Exploring the process of note-taking and consecutive

interpreting:
A pen-eye-voice approach towards cognitive load

Sijia Chen
Bachelor of Arts
Master of Arts

A thesis submitted in fulfilment of the requirements for the degree of

Doctor of Philosophy

Department of Linguistics
Faculty of Human Sciences
Macquarie University
Sydney, Australia

September 2017
Table of contents

Table of contents ................................................................................................................ i

List of tables..................................................................................................................... vi
List of figures ................................................................................................................. viii
List of appendices ............................................................................................................ ix
Abstract ............................................................................................................................. x
Declaration ...................................................................................................................... xii
Acknowledgements ........................................................................................................ xiii
Chapter 1 Introduction ...................................................................................................... 1
1.1 Literature and theoretical underpinnings ............................................................... 1
1.1.1 Investigating the process and cognitive aspects of interpreting..................... 1
1.1.2 Note-taking in CI ........................................................................................... 2
1.1.3 The construct of cognitive load in interpreting and its measurement ............ 4
1.2 Research scope, research purpose and research questions .................................... 5
1.3 Research design ..................................................................................................... 6
1.3.1 Pen recording ................................................................................................. 7
1.3.2 Eye tracking ................................................................................................. 10
1.3.3 Voice recording ............................................................................................ 10
1.3.4 Participants, tasks and procedures ............................................................... 11
1.3.5 Data and analysis ......................................................................................... 12
1.4 Outline of the thesis ............................................................................................. 12
1.5 References ........................................................................................................... 15
An introductory note to Chapter 2 .................................................................................. 20
Chapter 2 Note-taking in consecutive interpreting: A review with special focus on
Chinese and English literature ....................................................................... 21
2.1 Introduction ......................................................................................................... 21
2.2 Note-taking systems and principles: a prescriptive starting point ....................... 23
2.3 Note-taking didactics: the beginning of a shift from prescriptive to descriptive 26
2.4 Cognitive and linguistic aspects of note-taking: a theoretical drive.................... 29
2.5 Exploring the key note-taking features: descriptive studies on notes and
quality .................................................................................................................. 30
2.5.1 The choice of form and language in note-taking ......................................... 31
2.5.2 The relationship between note-taking and interpreting quality ................... 34

i
 2.6 Limitations of previous studies ............................................................................ 36
2.7 Cognitive load: a promising avenue for investigation ......................................... 37
2.8 Bibliography ........................................................................................................ 38
An introductory note to Chapter 3................................................................................... 45
Chapter 3 The construct of cognitive load in interpreting and its measurement ............ 46
3.1 Introduction .......................................................................................................... 46
3.2 Investigating cognitive load in interpreting ......................................................... 48
3.3 Mental workload in human factors research and cognitive load in Cognitive
Load Theory ......................................................................................................... 49
3.4 Cognitive load in interpreting: an illustration of the construct ............................ 51
3.4.1 Task and environmental characteristics ....................................................... 51
3.4.1.1 Task characteristics ............................................................................... 51
3.4.1.2 Environmental characteristics ............................................................... 53
3.4.2 Interpreter characteristics ............................................................................. 54
3.4.3 Interactions ................................................................................................... 55
3.5 Measuring cognitive load in interpreting: a methodological discussion ............. 55
3.5.1 Cognitive load measures .............................................................................. 55
3.5.2 Selection criteria ........................................................................................... 58
3.6 Techniques for cognitive load measurement in consecutive interpreting ............ 59
3.6.1 Subjective rating scales ................................................................................ 59
3.6.2 Primary task performance, pen recording, and eye tracking ........................ 60
3.6.3 Models .......................................................................................................... 63
3.7 Conclusions .......................................................................................................... 64
3.8 References ............................................................................................................ 65
An introductory note to Chapter 4................................................................................... 72
Chapter 4 Note-taking in consecutive interpreting: New data from pen recording ........ 73
4.1 Introduction .......................................................................................................... 73
4.2 Note-taking in CI: a brief review ......................................................................... 74
4.3 Method ................................................................................................................. 78
4.3.1 Participants ................................................................................................... 78
4.3.2 Apparatus...................................................................................................... 78
4.3.3 Tasks ............................................................................................................. 79
4.3.4 Procedures .................................................................................................... 80
4.3.5 Data and analysis .......................................................................................... 80

