探討社群媒體使用者對於口譯品質之觀感

國立臺灣大學文學院翻譯碩士學位學程 碩士論文

Graduate Program in Translation and Interpretation College of Liberal Arts

National Taiwan University Master Thesis

探討社群媒體使用者對於口譯品質之觀感 Exploring Social Media Users’

Perception of Interpreting Quality

林依瑩

Serena Yi‐Ying Lin

指導教授：張嘉倩 博士 Advisor: Chia‐Chien Chang, Ph.D.

中華民國 104 年 2 月

February 2015

i

Acknowledgments

I am very fortunate to have had the help of many people that made possible the completion of this thesis. First, this document wouldn’t exist if it were not for one person: Prof. Chia-Chien Chang. As a teacher, mentor, and friend, Prof. Chang has gone beyond the call of duty throughout my entire period at GPTI, and has patiently offered me both academic supervision and personal support through quite a long and challenging process. For many times I thought about postponing or giving up, and it was Prof. Chang who helped set me on the proper path. I owe my deepest gratitude to you.

I would also like to express my sincere appreciation to my committee members, Prof. Damien Fan and Prof. Elma Ju. Thank you for your valuable critiques and suggestions that helped shape the direction of this thesis. I am also very grateful to Prof. Michelle Wu for encouraging me to explore this research topic.

My sincere thanks also go to Vicky Li, who helped me navigate through the degree requirements, and my GPTI classmates, Ann Hsu, Veronika Song, Tony Tsou, and Iris Wang, for sharing your

enthusiasm. I also have to thank all the anonymous YouTube viewers whose comments made possible this research.

Finally, I would like to thank my family. The birth of my two daughters, Pei and Shin, marked the beginning and end of this degree, and I will miss carrying them to class (both in belly and carrier). My grandfather, who often talked about his granddaughter proudly, passed away two weeks before the professional exam. My parents, who have always given me unconditional love and support, stood even firmer with me during struggling times. My husband, Mr. Wind, never stopped believing in me. My gratitude to them is beyond words.

The completion of this thesis and degree turned out to become not only a pursuit of knowledge but a journey of personal introspection, and I feel very fortunate to have accomplished this journey with the help of so many people.

ii

Abstract

This paper examines social media users’ perceptions of simultaneous interpreting

quality in a live streaming interpreting event on YouTube. On December 11, 2012,

Harvard Professor Michael Sandel was invited by Taiwan’s Ministry of Culture to give

a lecture about his new book, What Money Can’t Buy: The Moral Limits of Markets.

The lecture was held at a stadium with an audience of 6,000 people. At the same time,

the lecture was broadcasted live on YouTube, so simultaneously there was a group of

online users listening to the lecture. Because the YouTube streaming did not provide

dual channels, the online audience had no choice but to listen to the simultaneous

interpretation, while the original source speech was broadcasted at the backdrop at a

lower volume. In other words, online YouTube audience became default absent users of

the interpreting service. Most importantly, the YouTube audience not only watched

Sandel’s lecture and listened to the interpreters online, but many of them posted online

comments simultaneously. The comments comprised the basis of this research.

A total of 233 comments from 134 unique users were collected to form a rich set of

uncontrolled, natural data on users’ perception of interpreting quality. This data was

closely examined to understand the different quality criteria used by different users.

iii

Overall findings show that delivery-related criteria, as opposed to content-related ones,

were most often cited by the interpreting users. Context does matter, so given the

highly-interactive nature of the interpreting event, the data shows that users paid much

attention to prosodic features, such as fluency of delivery, smooth alternation between

source and target languages, and lively intonation. Technology is another important

element in this research. New electronic media has enabled interpreting events to

broadcast live on-line, expanding the number of interpreting users and also changing the

nature of speaker-listener-interpreter relationship. Adequate technical support is

necessary in this case to ensure the quality of interpretation delivery, otherwise technical

shortfall affects users’ quality perception, as reflected in a large portion of comments

that preferred listening to the lecture without the interpretation.

Keywords: quality of interpretation, quality perception, Internet interpreting, social

media.

iv

中文摘要

本篇論文旨在探討一群YouTube 社群媒體使用者對於同步口譯品質的觀感。

2012 年 12 月 11 日美國哈佛大學邁可·桑德爾教授(Michael Sandel) 應文化部邀請 來台演講並談論他的著作－“錢買不到的東西：金錢與正義的攻防＂。該演講在

台灣大學體育館舉辦，現場觀眾共6,000 人，備有同步口譯服務。該演講同時也有

YouTube 網路即時轉播，所以有為數可觀的線上觀眾。當時 YouTube 並沒有中英 雙聲道服務，所以線上觀眾只能聆聽現場同步口譯版本，而演講原音則以較低音

量在背景播放，因此該YouTube 聽眾成了口譯服務遠端使用者。值得一提的是這

群口譯使用者不但透過YouTube 聆聽 Sandel 演講及口譯內容，更有許多聽眾在

YouTube 網頁上留下大量口譯相關意見，這些寶貴意見遂形成本研究的分析基礎。

總共有134 位 Youtube 使用者提供了 233 個意見，形成豐富、未經研究控制的

自然資料，有助瞭解使用者的口譯品質觀感。透過嚴謹資料分析，本研究嘗試了 解使用者不同的口譯品質衡量標準。研究結果顯示大部份口譯聽眾注意到表達相 關的衡量標準，而非過去研究中較受重視的實際口譯內容。口譯情境(Context)也很 關鍵。由於該演講場合高度互動的特質，資料顯示許多聽眾重視口譯訊息的表達，

包括聲韻、流暢度、生動語調以及口譯員目標語和來源語間的轉換。科技與口譯 也是本研究的重要議題，特別是科技如何影響聽眾的口譯品質觀感。本研究彰顯 新媒體以及新型傳播科技如何為更多口譯活動提供實況轉播，進而擴大口譯使用 者群也改變了講者、觀眾和口譯員三者間關係。為確保口譯轉播品質，需有適切 的技術支援，否則就如同許多使用者表達他們希望只聽原音而不要聽口譯，技術 上的不足有可能也會影響使用者的口譯品質觀感。

