language interpreting services? not really// Not many countries reported having signed language interpreting services. 108 people community know about deaf people? no// We found that the hearing community was not very aware of the situation of Deaf people. 109 means many those deaf people their lives can access t-o services really oppressed// This, in turn, meant that for many of these Deaf people, they were experiencing oppression in terms of accessing services. 110 so those deaf people true enjoy enjoy enjoy equal human rights? no// So, all in all, it seems that Deaf people do not yet enjoy the full and equal human rights. 111 so summarise that say deaf people equal like other round world have have have? no// So, in closing, it is fair to say that Deaf people the world over are not yet equal to their hearing counterparts. 112 still more work t-o lobby lobby lobby// and that more lobbying work on many fronts needs to be carried on. 113 hope really ideal that convention that future can change change// In an ideal world, the UN Convention will change this situation in the future. 114 thank you// Thank you! [00:20:46.700]

Appendix C Pseudonyms and Educational/Professional Details of Participants Pseudonym

Alex Amber Annie Emily Lauren Kay Monica Shannon Tiffany Zoe

Native signers (N = 10)

Pseudonym

University qualification

Interpreter education

Years of interpreting experience

Undergrad None Postgrad None None Postgrad None Postgrad None Postgrad

None A A, C None A None A A, C None C

23 3 15 14 3 25 12 14 22 24

Bernie Claire Debbie Helen Jane Linda Liz Molly Sophia Wendy

Non-native signers (N = 10) University qualification

Interpreter education

Years of interpreting experience

Undergrad Postgrad PhD Postgrad PhD Postgrad Undergrad Postgrad None Postgrad

A None None A A, C A, C A A, C A A, C

20 25 27 21 5 6 9 18 10 18

Appendices

Note: A: Diploma of Interpreting (Auslan/English); B: Advanced Diploma of Interpreting (Auslan/English); C: Postgraduate Diploma in Auslan-English Interpreting or equivalent

247

1. Accuracy/Equivalence and Poor completeness of message (50) (< 50%) < 25

Average (50%–64%) 25–32

Good (65%–74%) 32.5–37

Very Good (75%–84%) 37.5–42

1.1 m  inimal unjustifiable additions – Numerous its number and degree significant errors 1.2 minimal unjustifiable omissions in transfer of 1.3 minimal unjustifiable meaning. Poor substitutions cohesion and 1.4 minimal unjustifiable intrusions coherence. 1.5 cohesion and coherence – links, easy to follow

Many significant errors in transfer of meaning. Adequate cohesion and coherence.

Some significant Some minor errors Few to no errors errors in transfer in transfer of in transfer of of meaning. Good meaning.Very meaning. cohesion and good cohesion Excellent coherence. and coherence. cohesion and coherence.

2. Target text features (20)

Poor (< 50%) < 10

Average (50%–64%) 10–12.8

Good (65%–74%) 13–14.8

2.1 grammatical – syntax, tense, complete sentences 2.2 appropriate vocabulary 2.3 convey signer register (formal) and affect (emotion)

Numerous Many significant Some significant Some minor errors Few to no errors significant errors errors in the errors in the in the production in the in the production production of production of of grammatically production of of grammatically grammatically grammatically correct sentences. grammatically correct sentences. correct sentences. correct sentences. Very good lexical correct sentences. Significant errors Adequate lexical Good lexical choice. A few Excellent lexical in register and choice. Numerous choice. Some minor errors in choice. Accurate affect. errors in register errors in register register and affect. register and and affect. and affect. affect.

Very Good (75%–84%) 15–16.8

Excellent (85%–100%) 42.5–50

Excellent (85%–100%) 17–20

Mark (out of 50)

Mark (out of 20)

248 Appendices

Appendix D A Rubric for Assessing Auslan-to-English Simultaneous Interpreting Performance

3. Delivery features/public speaking Comfortable to listen to (15)

Poor (< 50%) < 7.5

Average (50%–64%) 7.5–9.6

Good (65%–74%) 9.75–11.1

Very Good (75%–84%) 11.25–12.6

Excellent (85%–100%) 12.75–15

3.1 clarity of articulation – Poor clarity of pronunciation, volume, pace articulation, 3.2 fluency – smooth, number of fluency, and hesitations, length of pauses, prosody. frequency of self-repairs, frequency of fillers such as ‘um, ah’ 3.3 prosody – rhythm, illocutionary force, intonation

Average clarity of articulation, fluency, and prosody.

Good clarity of articulation, fluency, and prosody.

Very good clarity of articulation, fluency, and prosody.

Excellent clarity of articulation, fluency, and prosody.

4. Processing skills (15) Strategic interpreting skills

Poor (< 50%) < 7.5

Average (50%–64%) 7.5–9.6

Good (65%–74%) 9.75–11.1

Very Good (75%–84%) 11.25–12.6

Excellent (85%–100%) 12.75–15

4.1 time lag 4.2 a free-literal interpretation continuum with appropriate shifts in between 4.3 strategic additions 4.4 strategic omissions 4.5 successful anticipations

Little or no Adequate evidence Some evidence of Much evidence of Substantial evidence of of strategic strategic strategic evidence of strategic interpreting skills. interpreting skills. interpreting skills. strategic interpreting skills. interpreting skills.

Mark (out of 15)

Mark (out of 15)

Appendices 249

250 Appendices

Note 1 The bold parts in small capital letters are omission potential in the Auslan source text.