ii
 4.3.5.1 Categorisation of note units.................................................................. 81
4.3.5.2 Calculation of the ear-pen span ............................................................ 82
4.3.5.3 Human evaluation................................................................................. 84
4.4 Results and discussion of note-taking choices and their relationship with
interpreting performance ..................................................................................... 85
4.4.1 Choice of form and language ....................................................................... 86
4.4.1.1 Choice between language and symbol ................................................. 86
4.4.1.2 Choice between abbreviation and full word ......................................... 87
4.4.1.3 Choice of language ............................................................................... 88
4.4.2 Relationship between note-taking and interpreting performance ................ 88
4.4.3 Discussion .................................................................................................... 89
4.5 Results and discussion of the pen data: potential indicators of cognitive load
in CI ..................................................................................................................... 92
4.5.1 Pen data on the choice of form .................................................................... 92
4.5.1.1 Between language, symbols, and numbers ........................................... 92
4.5.1.2 Between abbreviations and full words ................................................. 93
4.5.2 Pen data on the choice of language .............................................................. 93
4.5.3 Interpreting the findings from a cognitive load perspective ........................ 94
4.6 Conclusions ......................................................................................................... 96
4.7 References ........................................................................................................... 97
An introductory note to Chapter 5 ................................................................................ 101
Chapter 5 The process of note-taking and consecutive interpreting: Evidence from
digital pen recording .................................................................................... 102
5.1 Introduction ....................................................................................................... 102
5.2 Research background......................................................................................... 103
5.2.1 A review of studies on note-taking in CI ................................................... 103
5.2.2 Research questions ..................................................................................... 107
5.3 Method ............................................................................................................... 107
5.3.1 Participants ................................................................................................. 107
5.3.2 Apparatus ................................................................................................... 108
5.3.3 Tasks .......................................................................................................... 108
5.3.4 Procedures .................................................................................................. 109
5.3.5 Data and analysis ....................................................................................... 110
5.3.5.1 Tagging the source speech ................................................................. 110

iii
 5.3.5.2 Calculating the EPS ............................................................................ 110
5.3.5.3 Categorising the notes ......................................................................... 111
5.3.5.4 Detailed pen recording data and the interpreting performance........... 111
5.4 Results ................................................................................................................ 113
5.4.1 What do interpreters note down ................................................................. 113
5.4.2 When do interpreters take notes ................................................................. 114
5.4.3 How do interpreters take notes ................................................................... 114
5.4.4 Why do interpreters make different note-taking choices ........................... 116
5.4.4.1 The physical, temporal and cognitive demands of different
note-taking choices ............................................................................. 116
5.4.4.2 Note-taking and interpreting performance .......................................... 118
5.5 Discussions ........................................................................................................ 119
5.5.1 The content of note-taking ......................................................................... 119
5.5.2 The timing of note-taking ........................................................................... 120
5.5.3 The choice of form and the choice of language ......................................... 120
5.5.4 Note-taking, cognitive load and interpreting performance ........................ 121
5.6 Conclusion ......................................................................................................... 123
5.7 References .......................................................................................................... 124
An introductory note to Chapter 6................................................................................. 129
Chapter 6 An eye-tracking approach to note-reading in consecutive interpreting:
Reading patterns and cognitive load ............................................................ 130
6.1 Introduction ........................................................................................................ 130
6.2 Note-taking and note-reading in CI ................................................................... 132
6.3 Eye tracking and cognitive processing in language tasks .................................. 134
6.4 Purpose of the study and research questions...................................................... 138
6.5 Method ............................................................................................................... 139
6.5.1 Participants ................................................................................................. 139
6.5.2 Apparatus.................................................................................................... 140
6.5.3 Tasks ........................................................................................................... 140
6.5.4 Experimental procedure ............................................................................. 141
6.5.5 Data and analysis ........................................................................................ 142
6.5.5.1 Semantic Gaze Mapping and AOI drawing ........................................ 142
6.5.5.2 Eye tracking measures used in this study ........................................... 145
6.5.5.3 Statistical analysis ............................................................................... 145

iv
 6.6 Results ............................................................................................................... 146
6.6.1 How interpreters read back their notes ...................................................... 146
6.6.2 The relationship between the note-taking choices and cognitive load in
Phase II of CI ............................................................................................. 147
6.6.2.1 The choice between language and symbol ......................................... 147
6.6.2.2 The choice between abbreviation and full word ................................ 148
6.6.2.3 The choice between Chinese and English .......................................... 150
6.7 Discussion.......................................................................................................... 151
6.7.1 Note-reading, reading for comprehension and reading for translation ...... 151
6.7.2 Note-taking choices and cognitive load during Phase II of CI .................. 152
6.8 Conclusions ....................................................................................................... 154
6.9 References ......................................................................................................... 155
Chapter 7 Summary, discussion and conclusion .......................................................... 161
7.1 Main findings of the research ............................................................................ 161
7.1.1 Note-taking preferences ............................................................................. 161
7.1.2 Note-taking and cognitive load in CI ......................................................... 164
7.1.3 Note-taking and interpreting performance ................................................. 165
7.2 Note-taking, cognitive load and interpreting performance: the interplay and
implications ....................................................................................................... 166
7.3 Strengths and limitations of the study ............................................................... 167
7.4 The way forward................................................................................................ 169
7.5 References ......................................................................................................... 170
Appendices.................................................................................................................... 173

v
List of tables

Table 1.1 Criteria for recruiting participants ................................................................... 11