關鍵字：口譯品質、使用者品質觀感、網路口譯、社群媒體

v

Table of Contents

Chapter 1 Introduction ... 1

1.1 Background of the Study ... 2

1.2 Statement of the Problem ... 7

1.3 Purpose of the Study ... 8

1.4 Significance of the Study ... 9

1.5 Primary Research Questions... 12

1.5 Assumptions, Limitations, and Scope ... 13

1.6 Thesis Organization ... 14

Chapter 2 Literature Review ... 16

2.1 Defining Interpreting Quality ... 16

2.2 Measuring Interpreting Quality ... 18

2.2.1 Different Quality Criteria ... 18

2.2.2 Perceived Importance of Different Quality Criteria ... 20

2.2.3 Challenges in Measuring Quality ... 22

2.3 Interpreting Quality in the Ears of Different Users ... 24

2.3.1 Importance of Users’ Perspectives ... 24

2.3.2 Different Perspectives between Professional Interpreters and General Users ... 26

2.4 Broader Factors Affecting Interpreting Quality... 32

2.5 Media Interpreting ... 35

2.5.1 Definitions and Context ... 35

2.5.2 Users in Media Interpreting ... 39

2.5.3 Quality Studies in Media Interpreting ... 42

2.5.4 Gap in Existing Literature ... 45

Chapter 3 Methodology ... 46

Chapter 4 Findings and Discussions ... 53

4.1 Establishing Interpreting Context: Findings from Transcription Data ... 53

4.2 Understanding Quality Criteria-Findings from Comment Data ... 61

4.3 Technology and Interpreting Quality ... 64

4.4 Delivery-related Quality Criteria ... 73

4.4.1 Synchronicity ... 77

vi

4.4.2 Fluency of Delivery ... 83

4.4.3 Lively Intonation ... 88

4.4.4 Pleasant Voice ... 91

4.5 Content-related Quality Criteria ... 91

4.5.1 Logical Cohesion ... 92

4.5.2 Correct Terminology ... 95

4.5.3 Appropriate Style ... 96

4.6 Broader Factors Affecting Quality Perception ... 97

4.6.1 Quality Comparison: Comparing Interpreting Quality of Male and Female Interpreters .... 97

4.6.2 Quality Judgment: General Comments ... 110

4.6.3 Users’ General Perception and Understanding of the Role of Interpreters ... 115

4.7 English as Lingua Franca ... 124

4.7.1 Users’ Language Preference ... 125

4.7.2 Users’ Preference for Original English Lecture ... 127

Chapter 5 Conclusions ... 132

5.1 Summary of Study ... 132

5.2 Limitations of the Current Study ... 143

5.3 Directions for Future Research ... 145

5.4 Conclusion ... 148

References ... 150

Appendix ... 157

Appendix A: Lecture Transcription ... 157

Appendix B: List of 24 Audience Members ... 212

Appendix C: Interpretation-related YouTube Comments ... 213

vii

List of Tables

Table 1: Example of transcription segments ... 50

Table 2: A quantitative analysis of lecture transcription data ... 55

Table 3: Main lecture segments ... 58

Table 4: Dialogue #8 ... 59

Table 5: Interpretation-related comments by type and number ... 64

Table 6: Comments about technical arrangement and its effect on listening quality ... 68

Table 7: A summary of quality-criteria related comments ... 74

Table 8: Quality criteria- synchronicity ... 80

Table 9: Frequent turn example 1 ... 82

Table 10: Frequent turn example 2 ... 83

Table 11: Quality criteria- fluency of delivery ... 85

Table 12: Awkward word order ... 85

Table 13: Sentence repair ... 86

Table 14: Sentence segmentation ... 87

Table 15: Quality criteria- lively intonation ... 88

Table 16: Contrasting turn from male to female interpreters ... 89

Table 17: Lively intonation ... 90

Table 18: Quality criteria- pleasant voice ... 91

Table 19: Quality criteria- logical coherence of utterance ... 92

Table 20: Factors that impair logical cohesion ... 94

Table 21: Quality criteria- correct terminology ... 96

Table 22: Quality criteria- appropriate style ... 97

Table 23: Specific comments about male and female interpreters ... 101

Table 24: A quantitative comparison of female and male interpreter’s rendition ... 105

Table 25: Selected speech segments for speed comparison by order of target text rate ... 108

Table 26: General comments ... 110

Table 27: Code-switching audience member ... 114

Table 28: Discussion and general understanding of the work and role of interpreters ... 118

Table 29: Interpreter’s intervention example 1 ... 121

Table 30: Interpreter’s intervention example 2 ... 122

Table 31: Interpreter’s intervention example 3 ... 123

Table 32: Comments related to English as lingua franca ... 128

1

Chapter 1 Introduction

Interpreting quality research is one of the most extensively investigated subjects in

interpreting studies. Understanding how to objectively define and evaluate the quality of

interpretation has profound implications for both interpretation as a profession and an

academic discipline. When AIIC admits new members to the profession, it has to ensure

that the members meet stringent quality criteria and that there is a set of common

standards in the screening system (Kurz, 1993). Interpreting quality is also relevant to

the education and training of professional interpreters, take for instance the assessment

criteria used in entrance exams or qualifying tests in interpretation schools. Yet is there

an objective definition of interpreting quality, what is it, who defines it, and what are

some of the empirically sound methods available when probing the issue? Descriptively,

what are the factors that decide the quality perceptions of interpreting users? Many

researchers have poured significant efforts into the study of interpreting qualities,

debated overtime the different meanings and methods, and generated a wealth of

knowledge. This research, set in a social media interpreting context with the use of new

sources of data, aims to use an observational, bottom-up approach to listen to what

2

online interpreting users are saying about the performance of two simultaneous

interpreters in a live interpreting event.

1.1 Background of the Study

Rich literature on interpreting quality has revealed many different dimensions to

define interpreting quality. As summarized by a model developed by Pöchhacker (2001),

interpreting quality ranges from accurate rendition of source, adequate target language

expression, equivalent intended effect, to successful communicative interaction. Good

interpretation has multiple-level of definitions, from the “lexico-semantic core” at

textual output level to the “socio-pragmatic sphere of interaction” (p. 143).