Author Index

Adam, R. 1, 9 Allen, C. 61, 232 Anderson, L. 32, 34, 143 Angelelli, C.V. 67 Baddeley, A. D. 24, 30 Baraldi, C. 51, 67 Barik, H. C. 31–32, 34, 36–37, 39, 50, 143, 211 Bartłomiejczyk, M. 47, 108, 211–212 Best, B. 5–6, 37 Bontempo, K. 66, 100 Braun, S. (Sabine) 50–51, 67, 211 Braun, S. (Susanne) 41–44, 46, 124, 145 Burke, T. B. 5, 19 Cartellieri, C. 37 Cattaneo, E. 28 Charles, R. L. 21–22 Chen, S. 30–31 Cheung, A. K. F. 41 Christoffels, I. K. 3, 21, 24, 32, 34, 143 Clarici, A. 41–44, 46, 124, 145 Clark, P. 110 Cokely, D. 1, 5, 26, 30–32, 37, 39–40, 50, 52, 66, 100, 114, 120–121, 138, 143, 147, 167, 207 Collard, C. 31–33, 36, 41, 45–46, 114, 120–121, 143–146 Collins, P. F. 6 Crasborn, O. 7 Dawrant, A. 32–33, 35, 37, 121, 146, 148, 213, 221 Defrancq, B. 31–33, 36, 38–39, 41, 45–46, 114, 120–121, 143–146, 207, 213 De Groot, A. M. B. 32, 34, 143 De Meulder, M. 5, 18, 19, 110, 212, 220, 228 Desmet, B. 41, 44–46, 173

Díaz-­Galaz, S. 32, 35–36, 143 Dickinson, J. 5–6, 19 Diriker, E. 1, 13 Emmorey, K. 6, 13–16, 21, 101–122, 157, 207 Fang, J. 50–51, 66–67, 70, 211 Favaron, R. 51, 67 Feyne, S. 6, 17–18, 52 Finch, K. L. 110 Finton, L. 107 Frittella, F. M. 41 Galy, E. 21–22 Gernsbacher, M. A. 207 Gerver, D. 25, 32, 34, 36, 143 Gevins, A. 21 Gile, D. 1–3, 5, 7, 13, 21, 25–32, 37, 41–42, 46–49, 52, 66–67, 72–73, 80, 92, 108–109, 140, 145, 147–148, 152–153, 160, 163, 207, 210–214, 219, 221–222, 230 Goldman-­Eisler, F. 33, 35, 38, 211 Goswell, D. 6 Hanson, T. A. 25, 30 Harrelson, P. 5–6, 19, 110 Hart, S. G. 22 Haualand, H. 1–2, 4–5, 61, 232 Haug, T. 5, 19–20, 110, 229 Hauser, A. B. 110 Hauser, P. C. 5–6, 19, 109, 110 Hitch, G. J. 24 Hurwitz, T. A. 109 Hutchinson, B. 66 Jacobsen, B. 51, 67 Johnston, T. 1, 4, 62, 64, 74, 85, 103–105, 107, 111, 170, 192

252 Author Index Jones, R. 38, 121 Jorg, U. 212 Kade, O. 37 Kalata-­Zawłocka, A. 5, 101 Kalina, S. 33, 50, 67 Kellett Bidoli, C. J. 13 Kelly, J. 8 Kerzel, D. 7, 24, 108, 207, 208 Kim, H.-R. 32, 35, 37, 143 Kohn, J. 33, 67 Korpal, P. 41, 44, 144 Lamberger-­Felber, H. 33–34, 39, 46, 143, 147, 173, 207 Lawrence, S. 99 Lee, J. 66–68 Lee, T. 31–32, 34, 37–40, 120–121, 140, 143, 208 Leeson, L. 4–5, 13–14, 49, 107 Leneham, M. 37 Lillo-­Martin, D. 4 Liontou, K. 212 Liu, M. 2–3, 5, 21, 24, 26, 28, 31, 66, 210 Lucas, C. 103 Mackintosh, J. 28, 207 Macnamara, B. N. 3 Major, G. 50–51, 67, 211 Malcolm, K. 66 Mazza, C. 28, 41–46, 144–145, 207 McKee, R. 6, 32–33, 37, 40, 114, 143, 207 McNamara, T. 69, 77 Mead, P. 41 Mellinger, C. D. 25, 30 Merlini, R. 51, 67 Mikkelson, H. 52 Miller, G. A. 21 Morgan, E. F. 168 Nanpon, H. 32–33, 37–38, 143 Napier, J. 1, 4–6, 9, 13–16, 18–19, 26, 32–33, 37, 40, 49–52, 60, 67, 84–85, 95, 99, 101–108, 110, 114, 143, 146, 167, 207, 211–212, 219–220, 226, 228 Nicodemus, B. 5–6, 9, 13–16, 19, 21, 70, 101–102, 105, 121, 157, 207 Nilsson, A.-L. 1–2, 4–6, 13, 38, 101, 121 Nixon, J. 21–22 Nygren, T. E. 22 Oléron, P. 32–33, 37–38, 143 Ono, T. 31–32, 35 Owen, A. 8