Table 1.2 Data collected in the study .............................................................................. 12
Table 2.1 A summary of studies on key note-taking features ......................................... 33
Table 2.2 Findings on the choice of form in note-taking ................................................ 33
Table 2.3 Findings on the choice of language in note-taking ......................................... 34
Table 2.4 A summary of studies on the relationship between note-taking and
interpreting quality ......................................................................................... 36
Table 3.1 Major categories of cognitive load measures and their applications .............. 57
Table 4.1 Text analysis results ........................................................................................ 79
Table 4.2 A summary of the tasks ................................................................................... 80
Table 4.3 Data used for analysis ..................................................................................... 81
Table 4.4 Categories and definitions of note units .......................................................... 82
Table 4.5 Distribution over form: language vs. symbol .................................................. 86
Table 4.6 Distribution over form: abbreviation vs. full word ......................................... 87
Table 4.7 Distribution over language .............................................................................. 88
Table 4.8 Relationship between note-taking and interpreting performance ................... 89
Table 4.9 Pen data on the choice of form ........................................................................ 93
Table 4.10 Ear-pen span data on the choice of language ................................................ 94
Table 4.11 How physical, temporal and cognitive demands affect interpreters’
note-taking choices ......................................................................................... 96
Table 5.1 A summary of the two tasks .......................................................................... 109
Table 5.2 Categories of the source speech units and their frequencies of being noted
down ............................................................................................................. 114
Table 5.3 The choice of form ........................................................................................ 115
Table 5.4 The choice of language ................................................................................. 116
Table 5.5. A comparison of language and symbol notes in terms of distance,
duration and EPS .......................................................................................... 116
Table 5.6 A comparison of abbreviation and full word notes in terms of distance,
duration and EPS .......................................................................................... 117
Table 5.7 A comparison of Chinese and English notes in terms of distance, duration
and EPS ........................................................................................................ 118

vi
Table 5.8 The Pearson’s correlation between interpreting performance and
note-taking ................................................................................................... 118
Table 5.9 The physical, temporal and cognitive demands and interpreters' preferred
note-taking choices ...................................................................................... 122
Table 6.1 Note-taking choices and cognitive load in Phase I of CI.............................. 133
Table 6.2 Task specifications........................................................................................ 141
Table 6.3 Experimental procedure ................................................................................ 142
Table 6.4 Eye tracking measures used in this study ..................................................... 145
Table 6.5 Descriptive statistics of note-reading............................................................ 147
Table 6.6 The choice between language and symbol ................................................... 148
Table 6.7 The choice between abbreviation and full word ........................................... 149
Table 6.8 The choice between Chinese and English .................................................... 150
Table 6.9 Physical, temporal and cognitive demands of different note-taking choices
in both Phase I and II of CI .......................................................................... 154
Table 7.1 The interplay between note-taking, cognitive load and interpreting
performance ................................................................................................. 166

vii
List of figures

Figure 1.1 A screenshot of the tablet in the recording mode ............................................ 8

Figure 1.2 A sample screenshot of the Eye and Pen software in the analysis mode ......... 9
Figure 1.3 A sample data output of pen recording ............................................................ 9
Figure 1.4 Linkages between the article-based chapters in this thesis by publication .... 13
Figure 3.1 A graphical illustration of the construct of cognitive load in interpreting..... 51
Figure 3.2 A graphical illustration of the measurement of cognitive load in
interpreting ..................................................................................................... 57
Figure 3.3 A hypothetical relationship between cognitive load and interpreting
performance .................................................................................................... 61
Figure 4.1 Categorisation of note units ........................................................................... 81
Figure 6.1 An example of Semantic Gaze Mapping ..................................................... 143
Figure 6.2 An example of drawing the AOIs ................................................................ 144
Figure 6.3 Labelling the AOIs....................................................................................... 145
Figure 6.4 A sample AOI sequence chart ..................................................................... 147

viii
List of appendices

Appendix A: Macquarie University thesis by publication guideline ............................ 173

Appendix B: Ethics approval ........................................................................................ 176
Appendix C: Demographic questionnaire..................................................................... 179
Appendix D: Post-experiment questionnaire ................................................................ 181
Appendix E: Scripts for the two interpreting tasks ....................................................... 183

ix
Abstract

Interpreting is a cognitively challenging language-processing task. Ever since it became

a subject of scientific research, there has been a strong interest in finding out what
happens inside the black box of the interpreter’s mind as they perform this extraordinary
task. This thesis sets out to contribute empirical evidence to elucidate the process of
consecutive interpreting (CI) and note-taking, with a particular focus on cognitive load
inherent in these tasks. It collects data from pen recording, eye tracking and voice
recording to find answers to key questions revolving around CI and note-taking. This
thesis is presented in a thesis by publication format, with its chapters (except for the
introductory and concluding chapters) being stand-alone peer-reviewed journal articles.
The thesis begins with a review of the existing studies on note-taking in
consecutive interpreting. It identifies the key variables of research: the choice of form
(i.e., the choice between language and symbol, and the choice between abbreviation and
full word), the choice of language (i.e., the choice between source and target language,
and the choice between native and non-native language), and the relationship between
note-taking and interpreting performance. After diagnosing two important limitations
with previous studies – a lack of process research and a lack of empirical data – this
review pinpoints cognitive load as a promising avenue for future investigations.
Then, the thesis presents a theoretical and methodological discussion on the
construct of cognitive load in interpreting and its measurement. Borrowing from
adjacent fields in which cognitive load is more systematically studied, this thesis defines
cognitive load in interpreting as a multi-dimensional construct which reflects the portion
of an interpreter’s limited cognitive capacity devoted to performing an interpreting task
in a certain environment. It introduces the categories of cognitive load measures and a
series of selection criteria. Considering that previous cognitive studies mostly focus on
simultaneous interpreting, this thesis introduces techniques that can be used to study
cognitive load in CI.
To test the usefulness of some of the techniques proposed in the methodological
discussion, a pilot study is conducted, the purpose of which is to devise a design that
allows synchronised recording of pen and voice data, a combination that has been rarely
applied in the field. This pilot study provides evidence that pen recording is a powerful
method to tap into the process of note-taking and interpreting, thus paving the way for