Similarly, researchers have also developed various quality criteria to operationalize

and measure the different aspects of quality. Some are linguistic, output-oriented, and

content-related quality criteria, such as logical cohesion of utterance, sense consistency

with original message, completeness, correct grammar, and correct terminology. On the

other end of the spectrum, there are extra-linguistic, pragmatic, service-orientated, or

form-related quality standards. Take the following criteria for example, pleasant voice,

lively intonation, poise, and pleasant appearance.

Many empirical studies have been conducted to test how interpreters and users

3

perceive the importance of these different quality criteria. In some survey-based studies

(e.g. Büler, 1986; Marrone, 1993; Kopczyński, 1994; Moser, 1996) there seemed to be a

general consensus that content-related criteria are more important than form-related

ones. Meanwhile, form-related, delivery or extra-linguistic criteria have been

considered desirable, but not essential. Yet this is not a final conclusion. With the

accumulation of more literatures, many researchers have gradually realized that

interpreting quality is very multi-dimensional, and differs with at least two key

factors—the users and the interpreting context.

Regarding the role of users, Gile (1991) stated that quality assessment is “affected

by the different actors’ respective positions in the communication configuration and by

the limitations of these positions” (p. 196). Kahane (2000) exemplified how actors hold

different perceptions about what constitutes good interpretation. When empirical

research on interpreting quality started to prosper from the late 1980s, quality studies

started from understanding the perspectives of professional interpreters, but

subsequently more studies focused on surveying general end-users and even

distinguished between user groups of different backgrounds.

Context is another crucial factor affecting interpreting quality. Kopczyński (1994)

4

defined quality as a function of situation and context. Straniero-Sergio (2003) held that

without the situational context in which the interpreted event takes place, there would

be “a gap between the ideal (academic) quality and situated (real-world) quality”

(p.135). Because interpreters’ performance is largely dependent upon the context in

which the event took place, one has to frame the situation first, and then consequently

decide the realistic, achievable quality of the interpretation. With this theoretical

foundation in mind and also as professional interpretation expands to serve different

social contexts and communities, there have been emerging quality studies in specific

contexts, such as in community, court or media interpreting, just to name a few.

Researchers have attempted to define media interpreting, mainly by describing the

different scenarios where interpretation services are required. According to Pöchhacker

(2010), media interpreting is “a form of language transfer in the media used primarily

for live mass media broadcasts” (p. 224). So far, media interpreting studies focused on

simultaneous interpreting events on the television.

There are mainly two types of television interpreting (Mack, 2002). One typical

example is foreign events with special significance broadcasted on the television, such

as a presidential inaugural ceremony, or a state memorial service. Another typical

5

television interpreting scenario is the in-studio talk shows or interviews where

interpreters provide consecutive interpretation for the program host, interviewees, and

sometimes a live audience. In some countries, such as Japan, there are broadcast

interpreters who play the dual-role of television reporters and interpreters, and are

expected to meet the quality standards of both professions (Snelling, 1997).

Mack (2002) also pointed out two prominent features of media interpreting that set

it apart from simultaneous interpreting in a conference setting. First is the unique

participant relationship with users in media interpreting. The users in media interpreting

are often a group of off-line audience. How will the interpreters be able to gauge the

need of these remote users? Another feature is exposure--the nature of the media

enables the performance of a single interpreter to reach thousands, if not millions of

viewers and listeners. In addition, the exposure factor might increase the stress level on

interpreters (Kurz, 2002) as well as the difficulty level of the interpreters’ works

(Pöchhacker, 2011). These unique features of media interpreting could result in

insightful quality-related findings, contributing to making the literatures on interpreting

quality more comprehensive and holistic.

However, in the context of media interpreting, there have only been a small number

6

of studies on interpreting quality. A general consensus seemed to suggest that the unique

context of media interpreting does affect users’ quality perception. Because of the

unique context, delivery- or performance-related quality criteria were deemed to be

more important than content-related ones (Pignataro & Velardi, 2011; Straniero-Sergio,

2003; Pöchhacker, 1997; Kurz & Pöchhacker, 1995). This means that when evaluating

the quality of media interpreters, interpreter’s ability to speak convincingly can

sometimes be even more important than to speak correctly.

However, one has to note the methodological limitations in these small number of

quality studies. Kurz and Pöchhacker (1995) used Büler’s eight quality criteria and

asked survey participants to assess the relative importance of different quality, and

compared with results with Kurz’s study done in 1993. However, the sample size of TV

professionals was only 19 people, and it was unclear from the study whether the

differences between the scores given by the two comparison groups were of statistical

significance or not. Pöchhacker’s (1997) and Straniero-Sergio (2003) respectively

focused on one media interpreting event to analyze and compare the source and target

texts to assess interpreters’ performance. They differed in the sense that Pöchhacker’s

(1997) still applied Büler’s quality criteria, while Straniero-Sergio (2003) identified and

7

described the norms of media interpreters. Pignataro and Velardi (2011), although only a

preliminary study with limited data, observed how interpreters used rhetorical tools to

gain affiliative responses from the audience, which was a measurement of successful

interpretation. With these limited number of studies, the investigation of quality in

media interpreting context is not final, especially as new forms of media and web-based

broadcasting technologies continue to evolve.

1.2 Statement of the Problem

The previous section has underscored that interpreting quality has to be studied

under different contexts and from the perspectives of different users in order to develop

a more balanced, nuanced portrait of interpreting quality. However, as mentioned, only

a handful of studies have been devoted to the issue of quality studies in media

interpreting. Even fewer studies documented the phenomenon of simultaneous

interpreting conducted over the Internet, not to mention quality studies in a social media

context. A clear gap in existing literatures is about how the needs and perceptions of

online interpreting users might change or differ from conventional interpretation users

because of a new interpreting context and an unprecedented speaker-listener-interpreter

relationship. This question would only become more relevant as the usage of Internet

8

and web-based broadcasting technologies become more widely applied in more

simultaneous interpreting events, having an even greater impact of the interpretation

profession.

Many previous interpreting quality studies also contained a number of

methodological limitations. This study expands the territory of media interpreting with

new methodological options presented in this particular context. Fundamentally, instead

of asking users what quality criteria, specific or general, they value the most, the

researcher tries to observe and listen to what users are saying about the interpreters.