Paas, F. G. W. C. 22 Padden, C. A. 5 Paneth, E. 32, 143 Patrie, C. J. 102 Pinochi, D. 41–45, 124, 144–145 Pöchhacker, F. 1, 4–5, 13, 31, 66 Podhajská, K. 31–35, 37, 143 Pointurier-­Pournin, S. 29 Reid, G. B. 22 Rosenstock, R. 6 Russell, D. 6, 66 Salvendy, G. 21–22 Sandler, W. 4 Schembri, A. 1, 4, 62, 64, 74, 85, 103–105, 107, 111, 170, 192 Seeber, K. G. 7, 21–24, 28, 35, 108, 207, 208, 210 Setton, R. 32–33, 35, 37, 121, 146, 148, 213, 221 Shaw, R. 66, 85 Sheneman, N. 6 Shlesinger, M. 207 Sloetjes, H. 7 Smith, M. E. 21 Smith, R. 107 Stachowiak-­Szymczak, K. 41, 44, 144 Staveland, L. E. 22 Stone, C. 6 Taylor, M. M. 5, 8, 16–17, 41, 64, 66–67, 100, 102, 105–108, 115, 121, 138, 146, 157, 167–168 Timarová, Š. 1, 3, 21, 30–37, 39–40, 45–46, 108, 114, 120–121, 143–144, 146–148, 221 Treisman, A. M. 31–32, 34, 143 Valli, C. 103 van den Bogaerde, B. 5 van Dijk, R. 16, 99, 101–102, 105 van Merriënboer, J. J. G. 22 Wadensjö, C. 50–51, 67, 70, 211 Wang, J. 1, 3, 15–16, 21, 25–26, 28–31, 33, 37, 41, 50–51, 60, 62, 66–67, 70, 79, 99, 101, 102, 105, 108, 114, 144, 147–148, 180, 207, 211, 221 Whynot, L. 6 Wickens, C. D. 23 Xie, B. 21–22 Zhang, W. 24

Subject Index

Page numbers followed by ‘n’ refer to notes numbers acceptability 227; see also accuracy; completeness of message; equivalence; faithfulness; fidelity; sentence-­level accuracy accuracy 2, 14, 16, 25, 28, 33, 36, 39–40, 42–46, 49, 60, 65–66, 68, 78–80, 124–129, 131, 140, 143–146, 148, 152–160, 165–166, 169, 173, 175, 181–182, 188–190, 192–193, 195–196, 198, 202, 206–209, 211–214, 218–222, 224–225, 227, 248; see also acceptability; completeness of message; equivalence; faithfulness; fidelity; sentence-­level accuracy adapting 13, 20, 29, 48, 52 addition 39, 49, 51–52, 66–67, 70, 140; downtoning 51; elaborating 51; emphasising 51; emphatic 51; explanatory; 51; explicating 51; minor 70, 180; new-­information 51; phatic 51; unjustifiable 66, 139, 248 A language 44, 52; see also native language ambiguity 38, 68, 71, 124, 137, 174, 182–183, 190, 219 American Sign Language (ASL) xv, 2, 8, 14, 16–17, 30, 32, 39, 70, 81–13, 101, 103, 105, 107, 111, 231 analytic assessment 60, 65, 68–69, 77, 80, 110, 168, 225, 227 anticipation 23, 33, 41, 48, 67, 86, 109, 143, 205, 210, 249; see also prediction approximation 42–46, 213 assessment criteria 17, 65–66, 68–70, 78 assessment process xiv, 68–69 assessment rubric xiv, 65, 69, 74, 230, 248–249

assessors 17–18, 39, 78, 198; see also raters attention 7, 21, 26–28, 36, 41–42, 48, 94, 98, 162, 164, 168, 178, 181, 185, 187, 196, 199, 202, 206, 207, 210, 214, 228, 230; see also attentional resources; mental energy attention(al) deficit 27, 146, 196, 220 attention allocation (allocation of attention) 21, 27–28, 148, 210, 230; see also attention management; multitasking attentional resources 7, 26–27, 42, 48, 87, 142, 145, 188, 211; see also attention; mental energy attention lapse 27, 162 attention management 3; see also attention allocation; multitasking audience design 51, 97, 186, 211, 225 Auslan (Australian Sign Language) xiii, xiv, 1 Auslan monologue 6, 61–62, 70–72, 81, 84, 87–88, 97, 105, 168, 226; see also Auslan presentation; Auslan source text; signed monologue; signed presentation Auslan presentation xiv, 60–62, 64, 69, 87, 104, 111, 153, 164, 168, 173, 181, 188, 190, 193, 199, 206, 219, 226, 232; see also Auslan monologue; Auslan source text; signed monologue; signed presentation Auslan source text xv, 7, 61, 64, 69–72, 81, 124, 144, 146, 149, 167, 169, 220, 235–246, 250; see also Auslan monologue; Auslan presentation; signed monologue; signed presentation

254 Subject Index Australian Sign Language Interpreters’ Association (ASLIA) 106 authenticity 225 automaticity 148 automatic processing 22 balanced bilinguals 14–15, 52 best practice 3, 84, 106, 229 bimodality 4, 224 B language 44, 52 briefing 86, 95, 106, 220, 224–226, 228–229 British Sign Language (BSL) 1–3, 81, 105, 231–232 certification 2, 104, 179 chunking 23–24, 47, 207, 210; see also segmentation C language 24, 52 clarification 51, 88, 95, 109, 226 close interpretation 70; see also close rendition; direct equivalence close rendition 70; see also close interpretation; direct equivalence code mixing 85, 91–92, 105 code switching 85, 91–92, 105 cognitive effort 21, 92; see also cognitive load; cognitive workload; mental effort; mental workload; processing effort cognitive load 2, 13, 21–25, 31, 33, 36, 42, 44, 47, 52, 60, 89, 91, 142, 153, 162, 172, 178–182, 186, 199, 207, 209, 211, 213, 231; current 72, 153; exported 72, 153; imported 72, 153, 163; local 24, 28, 214; see also cognitive effort; cognitive workload; mental effort; mental workload; processing effort cognitive overload; xiii, 5, 7–9, 23–24, 26–27, 35, 37, 46–47, 60, 72, 80–81, 88, 90–91, 96–97, 108, 110, 114–115, 139, 143, 152–154, 173, 176–178, 180–183, 188, 190, 192, 199, 202, 210, 212–213, 215, 219–220, 222, 224, 230–232; see also cognitive saturation cognitive process(es/ing) 2–4, 6–8, 21, 25, 28, 31–35, 60, 70–73, 90, 102, 110, 120, 140, 145–147, 153, 155, 160, 165, 178, 183, 188–190, 192, 202, 209, 214–215, 224, 230, 232 cognitive saturation 7, 27, 37, 73, 180, 210, 213; see also cognitive overload cognitive science 4, 7