x
the main study of this PhD project. Findings of the pilot study are also informative for
the hypotheses made in the next stage of the research.
The main study of the PhD project is carried out by triangulating the methods of
pen recording, eye tracking and voice recording to collect data on the process of note-
taking and CI. It is found that interpreters prefer language to symbol notes and English
(non-native language) to Chinese (native language) notes, regardless of the direction of
interpreting. This is also the first study to visualise the activity of note-reading, showing
that it proceeds in a non-linear fashion and requires significant cognitive cost. The pen
and eye movement data collected in this study provide important indicators of cognitive
load in note-writing, note-reading and interpreting. A combined analysis of the pen, eye
and voice data shows that the note-taking choices are mainly affected by the cognitive
demands, rather than the physical or temporal demands. However, the choices made by
interpreters to lower the cognitive load in the first phase of CI are sometimes at the
expense of interpreting performance. Furthermore, the study detects a trade-off between
the cognitive costs of the two phases of CI.
Understanding the nature of the cognitive processes involved in interpreting is not
only beneficial to the field itself – to inform interpreter education, testing and
continuing professional development – but also more generally enriches our
understanding of bilingual language processing and human cognition. The
methodological and empirical findings of the thesis contribute to that effort and outline
possible avenues for future research.

xi
Declaration

I certify that the work in this thesis is the result of my own research and that the work
has not been submitted for a higher degree to any other university or institution. I certify
that sources of information used and the extent to which the work of others has been
utilised are indicated in the thesis.
The research was approved by Macquarie University Ethics Committee (Human
Research) (see Appendix B) and conducted in accordance with the guideline stipulated.

Sijia Chen
September 2017

xii
Acknowledgements

I am immensely grateful to my principal supervisor Associate Professor Jan-Louis

Kruger. Words can’t even begin to describe how much help and support Jan-Louis has
provided me through the entire PhD candidature. His mentorship and high standard of
scholarship prepared me well for the various challenges that came along the journey. I
couldn’t ask for a better supervisor. The fulfilment of this PhD project would not have
been possible without him.
I would like to express my sincerest gratitude to my co-supervisor Dr Helen Slatyer
and my adjunct supervisor Dr Stephen Doherty. Without Helen, I wouldn’t even have
been able to start my PhD at Macquarie University. She helped me get through the
hardest times, especially the beginning stage which was filled with struggles. Stephen
offered me generous and invaluable support in the stressful moments of my final year.
His guidance, insights and feedback are indispensable to the successful completion of
this PhD research.
I would also like to thank Professor Jing Chen, who supervised my undergraduate
and postgraduate studies at Xiamen University and introduced me to the field of
interpreting research.
My sincere gratitude also goes to Professor Gile Daniel for providing the
Conference Interpreting Research Information Network Bulletins (CIRIN Bulletins),
one of the key databases for the literature review of this thesis. I am deeply touched by
his patience and generosity in sharing his insights with a young researcher like me.
I am grateful to all my colleagues and friends for their kind help and mutual
encouragement. A very special thank you goes to Dr Chao Han for inspiring and
encouraging me both as a friend and as a peer researcher, and for providing critical
feedback on the thesis chapters.
I also owe my gratitude to all the participants, speakers, technicians, administrators,
and all other university staff who have contributed their precious time to help me carry
out the study.
I greatly appreciate the valuable feedback from journal editors and anonymous
reviewers on my published and submitted papers. I also want to thank the reviewers of
this PhD thesis for dedicating their time and effort.
My sincere gratitude goes out to the China Scholarship Council and Macquarie
University for co-funding the PhD research.

xiii
 To my husband Dong Liu: thank you for your boundless love, enormous patience,
and constant encouragement. Thank you for helping create little master Yixuan Liu in
the middle of my PhD journey. I love you.
To my baby boy: your smile can easily chase away all my anxiety, even in the
toughest times.
致父母：亲爱的爸爸妈妈，你们的爱，伴我成长，给我力量。我爱你们。

xiv
Chapter 1 Introduction

This opening chapter introduces the literature and theoretical underpinnings of the thesis,
points out the limitations and gaps in previous studies, and articulates the aims and
questions of the research. It introduces the design of the project, with paticular attention
paid to the methodological triangulation of pen recording, eye tracking and voice
recording to study interpreting. Presented in a thesis by publication format, this thesis
consists of an introductory chapter providing the background and framework for the
study, followed by a series of individual research articles linked together into one
overall argument, and a concluding chapter providing the overarching conclusions of
the project as a whole.