Descriptive, instead of normative, interpreting quality criteria are explored through an

in-depth analysis of the YouTube user comments.

1.3 Purpose of the Study

It is with this background that the research takes place. This research seizes a

unique data collection opportunity and constructs an observational case study to

examine users’ perceptions of the quality of simultaneous interpreting in a live

streaming interpreting event on YouTube. On December 11, 2012, Harvard Professor

Michael Sandel was invited by Taiwan’s Ministry of Culture to give a lecture about his

new book, What Money Can’t Buy: The Moral Limits of Markets. The lecture was held

9

at a stadium with an audience of 6,000 people. At the same time, the lecture was

broadcasted live on YouTube, so simultaneously there was a group of online users

listening to the lecture. Because the YouTube streaming did not provide dual channels,

the online audience had no choice but to listen to the simultaneous interpretation, while

the original source speech was broadcasted at the backdrop at a lower volume in the

same channel. In other words, online YouTube viewers were default absent users of the

interpreting service. Most importantly, the YouTube audience not only watched Sandel’s

lecture and listened to the interpreters online, but many of them left online comments

simultaneously. This use of social media data is advancement from existing media

interpreting studies. The YouTube comments comprised the basis of this research. A

total of 233 comments from 134 unique users were collected to form a rich set of

uncontrolled, natural data on online users’ perception of interpreting quality. This data

was closely examined to understand the different quality criteria, both narrowly- and

broadly-defined, used by different users, including a cross-reference analysis between

the YouTube comments and lecture transcription.

1.4 Significance of the Study

The importance of exploring online users’ interpreting quality perception is

10

two-folds. First, although media interpreting is not a new phenomenon, simultaneous

interpreting broadcasted over the Internet is a relatively more recent development that is

under-studied in interpreting research. The interpreting event in this research is not a

conventional media interpreting event but one that was webcasted live on YouTube,

which further expands the boundary of quality studies in media interpreting. As the

phenomenon of Internet interpreting becomes more prominent, interpreting quality

studies also have to keep up with new norms and practices, while quality standards and

criteria used to measure interpreting performance also have to evolve accordingly. This

research explores the key issue of how quality criteria might remain the same or differ

in this new interpreting context and how a much larger group of online interpretation

users perceive interpreting quality.

Media interpreting researchers (e.g. Mack, 2002) used to think of remote users as

those who are “undifferentiated, anonymous and numerous, with no possibility of active

participation” (p. 207). In most situations, the role of remote online users is invisible

and their voices unheard. This is not only because of the remote, off-site nature of the

users, but also because the users are voluntary, unsolicited, and can come from

11

anywhere in the virtual world.¹ This makes it almost impossible to understand quality

from their perspectives. But the nature of a social media like YouTube allows users to

leave comments while viewing the interpreting event. And this is exactly the second

contribution of this study. In addition to documenting a new phenomenon in media

interpreting, this research also captured new data to shed light on new issues and

challenges relevant to Internet interpreting. New data is available because of the nature

of social media, where Internet users are no longer passive listeners but active

participants who can voice their opinions about the interpretation performance

simultaneously along with the interpreting event. These YouTube comments about the

interpreters’ performance might only be a tip-of-the-iceberg, but at least they can start to

draw researchers’ attention to listen to the voice of this group of interpreting users. And

by understanding their needs and the factors that affect interpreting quality, professional

interpreters who engage in these online interpreting assignments can play better roles or

further advance their profession.

Another equally important component of the new data is the complete transcription

data of both the source and target texts. Pöchhacker (2011) pointed that in many cases

1 One exception is the scenario of virtual conferences or webinars where online users have to obtain passwords to access the web content, including the interpreter’s rendition.

12

interpreted speeches were broadcasted in voice-over mode in a single audio channel so

the original speech was barely audible. But this study is able to overcome the

above-mentioned limitation, allowing the researcher to cross-examine the transcription

and users comment data, so that interpreting quality can be analyzed and

cross-examined from different perspectives.

1.5 Primary Research Questions

Based on the literature reviewed and the gaps identified in existing studies, the

researcher proposes the following research question in the hope of informing users’

quality perception in an online interpreting context:

Research Question: What is the interpreting quality perception of the social media users, and specifically, what quality criteria can be elucidated from the YouTube

responses and how do these quality criteria differ from or support previous findings?

The research question aims to find out what quality criteria were noticed by the

social media users who viewed the webcast of a highly interactive lecture, and how they

differ or align with the commonly used quality criteria categories documented in

existing literature.

13

1.5 Assumptions, Limitations, and Scope

This is a qualitative, observational study of an interpreting event that was

webcasted on YouTube. The event was a lecture by Harvard Professor Michael Sandel,

and the topic were relevant issues covered in his book, What Money Can’t Buy: The

Moral Limits of Markets. Sandel addressed an audience of 6,000 people in a stadium at

National Taiwan University. Two interpreters were invited to provide simultaneous

interpretation for both Sandel, the audience in the stadium, as well as the online viewers

who watched the YouTube webcast. The online viewers not only watched the event, but

many of them also left comments, and a significant number of comments (233

comments) were interpretation-related. The researcher collected these YouTube

comments on December 18, 2002, one week after when the lecture took place.

Comments were mainly posted simultaneously during the event, while a smaller number

of comments were posted shortly after the conclusion of the event. Comments posted

after December 18, 2012, were not included in this study. The researcher assumes that

the comments describe the interpreting perceptions of a group of YouTube listeners, but

do not reveal the normative standards of interpreting quality. In addition, information is

not sufficient to establish a user profile nor claim that the profile of the YouTube

14

listeners resembled that of the live audience at the stadium of National Taiwan

University, who were mainly composed of university students. However, the YouTube

users were vocal enough to post large number of comments, indicating that they were a

group of active “netizens,” which might result in a certain degree of selection bias. The

researcher also assumes that the user comments were unsolicited opinions expressed in

a most natural setting by audience members. So by studying these comments, the

research can have a glimpse into how users really think of or judge the performance of

interpreters in an authentic interpreting event. Meanwhile, the researcher also completed

a full transcript of the source text and target text. This research focuses mainly on the

YouTube user comments and transcription data to develop the main arguments.