cognitive workload 21; see also cognitive effort; cognitive load; mental effort; mental workload; processing effort coherence 23, 60, 66, 71, 81, 91, 94–95, 102, 106–107, 110, 141, 176, 180–182, 186, 190, 197, 199, 202, 207, 208, 211, 219–221, 225, 248 cohesion 17, 23, 66, 107, 248 collocation 48, 107, 168, 212 communicative effect 51, 91, 97, 169, 197–198, 211 communicative goal 19, 25, 52, 67, 71, 110 community interpreting 60, 64 competence 6, 17, 147 completeness of message 66, 248; see also acceptability; accuracy; equivalence; faithfulness; fidelity; sentence-­level accuracy comprehension 16–17, 27, 29–30, 33, 67, 72, 81, 85, 88, 104, 153, 181, 213, 228; difficulties 85, 88; problems 8, 80–81, 96, 99–100, 102, 105, 109–110, 128, 219; signed language 8, 19, 29–30, 80–81, 87, 96, 99–100, 102, 104–105, 142, 153, 219, 224, 231; source language 48, 95, 100, 102, 138, 162, 164, 187, 196, 207, 209, 214; source message 94; source speech 28, 30, 41, 48, 96, 146 compression 96, 98; see also condensing concurrent processing 3, 153, 162, 199, 210, 212 condensing 98–99 conduit 230 conference interpreter(s) 25, 32, 34, 38, 52, 108, 230–231 conference interpreting xiii, 5, 8, 60, 63–64, 88, 110, 153, 226 conscious awareness 40, 108, 225 conscious intentional omissions 50, 84, 97 conscious receptive omissions 50 conscious strategic omissions 50, 67, 98; see also strategic omissions conscious unintentional omissions 50 consecutive interpreting 13, 30, 34 constraints: cognitive 33; environmental 61; linguistic 104; time 3 context 17–18, 23–25, 34, 37–38, 41–42, 47–48, 50, 52, 66–67, 71, 82–83, 87, 93, 98, 100, 106–107, 124, 134, 149, 162, 164, 167–168, 173, 198, 201, 204, 208, 212, 229, 232

Subject Index cooperation strategies 5, 18, 21, 95, 106, 110, 153, 183, 226, 228; see also look-­pause-­nod coordination (of attention) 26, 29–30, 48, 72, 146, 148, 153, 214 coping strategies/tactics 9, 46, 48–49, 61, 72, 74, 77, 102, 108, 110, 138, 149, 152, 169, 176–177, 181–182, 188–189, 192, 194–195, 213, 215, 218, 231; see also interpreting strategies; processing skills; strategies; tactics corpus 16, 60, 62, 65, 70, 121, 145; experimental 4, 7, 60, 62, 65, 69, 153, 206; interpretation 8–9, 61, 69, 74, 192, 214, 226 corpus-­based approach/research/study 35–36, 38, 45–46, 121 correspondence (source-­target) 39 creativity xiv, 170, 189, 195, 202, 207, 225 culturally bound concept (culture-­bound information) 18, 51 current load 207 data xiii, 9, 22, 24, 63, 74, 87, 99, 131, 227, 231, 235, 237–239; analysis 24, 62, 149, 152, 227; capture/collection 64, 74, 169; interpretation xiv, 9, 50, 60, 65, 68–69, 72, 110, 115, 121, 145, 147–148, 152–153, 206, 215, 220–222, 225, 227, 231–222; interview 60–61, 74, 80, 101–102, 104, 106, 110, 218–220, 222, 226 data-­driven study 45 Deaf community 4, 9, 19, 61, 63–64, 97, 99, 102–103, 106–107, 197, 220, 239 Deaf culture 61, 106 Deaf education 61, 86, 89, 97, 117, 243–234 Deaf professional(s) 1, 4–8, 17–21, 61, 64, 80–98, 102, 104–107, 109–111, 153–154, 159, 167–168, 176, 182, 188, 190, 192, 218, 225, 227, 229, 231–232 Deaf (signed language) interpreter(s) 1, 2, 6, 106, 232 debriefing 18 décalage 1; see also ear-­voice span; eye-­voice span; lag time; processing time; time lag decision-­making 47, 70, 108, 182–183, 231–232 deeper level processing 49; see also free interpretation; meaning-­based processing

255

deliberate practice 108 delivery (interpreter’s or speaker’s) 3, 17–18, 36, 41, 65, 67–68, 91, 93, 221 delivery features 15–16, 36, 65, 67–68, 78, 80, 99, 219, 249 delivery rate 34, 36, 44–45, 142, 148; see also presentation rate dense information 2, 27, 41, 71, 88, 108, 152, 165, 169–170, 173, 176, 184, 189, 199, 207, 208, 210–211, 218–220, 222–223, 228; see also information density designated interpreter 6, 19, 21, 109 dialect 82, 85, 104, 106 direct equivalence 70; see also close interpretation; close rendition directionality 4, 13–16, 35, 43, 52, 63, 94, 100, 232 directionality effect 14, 16, 232 directionality preference 14–15, 63–64 discourse environment 18, 49–50, 219–220, 224, 232 distortion 25, 46, 66, 70–71, 73, 125, 138, 147, 154, 162, 221; see also substitution diversity 62, 104, 170, 188–189, 195, 202, 207, 224–225 ear-­voice span 34, 36, 39, 45–46; see also décalage; eye-­voice span; lag time; processing time; time lag educational background 50, 101, 106, 169, 227 effectiveness 51, 67, 182, 211, 219, 225 Effort Model(s) 13, 25–30, 52, 80, 92, 152–153, 160, 207, 215n1, 222, 230 ELAN 7, 25, 60, 62, 69–71, 73, 115, 120–121, 226, 230 end negation 9, 115–116, 118–119, 138, 140–141, 147, 149, 192–215, 218, 221–224 enumeration 27 equivalence 37, 49, 66, 248; direct 70; dynamic 49, 198; formal 49, 52n2; see also acceptability; accuracy; completeness of message; faithfulness; fidelity; sentence-­level accuracy error analysis 42, 47, 49; see also error taxonomy; miscue taxonomy error taxonomy 42, 124, 145; see also error analysis; miscue taxonomy evaluation (of interpreting performance) 68, 71; see also analytic assessment; quality assessment