1.1 Literature and theoretical underpinnings

1.1.1 Investigating the process and cognitive aspects of interpreting

Interpreting is an intriguing, challenging, and complex language processing task. Ever

since interpreting research became established as a field of study in its own right in the
mid-1970s (Pöchhacker, 2004, p. 81), there has been a strong interest in uncovering
what is happening in interpreters’ minds while they perform this extraordinary task.
Researchers with a background in psychology have attempted to shed light on how the
human mind processes language under severe stress and while engaging in heavy multi-
tasking by investigating the cognitive processes in interpreting (e.g., Barik, 1973;
Christoffels, 2004; Christoffels & De Groot, 2004, 2005; De Groot, 1997; Gerver,
1974a, 1974b, 1976; Goldman-Eisler, 1972; Köpke & Signorelli, 2012). Researchers
from within the field of interpreting, in turn, approached the topic from an inter-
disciplinary perspective that benefits from the theoretical and empirical findings in the
cognitive sciences (e.g., Lambert, 1988; Moser-Mercer, 1997; Seeber, 2011, 2013;
Shlesinger, 2000).
However, most of the process-oriented research approaching interpreting from a
cognitive perspective focuses on simultaneous interpreting (SI), while consecutive
interpreting (CI) is often neglected. CI was the first form of interpreting used at
international conferences and dominated the market in the first half of the 20th century.
It gradually gave way to SI, which was made possible by the development of electronic
equipment, in multilateral and multilingual conference settings. However, CI remains

1
the preferred mode in the context of “bilateral interactions with only two languages
involved and in settings where confidentiality, intimacy and directness of interaction are
given priority over time efficiency”, such as high-level diplomatic encounters, business
negotiations, ceremonial speeches and press conferences (Dam, 2010, p. 76). CI
remains an important component in most interpreter training programmes. Its
significance is manifested in the large quantities of master’s theses on the subject1. Even
in places where the market is largely dominated by SI, training in CI is believed to be a
good way of preparing students for SI (Gile, 2001). Furthermore, CI is frequently
introduced to language students as a way of reinforcing language skills (e.g. Henderson,
1976; Hill, 1979; Paneth, 1984).
Given the important role CI plays in the above contexts, there exists a considerable
limitation in the literature in that process-oriented cognitive investigations have rarely
been carried out on CI. CI is an interesting activity from both a cognitive and a
linguistic point of view. Similar to SI, it requires a high level of bilingual language
processing and challenges the interpreter’s cognitive system by requiring multi-tasking
under strict time constraints. But CI also introduces a new challenge: note-taking2. In
addition to listening to the source speech and producing a target speech, CI requires the
interpreter to perform the tasks of note-writing and note-reading. In Phase I of CI,
interpreters listen to and analyse the source speech, keep parts of the speech in their
working memory, and write down notes. In Phase II, interpreters read back their notes,
retrieve information from their working memory, and produce a target speech. Both
phases depend heavily on note-taking – this unique and distinctive feature of CI.

1.1.2 Note-taking in CI

Note-taking has been a topic of interest in interpreting research for over half a century
(see Chapter 2). The well-developed volume of literature on consecutive note-taking
started with a series of books and articles introducing various note-taking systems and
principles. They were published in different languages, each generating a profound
influence in its own country and some even reached beyond (e.g., Allioni, 1989; Becker,
1972; Gillies, 2005; Gran, 1982; Ilg, 1988; Kirchhoff, 1979; Matyssek, 1989; Rozan,
1956/2002). Recommendations were made on such skills as noting the idea and not the

1
Interested readers can find the theses reported in various issues of the Conference Interpreting Research
Information Network Bulletin (CIRIN Bulletin) at www.cirinandgile.com.
2
In this thesis, consecutive interpreting refers to long consecutive where systematic note-taking is used.

2
word, how to use symbols, how to use abbreviations, and how to note links, negations,
and emphasis.
With such well-developed note-taking systems, it would seem that once students
are made aware of the systems and practice accordingly, note-taking should pose few
problems. However, when it comes to teaching and learning note-taking skills, both
teachers and students find it challenging. Various studies (e.g., Alexieva, 1994; Gile,
1991) identified that note-taking diverted students’ attention and even led to a
degradation in interpreting performance.
Researchers who have approached the topic form cognitive and linguistic
perspectives (Kirchhoff, 1979; Kohn & Albl-Mikasa, 2002; Seleskovitch, 1975) found
that there was a concurrent storage of information in notes and in memory, and a
competition for cognitive resources between note-taking and other activities in the
interpreting process. This has motivated subsequent research to target more specific
note-taking features and to examine them empirically.
Some of the most important variables investigated are: the choice of form (e.g., Dai
& Xu, 2007; Dam, 2004a), the choice of language (e.g., Abuín González, 2012; Dai &
Xu, 2007; Dam, 2004b; Szabó, 2006), and the relation between note-taking and
interpreting performance (e.g., Cardoen, 2013; Dai & Xu, 2007; Dam, 2007; Dam,
Engberg, & Schjoldager, 2005). The choice of form refers to the choice between
language and symbol, and the choice between abbreviation and full word. The choice of
language refers to the choice between source and target language, and the choice
between native and non-native language. Despite a couple of general trends, such as a
preference for language over symbol and a source language dominance in the notes
taken by student interpreters, the studies have reported inconsistent findings (see
Chapter 2).
The inconsistencies are potentially related to some shared limitations of these
studies. Of the limited empirical data that have been collected, a large portion has been
collected from students (varying in their stages of study and maturity of competence),
making the findings difficult to generalise (Gile, 2009, p. 179). Furthermore, many
studies only investigate one interpreting direction so the results are hard to compare.
More importantly, most of the studies are product oriented, which means that they only
look at the final product of note-taking (the notes produced), without an in-depth
analysis of the interpreting process. Last but not least, the research efforts are somewhat
scattered, without an overarching framework to pull them together.