1.6 Thesis Organization

The thesis is divided into five chapters and an appendix section. The first chapter

provides a brief introduction about the background of the study. It also describes the

objectives and significance of this research and presents the main research question.

Chapter Two presents a comprehensive review of related literature on interpreting

quality, both in terms of specific quality criteria commonly applied in previous

empirical works as well as broader interpreting quality factors that have been

15

documented. A thorough review of research works on media interpreting is also

included. The third chapter describes specific data collection and analytical methods

used in this research, and explains how such a research design is suitable to address the

main research question.

Chapter Four presents the findings and discussions of the research results, starting

with findings from the lecture transcription and YouTube comment data, followed by

detailed discussions on multiple aspects of interpreting quality, including major quality

criteria and broader factors affecting quality perception. Finally, Chapter Five concludes

the thesis by presenting a summary of the main findings, limitations of this study, and

suggestions for future research. There is also an appendix section, which includes the

complete lecture transcription, original YouTube comments, as well as a profile of the

audience members who interacted with Sandel.

16

Chapter 2 Literature Review

The aim of this chapter is to establish a theoretical foundation for this study and to

identify the gaps to be filled in the existing research methods and knowledge concerning

users’ perception of interpreting quality. The review starts from understanding existing

definitions and constructs of interpreting quality, followed by mapping out broader

factors that affect interpreting quality, and in particular, the increasing importance of

technology and media.

2.1 Defining Interpreting Quality

For a very long time good interpretation was merely a subjective judgment. There

lacked a consensus about what constitutes interpreting quality and how it should be

assessed. No wonder AIIC referred to quality as “that elusive something which

everyone recognizes but no one can successfully define” (AIIC, 1982, as cited in

Zwischenberger, 2010, p.128).

Although quality can be very subjective, since the late 1980s more and more

researchers started to address the challenge of defining and understanding interpreting

quality (Kurz, 2001). Earlier researchers defined good interpretation as one that does

17

not deviate or differ from the original text. Quality criteria are at the comparative

linguistic level, focusing on syntactic and semantic parameters (Barik, 1994; Altman,

1994). At this stage, researchers seldom took into account broader factors, such as the

context, actors, text type, or process of the interpreted event, so the findings often were

not readily applicable to real-life interpreting norms. Déjean Le Féal (1990) argued that

good interpretation carries an equivalent effect on the listeners as the source

text—“what our listeners receive through their earphones should produce the same

effect on them as the original speech done on the audience” (p.155). Moreover, the

target text should present the original content with the same level of clarity and

precision, and its “language and oratory quality should at least be on the same level, if

not better” (p.155). Moving beyond comparing source and target text, more researchers

started to view good interpreting from broader and more holistic perspectives. Gile

(1998, as cited in Kalina, 2005) took a processing viewpoint and defined interpreting

quality as “the optimum balance between different processing efforts.” Interpreting

quality deteriorates when one of the processing efforts (listening, memorization, or

speaking) is overloaded, straining the total processing capacity. Taking an encompassing

viewpoint, Pöchhacker (2001) developed a model of quality standard, where

18

interpreting quality ranges from accurate rendition of source, adequate target language

expression, equivalent intended effect, to successful communicative interaction. This

model summarizes the fact that good interpretation has multiple-level of definitions,

from the “lexico-semantic core” at textual output level to the “socio-pragmatic sphere of

interaction” (p. 413).

2.2 Measuring Interpreting Quality

2.2.1 Different Quality Criteria

To assess interpreting quality empirically and to actually measure quality standards,

there have been many endeavors and debates about research methods. One of the most

common empirical studies on interpreting quality has been survey research, as

compared to experimental studies, corpora analysis or case studies (Pöchhacker, 2002).

One of the pioneering survey-based research works started from Büler (1986). Since

then researchers have developed and applied a range of quality criteria, and asked the

respondents to rate the criteria by order of significance. These criteria reflected

researchers’ endeavor to operationalize the construct of quality by breaking it into more

tangible components or objective attributes.

There are basically two main categories of quality criteria that have been widely

19

used. The first category is related to accuracy and focuses on linguistic, semantic,

product-oriented, or content-related criteria. These are the core criteria of interpreting

quality. This product-oriented perspective reflects the reasoning that good interpretation

means that the target-text is a “faithful image” (Gile, 1991, p. 198) or “exact, faithful

reproduction” (Jones, 1998, p. 5) of the source text. Some quality criteria that fall under

this category include fluency of delivery, logical cohesion of utterance, sense

consistency with original message, completeness, correct grammar, and correct

terminology. The second category emphasizes on style and looks at extra-linguistic,

pragmatic, service-orientated, or form-related standards. These criteria are more related

to “listener-orientation” and “target-text comprehensibility” (Pöchhacker, 2001, p. 413).

Take the following criteria for example, pleasant voice, lively intonation, poise, and

pleasant appearance. Interestingly, some criteria, such as fluency of delivery,

synchronicity, or native accent can be categorized as either content- or form-related,

depending on the researchers’ interpretation.

In addition, there can be a sub- or third category, which refers to personality traits

(Büler, 1986). These criteria include those such as thorough preparation, endurance,

reliability, and ability to work in a team. Note that these criteria cannot be directly and

20

explicitly observed or experienced by the audience.

2.2.2 Perceived Importance of Different Quality Criteria

There has been a constant debate about which category of quality criteria are

considered more important. In many studies (e.g. Büler, 1986; Marrone, 1993;

Kopczyński, 1994; Moser, 1996) there seemed to be a general consensus that

content-related criteria are deemed more important than form-related ones. The

linguistic criterion sense consistency with the original has constantly been given the

highest ratings, as well as the accuracy criterion. Meanwhile, form-related, delivery or

extra-linguistic criteria have been considered desirable, but not essential.