256 Subject Index expanded renditions/expansion 51, 99, 185, 210; see also strategic additions experimental study 15, 24, 44, 96, 226 expertise xiv, 2, 6, 18–19, 61, 65, 148, 229 explicitation 51 exported load 207 external factor(s) 31–32, 34, 37, 40, 142, 144, 212 extralinguistic knowledge 30, 212 eye contact 3, 17–20, 29, 190, 212, 220, 226, 228–230 eye-­voice span 1; see also décalage; ear-­voice span; lag time; processing time; time lag facial expression 13, 20, 89, 206, 230, 234 failure sequence(s) xv, 214, 230 faithfulness 66; see also acceptability; accuracy; completeness of message; equivalence; fidelity; sentence-­level accuracy false start(s) 18, 36, 66, 135, 138–139 familiarity 19, 26, 33–34, 37, 39, 50, 67, 86, 88, 97, 109–110, 144, 148 fatigue 33, 180–182, 196 feedback xv, 20, 29, 230 fidelity 66; see also acceptability; accuracy; completeness of message; equivalence; faithfulness; sentence-­level accuracy filler(s) 17, 23, 51, 107, 162, 164, 166, 178, 197, 249 fingerspelling 15–16, 20, 74, 82, 92, 102, 104, 107, 141, 157, 164, 179, 185, 228 fluency 36, 67, 70, 99, 102, 214, 220, 225, 249 focus of attention 214 formal register 17, 60, 66, 81, 84–85, 92, 94, 96, 100, 102, 105, 107–108, 110, 167, 180–181, 186, 207, 219–120, 225–226, 231 formal (interpreting/interpreter) training 2, 19, 64, 70, 101, 108, 148, 169, 179, 185; see also interpreter education form-­based (information) processing/ interpreting/translation 48–49, 91; see also literal interpretation; transcodage; transcoding; tranlisteration free interpretation 20, 49, 67, 70, 73, 92, 96, 98, 108, 160–161, 163–165, 189, 195, 198–203, 207, 209–211, 222–223, 225; see also deeper level processing; meaning-­based processing

generalisation/generalising 48, 90, 91, 96, 98, 108, 124–129, 145, 154, 213, 222 gesture(s) 18, 82, 85, 190, 226, 229 hesitation 18, 116, 166, 213, 241, 249 illocutionary force 51, 71, 249 imported load 207 inference 164, 208 information density 81, 144, 154, 207; see also dense information informal register 167–168, 207 instantaneous workload 22 interaction 19, 20, 29, 50–51 internal factor(s) 32, 34–35, 37, 40, 115, 143–144 International Sign 6, 40, 81–84, 92, 105, 111, 187, 225, 236 interpreter education 9, 65, 103–104, 227, 247; see also formal (interpreting/ interpreter) training interpretese 213 interpreting from a signed language into a spoken language 4, 6–8, 16, 18, 29, 41, 52, 105, 115, 152, 220, 222, 224, 232 interpreting from a spoken language into a signed language 4, 15, 29, 232 interpreting skills 16–17, 19, 27, 100, 104, 108–109, 209, 249 interpreting strategies 25, 32–37, 40–41, 47, 65, 96, 99, 108, 110, 114, 140–142, 144, 148, 225, 230–231; see also coping strategies; strategies; tactics inter-­rater reliability 68–69, 77–79 interrelation 32, 140, 153, 160, 189, 218, 232 intonation 25, 66–67, 137, 164, 168, 197, 213–214, 231, 234, 249 key message 73, 202 knowledge: cultural 50, 67; domain-­ specific 86, 101; encyclopedic 67, 86; extralinguistic 30, 212; general 67; linguistic 30, 194, 224; requisite 19; specialised 18, 108; world 101, 108 knowledge-­lean skills 17 knowledge-­r ich skills 17 lag time 1, 89, 97; see also décalage; ear-­voice span; eye-­voice span; processing time; time lag language acquisition xiii, 103, 219