3
 A possible solution is to examine the topic from a cognitive perspective. As Gile
(2009, p. 178) points out, “note-taking is an area in which the concept of processing
capacity can be useful.” Viewed from a cognitive perspective, all discussions on note-
taking could boil down to one fundamental question: how to reduce the cognitive load
of note-taking while maintaining the efficiency of notes. If cognitive load can be
measured while interpreters perform note-taking and CI, some fundamental principles
underlying the note-taking choices might be unveiled. It is possible, for example, that
the different observations reported by previous studies might not be controversies, but
rather converging evidence in proving that interpreters make choices according to their
own characteristics to reduce cognitive load.
The current research attempts to revisit the topic of note-taking and CI and address
some of the existing limitations by (1) using professional interpreters as participants; (2)
investigating both directions of interpreting; (3) combining product analysis with the
process research methods of pen recording, eye tracking and voice recording; and (4)
performing analyses and illustrations under a common cognitive framework and
focusing on cognitive load in the process.

1.1.3 The construct of cognitive load in interpreting and its measurement

The construct of cognitive load was not new when it was introduced to the field of
interpreting in the 1980s. Similar constructs had already been investigated, with two of
the most relevant ones being mental workload in human factors research and cognitive
load in Cognitive Load Theory. Human factors research centres on how humans
accomplish tasks in the context of human-machine system operation and how different
variables affect that accomplishment (Meister, 1989). Cognitive Load Theory is a
theory of learning, focusing on how instruction affects the load on learners’ cognitive
systems (Sweller, Ayres, & Kalyuga, 2011). The two fields define their central
constructs using different terms (see O'Donnell and Eggemeier (1986) for mental
workload and Paas and Van Merriënboer (1994) for cognitive load), but both constructs
capture the interactions between a task with specific characteristics and a human with
limited cognitive capacity. In that sense, both constructs are relevant and both fields
serve as good references for the definition and illustration of cognitive load in
interpreting. Borrowing from these two fields, this research therefore builds a
theoretical and methodological framework for defining and measuring cognitive load in
the field of interpreting (see Chapter 3).

4
 Cognitive load in interpreting has been operationalised as a multi-dimensional
construct depending on the interactions between two groups of variables: the
task/environmental characteristics and the interpreter characteristics. Since cognitive
load is a theoretical construct, it cannot be observed or measured directly. Its
measurement relies on an arrangement of surrogates that are indicative of cognitive
load. Moreover, cognitive load is a multi-dimensional construct, so any single measure
cannot provide a comprehensive picture. Due to such challenges, only a few pioneering
studies have investigated cognitive load in interpreting (see Chapter 3). These studies
usually apply only one technique to examine the cognitive load, and are almost
exclusively on SI. This research directly contributes to that effort by using a
combination of methods, including pen recording, eye tracking and voice recording, to
explore the cognitive load in CI.

1.2 Research scope, research purpose and research questions

The scope of this process-oriented cognitive research on interpreting is limited in four

respects. First, the study focuses only on CI – a spoken-language interpreting in the
consecutive mode – and not on other interpreting modalities. This is mainly due to a
lack of research on the processing and cognitive aspects of CI and the fact that the
author was only trained and practiced in spoken-language (as opposed to sign-language)
interpreting. Second, the research only involves the language pair of Chinese and
English. This is because the author only has the linguistic prerequisites to study and
analyse this particular language pair at the required level. Third, the project only
recruited professional interpreters to participate in the experiment. This is to address the
problem that most previous studies on note-taking and CI only involved student
interpreters. Despite its necessarily limited scope, this study is expected to make
methodological contributions to and generate meaningful findings for future research
involving other interpreting modalities, language pairs, and interpreter types.
The research has three main purposes. First, it aims to make a methodological
contribution to bilingual language processing research (especially translation and
interpreting research) by triangulating the methods of pen recording, eye tracking, and
voice recording. Second, it seeks insights into the interpreter’s black box by examining
the cognitive processing and cognitive load in interpreting. Third, through observing
and analysing how professional interpreters take notes in CI, the study searches for

5
empirically-based recommendations for interpreter education and continued
professional development.
The PhD study sets out to answer three main research questions (RQs)
corresponding to the key variables in consecutive note-taking identified in the literature.
RQ 1: What are the preferred note-taking choices of professional interpreters in
terms of form and language?
This RQ subsumes two sub-questions, including:
RQ 1.1 What do interpreters prefer when choosing the form of note-taking:
language or symbol; abbreviation or full word?
RQ 1.2: What do interpreters prefer when choosing the language of note-taking:
source or target language; native or non-native language?
After identifying the patterns of note-taking choices, a second step is to examine the
relation between these note-taking choices and cognitive load in CI.
RQ 2: What is the relationship between note-taking choices and cognitive load in
CI?
More specifically, the study is interested in how the choices made by interpreters
during Phase I of CI affect the level of cognitive load in Phase II. So RQ2 also consists
of two sub-questions:
RQ 2.1: What is the relationship between the note-taking choices and cognitive
load in Phase I (the listening and note-writing phase) of CI?
RQ 2.2: What is the relationship between the note-taking choices and cognitive
load in Phase II (the note-reading and production phase) of CI?
The third research question concerns a central issue in interpreting research, namely
the quality of performance, and its relationship with note-taking. The aim is to see
whether specific note-taking choices help or hinder the interpreting performance.
RQ 3: What is the relationship between note-taking and interpreting performance?