However, there are also some studies that found that users might consider more

important “superficial” criteria such as native accent, pleasant voice, and fluency of

delivery, as opposed to linguistic criteria. For instance, Kurz (1993) mentioned that

extra-linguistic criteria might be considered more important in certain situations or

contexts, such as media interpretation or in conferences filled with lively discussion and

spontaneous exchange. There are also some other possible explanations. Users do not

normally understand the source language and its linguistic quality, so they are more

likely to judge interpreting quality based on criteria that they can directly feel or

21

experience (Büler, 1986). Another explanation is that some users have less tolerance

listening to a halting interpretation, even if it is logically coherence and correct.

Although formal, delivery related criteria are not considered essential, they still affect

how the users evaluate the real performance of the interpreters (Garzone, 2002).

Looking at each specific quality criterion, one observes an even more complicated

picture. According to user comments in a study by Pöchhacker and Zwischenberger

(2010), the criterion synchronicity is more important in certain situations or in speeches

with certain features. Take for example speeches with jokes, punch lines, or lots of

numerical data. The criterion native accent is often considered less important or

unimportant. But in the same research by Pöchhacker and Zwischenberger, some survey

respondents expressed the viewpoint that the importance of native accent depends on

the target language. Another user in the same study pointed out that accent might be

related to prosodic quality (e.g. native intonation), which elevates the importance of

accent. These are only observational comments given by individual survey respondents.

However, empirical evidence can be found in a recent experimental study by Cheung

(2013). When judging three versions of an English-to-Cantonese SI differing only in

accent, participants rated the native accent SI quality more favorably than the two other

22

SI that had Mandarin and English accent respectively. The researcher cannot review and

discuss every single criterion in existing studies but the isolated examples listed here are

meant to highlight how and why some form-related criterion were attached significant

importance.

Whether content- or form-related, these quality criteria studies reflect researchers’

efforts to operationalize the construct of quality by breaking it into more tangible

components or attributes. The goal of many of these studies is not necessarily to reach

an absolute dichotomic conclusion about which category of quality criteria is more

important, but to highlight the fact that the discussion of interpreting quality should also

consider the type of interpreting event or assignment, who the respondents are, the

expectations of the respondents, how the respondents perceive the interpreter’s role, and

many other contextual factors.

2.2.3 Challenges in Measuring Quality

Despite the fact that the topic of interpreting quality criteria has been extensively

studied, there still exist quite a number of research challenges. As Moser-Mercer (2009)

pointed out in her article “Construct-ing Quality,” many existing studies were “overly

ambitious”, trying to cover too many attributes of quality in a non-specific way (p. 156).

23

When a study eliminates the situational context, it would create “a gap between ideal

(academic) quality an situated (real-world) quality” (Straniero-Sergio, 2003, p. 135). In

addition, the many quality attributes are often “suggested by researchers and not by

uses”—users mostly likely do not think in terms of quality (e.g. this is a good/bad

interpretation), but more in terms of other constructs such as comprehension (e.g. I

don’t understand what the speaker is trying to say) (p. 157). This causes a construct

validity problem. In addition, many quality criteria are not only subjective, but can also

be interpreted into different meanings. For example the criterion fluency of delivery can

be further characterized by frequent hesitations, pauses, or irregular, unnatural rhythms.

So when a respondent rates fluency of delivery as important, it might be due to a general

feeling or a certain characteristic that he/she thought about. To address this issue,

Macías (2006) used silent pauses as a sub-parameter of fluency to design an experiment.

Respondents were asked to rate the interpreting quality of three simulated simultaneous

interpretation videos, including the control video without additional silent pauses.

Findings indicated that silent pauses have a negative effect on the quality parameter of

fluency. Still, much research design and methodological work remains to be done to

ensure that the construct of quality is clearly measured.

24

2.3 Interpreting Quality in the Ears of Different Users

2.3.1 Importance of Users’ Perspectives

In addition to studying the various quality criteria, it is equally crucial to

understand quality of interpretation from different users’ perspectives. Gile (1991)

stated that quality assessment is “affected by the different actors’ respective positions in

the “communication configuration” and by the limitations of these positions” (p. 196).

Moreover, the actors’ motivation and attention would also shape how they judge the

interpreting quality. These actors include the Sender, the Receiver, the Client and the

interpreters themselves. More specifically, professional interpreters (senior or novice),

listeners (including professionals or experts who attend conferences regularly and are

familiar with interpretation service versus general audience who seldom have the

chance to receive interpretation service), and clients who pay for the interpreting

services (e.g. agents and event organizers) might all have different perceptions or

understanding about what constitutes good interpretation (Kahane, 2000). If

interpretation is perceived as a type of service, when trying to evaluate the quality of the

service, one can either try to understand the end-users’ perceptions, or those of the

service providers, namely the interpreters (Moser, 1996; Kurz, 2001). These viewpoints

25

are usually mutually complementary rather than exclusive. In addition, Viezzi (1996, as

cited in Pöchhacker, 2001, p. 412) not only distinguished between the viewpoints of the

interpreters and users, but also included the perspective of the “external observer” –

namely those are studying the subject of interpreting and measuring its objective

features. And even if the user groups are categorized clearly, there are still many

individual factors that shape the users’ preferences, such as “cultural habits, knowledge

of and interest in a given subject, personal attitudes and subjective impressions” (Kalina,

2005, p. 774).

To model the rather complex relationships between these various positions and

perspectives on interpreting quality, Pöchhacker (2001) developed a model to include

the triad of interactants, or actors, composed of the interpreter, the speaker, and the

listener. Additional actors are the client (employer) and colleague (fellow

interpreter/team member). Importantly, when a researcher studies these multiple

perspectives and perceptions, he/she can either investigate an abstract, off-site,

hypothetical event, or refer to a concrete communicative event in a specific context.

Moreover, when the researcher chooses a concrete event approach, he/she may take

either a product-orientation or interaction-orientation approach, focusing either on the

26

recordable interpreting output or the entire process of communicative interaction.

2.3.2 Different Perspectives between Professional Interpreters and General Users

When empirical research on interpreting quality started to prosper from the late

1980s, quality studies started from understanding the perspectives of professional

interpreters, but subsequently more studies focused on surveying general end-users or

listeners. Some studies also compared the quality expectations between professional

interpreters and conference participants. Ideally both viewpoints should be considered

simultaneously, and it is still debatable regarding whether it is best to understand

interpreting quality from the perspective of an interpreter or an end-user (Chiaro &

Nocella, 2004). The following sections summarize the quality perceptions of

professional interpreters and interpreting users respectively.