Subject Index language combination xiii, 8, 33, 35, 38–41, 46, 212, 230–231; see also language pair language pair 3, 8, 32–34, 37–40, 42–43; see also language combination lexical choice 66, 70, 81, 164–165, 167–169, 186, 197, 231, 248; see also lexical selection lexical selection 69; see also lexical choice linearity 173 lip reading 20, 244 literal interpretation 15, 17, 38, 41, 49, 52n2, 67, 73, 92, 96, 98, 108, 136, 147, 160, 163–166, 168–169, 179, 182–184, 188–190, 198–203, 207, 209–211, 222–223, 225, 249; see also form-­based (information) processing/interpreting/ translation; transcodage; transcoding; transliteration local analysis 72–73, 153–154, 160, 164, 169–171, 176, 182, 189–190, 192, 194–195, 202, 205, 206, 210, 212–215, 218, 221–224, 230–231 long-­term memory 21, 27, 30, 96, 99, 222 look-­pause-­nod 18, 106, 110, 153, 183, 220, 225–226, 228; see also cooperation strategies meaning 1–3, 13, 17, 19, 21, 23, 28–29, 40–42, 49, 51, 52n1, 66, 70–71, 82, 86, 90–92, 97, 107, 124, 137–140, 142–143, 147, 149, 154–155, 158, 163, 166, 169–170, 175–176, 179–181, 183, 185–187, 189, 192–193, 195–199, 202–205, 207, 221–222, 228–229, 237, 239, 245, 248 meaning-­based (information) processing 49, 91, 101, 203; see also deeper level processing; free interpretation meaning transfer 4, 67 meaning unit 66, 71, 90, 136, 138, 141, 163, 165, 173–175, 178–181, 183, 186–187, 194, 196, 199, 202–205, 223 memory overload 90, 94; see also short-­term memory overload; working memory overload mental effort 21–22; see also cognitive effort; cognitive load; cognitive workload; mental workload; processing effort mental energy 7; see also attention; attentional resources

257

mental workload 21–22; see also cognitive effort; cognitive load; cognitive workload; mental effort; processing effort message(s) 1–2, 21, 26, 29, 31, 35, 49, 51, 66–67, 71, 73, 97, 102, 121, 136, 140, 145, 154, 156, 162, 176, 183, 185–186, 188, 196–197, 199, 202–204, 207, 211, 219, 223, 227, 232, 248; Auslan 93, 135, 179, 183, 187, 202, 218; signed 4–5, 20, 29, 98, 100, 102, 109, 135, 199, 224, 226–228, 232; source (language) 1–3, 23, 48–50, 52, 66, 70, 94, 98, 122, 136, 146–147, 182, 189–190, 207, 209–210, 219, 221, 224, 229; target language 49, 52n2 message transfer 62 microanalysis 60, 69–70, 72, 74 miscomprehension 146–147, 182, 190, 209, 224 miscue 39–40, 50, 52n2, 66, 70, 178, 182 miscue taxonomy 49; see also error analysis; error taxonomy modality (of interpreting) 4, 13, 15 mode (of interpreting) 1, 13, 29, 44, 52, 229–230 monitoring 21, 31, 69, 162, 168, 201; see also self-­monitoring multitasking 128, 143, 153, 174, 182, 199, 211–214, 219; see also attention allocation; attention management name sign/sign name 16, 107, 235, 238 National Accreditation Authority for Translators and Interpreters (NAATI) 62–63, 65, 69–70, 74 National Aeronautics and Space Administration Task Load Index (NASA-­TLX) 22 native language 2, 14, 24, 43, 52n1, 93–94, 100, 102, 105–108, 110; see also A language native signer 15–6, 19, 62–65, 74, 79–80, 99–102, 104–105, 107, 110, 125, 219, 247 New Zealand Sign Language (NZSL) 1, 105 non-­automatic operations 26 non-­manual features 104, 107, 206, 212, 230 non-­native language 2, 14–15, 43, 52n1, 63, 93, 100, 103, 105; see also B language

258 Subject Index non-­native signer 15–16, 62–64, 79–80, 93, 99–102, 105, 107, 110, 125, 219, 247 non-­renditions 51 non-­verbal information 51 norm(s) 17–18, 33, 92 note-­taking 30, 34, 43, 46 number processing time 114, 120, 131–136, 141–142, 144, 146, 165, 167, 169, 180–181, 221; see also number time lag; processing time for/regarding numbers numbers 9, 13, 16, 27–28, 31, 35, 39, 41–48, 52, 71, 73, 81, 87–91, 95–96, 98, 108, 114–115, 120, 124–135, 142–149, 152–162, 164–167, 169–174, 176–177, 181–182, 184, 187–189, 192–193, 205, 207–210, 212, 214, 218–223, 228–229, 245 number time lag 120–121; see also number processing time; processing time for/ regarding numbers NVivo 74, 80 observation 74, 146 omission 39, 42–46, 49–50, 52, 66, 71–73, 124–129, 136, 142, 145, 167, 182, 190, 196, 199, 213; long 39; major 208; minor 71, 196, 199, 202, 213; significant 182; unjustifiable 25, 39–40, 46, 66, 70, 138–139, 147, 154, 182–183, 187–188, 190, 196, 198, 199, 210, 212–213, 221, 224, 248 omission potential (in the source text) 250 onset processing time 31, 39, 45–46, 114–115, 120–123, 131–144, 146–148, 154, 160–161, 163–170, 172–176, 178–183, 185–187, 189–190, 192, 194–202, 204–205, 207–210, 212, 221–223, 228; see also onset time lag onset time lag 31, 114, 120–121; see also onset processing time padding 23, 140, 166, 170, 172, 197, 223 paraphrasing 41, 48, 96–97, 124–129, 222; see also rephrasing pause (pausing) 3, 17–18, 36, 66, 70, 73, 107, 137, 139–140, 148, 160, 167–168, 170, 172, 175, 180–182, 185–187, 195–199, 202–205, 207–211, 213, 223, 229, 249; filled 36, 179–180, 213; silent 36 peak workload 22