1.3 Research design

This PhD research used an exploratory design to gain insights into the process of note-
taking and CI. In particular, it triangulated the methods of pen recording, eye tracking
and voice recording to collect data on cognitive processing and cognitive load in
interpreting. To make the data more generalizable, professional interpreters rather than
student interpreters (whose interpreting competence are greatly varied and not yet

6
mature) were recruited. Two CI tasks covering both directions of interpreting (between
Chinese and English) were involved to account for both the source/target language
status and the native/non-native language status. The order of the two CI tasks was
randomised to eliminate the impact brought by task order. A retrospection was designed
to collect additional qualitative data following the completion of the two tasks. The
note-writing process was recorded via pen recording; the note-reading process was
recorded via eye tracking; the interpreting process and retrospection were recorded via
voice recording.

1.3.1 Pen recording

The apparatus used for pen recording was the Cintiq 13HD (a 13-inch LCD tablet with a
resolution set at 1366×768 pixels) and the Wacom Pro Pen. The system was chosen
because it targets graphic designers who have very high requirements in terms of the
precise control of the pen on the tablet surface. It is ergonomically designed to mimic
natural writing and painting. Another reason for choosing this system is because it is
compatible with the Eye and Pen software3, one of the core software products powering
the experiment. The software piloted a laptop computer which was linked to the pen
recording apparatus. The software carried out three tasks: controlling the experiment,
collecting the pen data, and processing the pen data.
Controlling the experiment. The experiment and its procedures were programmed
into the software, which then controlled the progress of the experiment and interacted
with the participant. For example, in Phase I of CI, when finishing one page of note-
taking, the participant could use the pen to click on a button displayed on the tablet
screen called “New Page” (Figure 1.1) and the software would create a new blank page
for note-taking. The participant could use as many pages as needed. When the listening
and note-writing phase was finished, the participant only needed to click a button called
“Begin Interpreting” (Figure 1.1) and the software would automatically turn to the first
page of notes written by the participant. Then the participant could read back the notes
and produce a target speech. In this phase, new buttons such as “Turn Page” (which
turns to the next page of written notes) and “Next part” (which plays the next segment
of the source speech) would appear on the screen and the participant could interact with
the software to navigate through the pages of written notes. The tablet screen would
only react to the tip of the digital pen, so the participant could write as naturally as
3
More detailed information about the software can be found on http://eyeandpen.net/en/.

7
possible and did not need to worry about triggering any buttons by touching the screen
with their hands.

Figure 1.1 A screenshot of the tablet in the recording mode

Collecting the pen data. The software collected the spatial and temporal data about
the pen as it moved across the tablet surface. For example, data was recorded for each
pen stroke in terms of the distance (how far the pen travelled across the surface),
duration (for how long the pen was in touch with the tablet), and speed (how fast the
pen was moving). Spatial data was reported in centimetres and temporal data was
reported in milliseconds. The software also kept a session log for each trial,
documenting the time every action took place during the recording (e.g., the source
speech segment started playing, the participant started writing, etc.). This function was
crucial for the calculation of an important cognitive load indicator, the ear-pen span,
which is the time span between the moment a speech unit is heard and the moment it is
written down in notes (see Chapter 4 and Chapter 5).
Processing the pen data. The software has many functions for displaying and
analysing the recorded pen data (Figure 1.2 is a screenshot of the software in the
analysis mode). The most useful function for this study is the “Word separation” tool,
which semi-automatically separates the written texts into words (in this study’s case,
note units). Although manual work was required to correct the separations, this function
allowed very accurate data to be reported for each individual note unit (e.g., start and
end time, duration, distance, speed, etc.). Labels could be created for each note unit so
that qualitative data could be added to each note and exported for further analysis. For
example, for note unit no.13 (see bottom left of Figure 1.2), texts 1 to 6 documented the
form and language of the note unit as well as its content, meaning, and corresponding
8
source speech unit. The labels indicated that this note unit was language (“L” in Text 1),
in English (“E” in Text 2), and an abbreviation (“A” in Text 3). It contained three letters
“svs” (Text 4), meaning “services” (Text 5) and corresponding to the word “services”
(Text 6) in the source speech. In this way, the exported file contained both quantitative
and qualitative data (Figure 1.3).

Figure 1.2 A sample screenshot of the Eye and Pen software in the analysis mode

Figure 1.3 A sample data output of pen recording

9
1.3.2 Eye tracking

There were a few prerequisites for selecting the type of eye tracker to be used in the
study. First, the eye tracker needed to allow the interpreter to speak freely, thus
eliminating the use of eye trackers that require chin rests. Second, the eye tracker
needed to be usable in a handwriting situation. In particular, the eye camera(s) could not
be masked by the participant’s forearms in movement. Head-mounted eye trackers
could meet the first two requirements. Third, for the comfort of the participant and
ecological validity of the experiment, a light-weight eye tracker that could be attached
to the participant easily was preferable.
The eye tracker used in this study was the SensoMotoric Instruments (SMI) Eye
Tracking Glasses 2 (ETG2). It is a light-weight (47 g), head-mounted eye tracker in the
shape of a pair of glasses. The eye tracker uses dark pupil tracking. It has a tracking
accuracy of .5°over all distances and a sampling rate of 60 Hz. The eye tracker has a
built-in high-definition camera for scene recording. This camera recorded both the
video and the audio during the entire note-taking and interpreting process. The SMI
software iView ETG and BeGaze were used with default settings for eye data recording
and analysis respectively. The experiment took place in a sound-proof studio with
constant artificial illumination to avoid any distractions or disruption to the recording of
eye data.