Some of the earlier quality research focused on the perspectives of professional

interpreters. Take Büler’s pioneering empirical study in 1986 for instance. She surveyed

AIIC sponsors, who were all experienced conference interpreters, regarding quality

evaluation standards they consider when endorsing potential candidates for their AIIC

membership. The quality criteria in her work were meant to “reflect the requirements of

27

the user as well as the fellow interpreter in a well-balanced mixture” (Büler, 1986, p.

233). She found that linguistic/semantic criteria were considered the most important,

with the criterion of sense consistency with the original message being ranked the

highest among all the criteria. Meanwhile, the extra-linguistic/ pragmatic criteria were

considered desirable, but not essential. In a more recent study by Pöchhacker and

Zwischenberger (2010), they also surveyed AIIC members and confirmed many

findings from Büler’s work. More details are included as below.

Even more interpreting quality studies looked at perspectives from interpreting

users. Kurz (1993) tested Büler’s hypothesis by using eight of Büler’s quality criteria,

but she surveyed three groups of conference attendees, including medical doctors in a

medical conference, engineers in an international conference on quality control, and

delegates at the Council of Europe. There was general agreement by all groups on the

importance/unimportance of some of the quality criteria (e.g. sense consistency with

original message, pleasant voice, and native accent). Meanwhile, there were also many

discrepancies in terms of overall ranking as well as scores attached to individual quality

criterion. For instance, Delegates at the Council of Europe considered the criterion

correct terminology as the most important, giving higher scores than did the two other

28

groups. On the surface such a result might be surprising, as one would expect correct

terminology to be considered more important in scientific or medical conferences. But

Kurz (1993) offered a likely explanation, arguing that the delegates at the meetings of

international organizations were used to a specific terminology and thus expected the

interpreters to use those technical jargons that they were most used to and thus were less

tolerant towards any deviations. In addition, because of the shared criteria with Büler’s

earlier study, Kurz could compare the quality expectations of both professional

interpreters and users. She found that while all user groups shared some agreement

when assessing the different quality criteria, the AIIC interpreters in Büler’s study

seemed to demand higher interpreting quality than the participants in Kurz’s research. In

other words, professional interpreters were more stringent and had higher quality

standards than common conference delegates.

Testing the same hypothesis between different user groups and quality expectations,

Moser (1995, 1996) conducted a large-scale study, holding 201 standardized interviews

with end-users at 84 different meetings. The interviewee profile was rather diverse.

They included first-time interpretation service users, meeting participants who only

have limited experienced in using interpretation service, and frequent attendees of

29

multilingual conferences with considerable experience of interpretation. Moser also

provided gender and age distribution of the participants. With this large and diverse

interviewee sample, Moser was able to examine the correlation between

conference-going experience and users’ quality expectations. The four quality criteria

categories that he used included faithfulness of interpretation to the original content,

synchronicity, rhetorical skills, and voice. He found that highly experienced users

attributed a much higher weight to the faithfulness to original content criterion than to

the three other criteria. The less experienced conference goers attached similar

importance to all the four criteria. Meanwhile, the newcomers, those respondents who

used interpretation for the first time, valued good rhetorical skills the most. When

comparing the requirement for faithful content across the three user groups, the highly

experienced users valued it more than the less experienced respondents. In addition to

conference-going experience, Moser’s study also looked into quality expectation

differences between different gender and age groups, and yielded interesting findings.

For example, women were more disturbed by lack of synchronicity and more sensitive

to “ums” and “aahs” or other fillers and pauses. Unfortunately, although this research

conducted interviews and open-ended questionnaires, this research did not provide

30

explanations for the different quality expectations across different user groups.

Similar to Kurz’s (1993) goal to compare quality criteria importance perceived

between professional interpreters and interpreting users, Ru (1996) surveyed 20

conference interpreters in Taiwan and 166 interpreting users. Respondents were first

asked to rate the importance of seven quality criteria, including pronunciation, fluency

of delivery, coherence, speech rate, faithfulness, professional terminology, and pleasant

voice. Then respondents were given hypothetical scenarios to understand their

expectations towards the function and role of interpreters. An example of a scenario was

“When the speaker talks in a lively intonation, should the interpreter mimic the

speaker’s tone and also deliver an animated rendition?” Ru found that both interpreters

and interpreting users ranked faithfulness as the most important quality criteria,

followed by coherence and fluency of delivery. Both groups also ranked pleasant voice

as the least important criteria. Ru studied the perspectives of interpreters and listeners in

Taiwan, but the findings align with previous studies, showing agreements between the

two groups.

In a more recent study by Pöchhacker and Zwischenberger (2010) that surveyed

AIIC members and yielded 704 responses, it highlighted how socio-demographic and

31

professional characteristics of the surveyed interpreters affected their quality

expectations or judgment. These characteristics included gender, age and years working

as a professional interpreter. Generally the survey participants were more demanding for

the form-related quality parameters than for the delivery-related criteria, but the degree

of importance also varied across different meeting types, domains, and individual

preferences or expectations. For example, when tested on two audio samples, one with

lively intonation and the other with monotonous intonation, female interpreters seemed

to be more generous judges than their male counterparts and were more appreciative of

lively intonation.

These studies validated the hypothesis that different user groups have varying

quality perceptions and attach different weights to the various quality criteria. Yet

understanding the quality perceptions of professional interpreters or conference

participants only provide a partial picture of interpreting quality. Some studies argue

that neither professional interpreters nor end-users are capable of making quality

assessment (Chiaro & Nocella, 2004). How should interpreters know what is good for

users, and what users need in the interpreting context? And if we apply the service

concept in business and marketing, since professional interpretation is a kind of service,

32

service providers have to understand the needs and satisfaction degree of the service

users (Kotler & Armstrong 1994, as cited in Kurz, 2001). Similarly, do the users know

what is good for them? (Shlesinger, 1997) Moreover, “How can they (users) know for

sure whether the service provided is adequate?” (Garzone, 2002, p. 118). Although

end-users are people who “consume” the interpretation service, they do not know the

source language nor do they know much about interpretation. So an average end-user

might not be the best person to judge whether the interpretation is good or bad. At best,

general users provide their perception of the interpreted texts, which is only one

dimension of the quality paradigm (Kalina, 2002). Because of these limitations, Kalina

(2002) called for a new model that “encompasses the communication situation, the

intentions and knowledge bases of the different actors (including the interpreters), and

any conditions liable to affect the interpreted event” (p. 123-124).