performance quality (interpreter’s) 6–8, 25, 60, 80, 102, 160, 224; see also simultaneous interpreting performance personality 37, 109, 231 predictability 33, 41, 143 prediction 67, 122, 166, 205–208, 212; see also anticipation preferred interpreter 5, 18–19, 109 preparation 3, 6, 18–19, 33–35, 49, 62, 64, 67, 86–87, 98, 101, 104, 106, 110, 115, 148, 195, 205, 206, 212, 220, 224, 228, 231 presentation rate (source speech) 34, 37; see also delivery rate; problem-­solving47 problem trigger 9, 13, 27–28, 41, 46, 52, 71, 108, 207, 211, 214, 230 processing capacity 7, 26–28, 40, 48, 92, 142, 162, 168, 181, 188, 195, 204, 207, 209, 211, 214; available 7, 26–27, 212, 214, 224; required 7 processing effort 34, 48; see also cognitive effort; cognitive load; cognitive workload; mental effort; mental workload processing load 22 processing skills 65, 67–68, 78–80, 219, 249; see also coping strategies/tactics; interpreting strategies; strategies; tactics processing speed 2, 31, 73, 94, 221 processing time 1–2, 7, 9, 13, 24–25, 27–28, 30–41, 45–47, 52, 60, 67, 70, 72–73, 81, 87–91, 95, 97–99, 108, 114–116, 120–121, 123–124, 128, 130–136, 138, 140–149, 152, 160, 162–169, 172–173, 176, 178–182, 185–190, 194, 196–197, 199, 201, 203–205, 208, 209, 212, 218–222, 224, 230–232; see also décalage; ear-­voice span; eye-­voice span; lag time; time lag processing time for/regarding numbers 31, 45, 114–115, 120, 124, 145–146, 188; see also number processing time; number time lag professional development 106 professionalism 17–18 proper names 41 proposition 71 prosody 14, 67, 93, 99, 102, 106, 249 pseudo-­cleft structure/sentence 154, 170–171, 176, 188–189, 192, 194, 198, 201, 209, 223, 229; see also question-­ and-­answer (syntactic) structure; rhetorical question

Subject Index psychophysiological measures 22–23 pupillometry 23, 231 qualitative analysis/analyses 36, 114–115, 120–121, 136, 138, 140, 147–149, 152, 189, 207, 215 qualitative results 99 quality assessment 3, 9, 16, 66 quantitative analysis/analyses 114–115, 120–121, 128, 142, 147–148, 152, 205, 207 quantitative results 9, 147, 188, 192, 214 question-­and-­answer (syntactic) structure 137–138, 170, 176, 178, 182, 188–189, 192, 201, 209, 210, 223; see also pseudo-­cleft structure/sentence; rhetorical question raters xiv, 60, 62, 65, 68–69, 77–79, 110–111, 164, 218–219, 225, 227; see also assessors redundancy 34, 41, 85, 141–142, 163–164, 167, 170, 179 repertoire (linguistic) 105–106 repetition 18, 51, 70, 109, 141, 179–180, 182, 185, 204, 207, 213, 226 rephrasing 20, 51, 94, 96, 98–99, 222; see also paraphrasing representation 17, 95, 98 restructuring 24, 47, 195–196 rhetorical question 84–85, 89, 92, 97, 104, 154, 166–168, 170, 176, 178, 181–182, 184, 189, 192, 194, 197, 199, 201–203, 205, 223, 228; see also pseudo-­cleft structure/sentence; question-­and-­ answer (syntactic) structure role-­shift 204, 229 segmentation 23, 31, 37, 47, 70; see also chunking self-­correction 15, 37, 49, 70, 136, 139, 147, 160–162, 164, 167, 176, 179, 181–182, 188, 190, 195–196, 203, 207, 209–210, 213, 224 self-­monitoring 29, 93, 95, 101, 109, 164, 174, 207, 208 sentence-­level accuracy xv, 71, 136, 152–154, 192, 227; see also acceptability; accuracy; completeness of message; equivalence; faithfulness; fidelity setting 3, 6, 17, 19, 67, 88, 95, 100, 108–109; conference xiv, 4, 6–7, 45, 64, 66, 80, 84–85, 91–92, 107–108, 110,

259

167–168, 180, 186, 190, 207, 218–219, 224; formal xv, 2–3, 6–8, 18, 52, 60–61, 65, 77, 101–102, 105–106, 110, 195, 207, 218, 224, 227, 229, 231, 240 shadowing 34 short-­term memory 26–30, 43, 47–48, 161, 214 short-­term memory overload 28 sign(ed) language interpreters xiii, 2–8, 13–15, 19–20, 25, 38, 46, 48, 51–52, 61–62, 65, 74, 82, 87, 100–110, 115, 118, 138, 144, 146, 152, 154, 157, 159, 161, 178–183, 188, 195, 197, 211–212, 214, 218–220, 222, 224–225, 227–232, 239–240, 245–246 sign(ed) language interpreting xiii–xv, 1, 4–9, 13–16, 19, 29, 39, 41, 49–50, 61, 65–66, 95, 100–101, 104, 106, 109, 118–119, 121, 154–155, 159, 167, 180, 182–183, 185–186, 193–194, 197, 219, 227, 230–232, 245–246 signed (language) output 15, 29 signed monologue 3–4, 6, 8, 18, 25, 73, 81, 85, 92, 101–102; see also Auslan monologue; Auslan presentation; Auslan source text; signed presentation signed presentation 3, 18, 20–21, 61, 88, 96, 105–106, 109–110, 115, 121, 136, 169, 202, 212, 220, 224–226, 228–232; see also Auslan monologue; Auslan presentation; Auslan source text; signed monologue signing style 20, 64, 81–85, 88, 95, 100, 103–106, 110, 219, 229 sign variation 19, 81–83, 102, 106, 219–220, 224, 228 simultaneous communication 103 simultaneous interpretation xiv, xv, 2–4, 6–7, 16, 25–27, 36, 38–40, 42, 45–46, 50, 60–61, 64, 65, 69, 72, 74, 79, 136, 145, 147–148, 152–153, 164, 167, 169, 182, 189, 192, 206, 209–210, 212–215, 218–222, 224–229, 231 simultaneous interpreting performance xiv, 7–9, 14–16, 25–27, 31–32, 37–39, 65, 68–69, 77–80, 99–101, 110, 136, 148, 182, 210–211, 213–215, 218–220, 225, 227, 230, 248; see also performance quality source language input 21, 23, 30–31, 41, 100 source language interference/intrusion 38–41, 66, 70, 136–138, 147, 167, 169, 207, 221