2.4 Broader Factors Affecting Interpreting Quality

The previous section highlights the fact that both the concept of quality and the

operation of quality criteria are subjective and only reflect certain aspects of interpreting

quality. In addition to the two main categories of quality criteria mentioned above, more

recently, some researchers started to broaden the research on interpreting quality,

33

looking at a broader set of factors that may affect interpreting quality in a

communication context. Garzone (2002) has noted this shift, indicating that research has

departed from “a purely “linguistic” and “technical” approach and moved towards a

broader view, based on a notion of interpretation as “a complex interactional and

communicative event encompassing pragmatic and sociolinguistic factors” (p. 107).

Many previous studies strived to understand the “ideal” quality standards with more

focus on the interpretation output. But subsequently researchers started to look at factors

in real conditions that interpreters have to face.

Moser-Mercer’s (1996) framed the concept of “optimum quality” in a context-- “an

interpreter provides a complete and accurate rendition of the original that does not

distort the original message and tries to capture and all extra-linguistic information that

the speaker might have provided subject to the constraints imposed by certain external

conditions” (p.44). Using media interpreting situations as examples, Straniero-Sergio

(2003) emphasized that quality studies have to be conducted in a particular

context—“frame the conditions in which the interpreter has to translate a particular text,

and, consequently, consider and decide the achievable quality of the interpretation”;

otherwise the quality standards are only “wishful thinking” that exist only in an ideal

34

world (p. 171).

Kalina (2002) looked at factors relevant to interpreting quality that belong to

different phases of an actual interpretation activity, covering a wide range of factors

from contract specifications, technical equipment, availability of source text

presentation, to output review (e.g. recording). Kellett Bidoli (2000) also provided a

thorough review of factors that shape interpreting quality, including environmental

factors, experience-related factors, inter-personal/social factors, linguistic features,

para-linguistic features, non-verbal communication features, physical and mental factors,

prosodic features, situational factors, task-related factors, technical features, textual

features, and time factors. In a pilot study by Moser-Mercer, Künzli, and Korac (1998),

the authors investigated the effect of prolonged turns, defined as those lasting more than

30 minutes, on interpreters’ quality output. They found that prolonged turns resulted

into more meaning errors and a higher level of mental fatigue. Moreover the interpreters

were not even aware of the decline of their interpreting quality. These long lists of

factors can never be exhaustive, yet they continue to broaden and deepen the field of

interpreting quality research.

Although this research is unable to review every single factor, we will review in

35

length one of the less-discussed quality-related factors that is relevant to this research,

namely, technology and broadcasting media (see 2.5).

2.5 Media Interpreting

2.5.1 Definitions and Context

Many previous studies chose conferences settings to test the various quality criteria.

Researchers have selected conferences of different topics and with different audience

size, but many findings are still rather similar (Kahane, 2000). In contrast to quality

studies in conference interpreting, there have been fewer interpreting quality studies in

the field of media interpreting. Yet as electronic media and telecommunications

technology become more sophisticated and prevalent, there are more topics and

methodological options to be explored in this field.

First of all, it is important to understand the different definitions and scenarios of

media interpreting. Pöchhacker (2010) defined media interpreting as “a form of

language transfer in the media used primarily for live mass media broadcasts” (p. 224).

Here, the form of media refers mainly to television. But earlier in history, simultaneous

interpreting was already broadcasted over the radio, when in the 1930s Hitler’s speeches

were interpreted for French radio by acclaimed interpreters such as André Kaminker

36

and Hans Jacob (p. 224). So the practice of media interpreting itself is not new, but the

scenarios, modes, and modalities are constantly changing, especially as new forms of

electronic media arise.

Mack (2002) also distinguished between “on-site scenarios that involve interpreters

in a studio-based communicative event, with or without the presence of an audience,

and simultaneous interpreting of broadcast events occurring in a remote location” (p.

208-209). An example for the former type of television interpreting would be an

interview or a talk show in a TV studio where an interpreter is required to provide short

consecutive interpretation for the host, the interviewee, and sometimes with a live

audience. As for the latter case when a distant event is provided for local audience via

live transmission, a typical example would be a foreign event with special significance,

such as a presidential inaugural ceremony. An interpreter provides simultaneous

interpretation for the audience in front of the television. In addition to these two main

forms of media interpreting, in some countries, such as in Japan, there is also need for

broadcast interpreters (Snelling, 1997). Some news programs produced in foreign

countries are broadcast with Japanese translation on a daily basis. These broadcast

interpreters prepare translation for recorded news programs and report news in a manner

37

that adheres to broadcast standard, be it voice quality, intonation, or pronunciation. In

the occurrence of significant occasions, such as the State of the Union address by the

U.S. president, live simultaneous interpretation is required.

These scenarios of media interpreting have been limited to the TV interpreting

context. But as more international conferences or media events get webcasted or stored

on the Internet, there are greater chances that interpreters’ performance would be

exposed beyond the physical conference site. As new forms of media and broadcasting

technologies emerge, the applicable definitions and likely scenarios of media

interpreting should be further expanded as well. For instance, in addition to the subject

of this study, Professor Sandel’s lecture broadcasted over YouTube, there have been

many recent highly publicized events that were live-streamed over the Internet, and the

interpreters all played visible roles or received much media attention. One is Apple’s

iPhone 6 release event on Sept. 9, 2014, where the voice of a Chinese interpreter was

broadcast to all the online users and overlaid the Apple presenters.² The Chinese

interpretation was supposed to broadcast to only those with an IP address in China, but

due to a technical shortfall, the entire online streaming to global audience was dubbed

2

http://www.forbes.com/sites/antonyleather/2014/09/10/apple-iphone-6-launch-event-technical-issues-and- a-huge-missed-opportunity/