260 Subject Index spillover effect 24 spoken language interpreters 4, 8, 14, 34, 48, 64, 114, 144, 146, 211, 214, 230 spoken language interpreting 1, 4, 8, 13, 121 stalling 23–24, 96–97, 166, 170, 172, 178, 181, 190, 197, 207, 209–210, 222–223 strategic addition 50–52, 67, 70, 73, 96–99, 108, 164, 166, 173–174, 181–183, 185–188, 190, 197, 199, 202, 205, 207, 209–213, 219, 222, 224–225, 249 strategic omission 47, 49–50, 52, 67, 70, 73, 96, 98–99, 108, 142, 163–164, 167, 173, 182, 211, 219, 222, 224–245, 249; see also conscious strategic omission; strategic reduction strategic reduction 50; see also conscious strategic omission; strategic omission strategies xiii, 3, 5–9, 13, 17, 20, 23, 37, 40–41, 47, 52, 60, 67, 73, 77, 90–91, 93–94, 96, 99, 102, 108–110, 115, 124, 136, 144, 147, 152, 160, 169–171, 175–176, 178, 181–182, 188–190, 192, 194–195, 198, 199, 202, 205, 207, 209–214, 220–225, 229–232; see also coping strategies/tactics; interpreting strategies; processing skills; tactics stress (occupational) 22, 95 stress (prosodic) 199 Subjective Workload Assessment Technique (SWAT) 22 subject matter 18, 27, 30, 33, 50, 67, 81, 85–86, 88, 100, 109–110, 219, 225 substitution 39, 45–46, 66, 107, 124–129, 139, 145, 248; see also distortion summarising/summarisation 73, 96, 98–99, 108, 196, 213, 222 suppression 160–161, 164, 208, 224 syntactical asymmetry 24, 28, 37, 89–90, 108, 147, 210; see also syntactic(al) difference syntactic(al) difference 9, 27, 32, 40, 71, 81, 88–89, 91–92, 101, 106, 114–115, 135, 137, 142, 144, 148, 152, 169–170, 173, 177, 181, 189, 192, 194, 199, 201, 207–212, 214–5, 218–20, 222–4, 228, 230; see also syntactical asymmetry tactics 7–8, 37, 47–48; comprehension 47; emergency 48; preventive 47–48; reformulation 47–48; see also coping strategies/tactics; interpreting strategies; processing skills; strategies

tail-­to-­tail span 31, 72–73, 120–121, 138–140, 172–175, 181, 185–187, 207, 212 target audience 13, 17, 51, 197 target language output 21, 30–31, 147, 162, 164, 168–169, 187; see also target language production target language production 23, 47–48, 92, 94, 100, 102, 138, 140, 145, 153, 162–164, 181, 187, 196, 199, 204, 205, 214; see also target language output target text features 15, 65–66, 68, 78, 80, 99, 219, 248 team interpreter 3, 46–47, 81, 87, 90–91, 93, 96, 128, 153, 181, 190, 207, 213, 219–220, 225–226, 229 teamwork 110, 225, 231 technologies 30, 104 temporal delay 31, 120, 143 temporal range 32–33 terminology 17–18, 35, 37, 39, 50, 84–85, 92, 96, 99, 106, 144, 220, 224 thematic analysis 60, 74, 80, 110, 222 think-­aloud protocol (TAP) 22 tightrope hypothesis 25, 27–28 time lag 1, 88, 90, 95, 97–98, 164, 194, 249; see also décalage; ear-­voice span; eye-­voice span; lag time; processing time topic-­comment structure/statement 104, 154, 168–171, 189, 209 transcodage 49; see also form-­based (information) processing/interpreting/ translation; literal interpretation; transcoding; transliteration transcoding 15, 48–49, 102, 169; see also form-­based (information) processing/ interpreting/translation; literal interpretation; transcodage; tranlisteration transcript 19, 62, 73, 104, 160 transliteration 49; see also form-­based (information) processing/interpreting/ translation; literal interpretation; transcodage; transcoding trust 19–20, 88, 95, 106, 110, 225 unbalanced bilinguals 15, 63, 105 unconscious omissions 50 unfamiliarity 50, 81, 85–88, 100, 109–110, 219 United Conventions on the Rights of Persons with Disabilities (UNCRPD) 2, 81–82, 91–92, 111, 162

Subject Index variability 22, 32–33, 64, 68, 77, 79, 114, 122, 143–144, 227 visual access 44, 128, 131, 143, 173, 190, 229 visual cues 18–19, 225 visual-­spatial language 4, 13 waiting 23–24, 37, 47, 96–97, 121, 147, 166, 190, 195–196, 201, 207, 209–210, 213 working conditions 48, 95, 227

261

working memory 21, 24–25, 28–30, 33–34, 42–44, 72, 142, 145–148, 153, 204, 214 working memory capacity 24–25, 33, 37, 40–41, 60, 136, 140, 144, 148, 180, 182, 188, 202, 209, 212, 221 working memory overload 48, 128, 146, 176, 180, 212; see also memory overload; short-­term memory overload working memory span task 60, 180 zero renditions 50