1.3.3 Voice recording

Voice recording was used for several different purposes in this study. First and foremost,
the interpreting performance was recorded. The audio recordings were later transcribed
and provided to a group of raters for evaluation (see Chapter 4 and Chapter 5). This
generated performance scores used for exploring the relationship between note-taking
and interpreting performance. Second, voice recording was used during cued
retrospection. Immediately after the interpreting tasks, the participants were provided
with their notes for cued retrospection. They were asked to provide as much information
as they could remember about the note-taking process, including but not limited to:
what each note unit was; what it stood for; whether it was symbol or language, and if
language, whether it was abbreviation or full word, Chinese or English. This is an
important step because note-taking in CI is highly individualised, and the handwriting
of interpreters could sometimes be difficult for others to decipher. Third, the source
speech audio files were used together with the session logs kept by the Eye and Pen

10
software to calculate the ear-pen span, an important indicator of cognitive load used in
this study (see Chapter 4 and Chapter 5).

1.3.4 Participants, tasks and procedures

A total of 26 professional interpreters were recruited using purposive sampling. Ethics

approval was obtained for conducting the research (see Appendix B). Table 1.1
summarises the criteria used to recruit the participants. Demographic questionnaire
(Appendix C) results showed that most of the participants had a postgraduate degree in
interpreting (65%); some had an interpreting diploma (15%); some attended an
intensive interpreting training course (15%); and one was self-trained (4%). The age of
the participants averaged at 36.4 years; their working experience (years of working as
either full-time or part-time interpreters) averaged at 7.4 years. The participants worked
most frequently in Australia (with only two exceptions who worked more frequently in
China).

Table 1.1 Criteria for recruiting participants

Professional Interpreter level accreditation from Australia’s National

Accreditation
Accreditation Authority for Translators and Interpreters (NAATI)
Working language Mandarin Chinese (native language) and English (non-native language)
A minimum of two years of experience working as part-time or full-time
interpreters;
Experience If working as a part-time interpreter, the other jobs have a bilingual feature (e.g.,
interpreter trainer);
Experience in working in the consecutive mode

Two CI tasks (see Appendix E) were carefully created through a series of

procedures to control for variance (see Chapter 4). The experiment took place in four
main procedures (see Chapter 6 for a detailed procedure): practice, task performance,
retrospection and post-experiment questionnaire. The practice session was designed to
familiarise the participants with the experimental procedures and the apparatus,
especially the digital pen and the eye tracker. The task performance session involved
two CI tasks, the order of which was randomised. Rest was allowed between tasks if
needed. The retrospection session was cued by the written notes and participants were
instructed to recall whatever they could remember about the note-taking process. This
was mainly designed to help the researcher accurately identify the note units and to
collect additional qualitative data for the interpretation of the results. The questionnaire
was designed to collect such information as the participants’ familiarity with the task

11
topics, how they felt about using the digital pen and the eye tracker, and other feedback
about the experiment.

1.3.5 Data and analysis

The main sources of data collected in this study and their corresponding RQs and
chapters are summarised in Table 1.2. Wherever applicable, the data were standardised
using a note unit as the unit of analysis. For example, if the number of Chinese notes
written by a participant is n, then the ear-pen span of Chinese notes (𝐸𝑃𝑆𝐶 ) of that
participant is calculated as:
1
𝐸𝑃𝑆𝐶 = 𝑛 × (𝐸𝑃𝑆1 + 𝐸𝑃𝑆2 + ⋯ + 𝐸𝑃𝑆𝑛 )
Paired-samples t-tests were conducted to compare between the note-taking choices
in different forms (language vs. symbol; abbreviation vs. full word) and languages
(Chinese vs. English). The Pearson’s correlation was used to explore the relation
between note-taking and interpreting performance. All statistical analyses were
performed by running the IBM SPSS Statistics 22. Two-tailed p values less than 0.05
were considered to be statistically significant. Cohen's d (the difference between the
means divided by the pooled standard deviation) was used to indicate the effect sizes,
which were classified as small (d = 0.2), medium (d = 0.5), and large (d = 0.8).

Table 1.2 Data collected in the study

Source Data RQ(s) Chapters

All written note units (including the form,
language, content, meaning, and corresponding RQ 1.1 and 1.2
source speech unit) Chapter 4
Pen recording
and 5
The distance, duration, speed and ear-pen span
RQ 2.1
of all note units
Eye movement measures such as regression
rate, first fixation duration, first-pass dwell time,
Eye tracking second-pass dwell time, total dwell time, RQ 2.2 Chapter 6
number of fixations, number of revisits, average
fixation duration, and skip rate
Interpreting performance and audio of Chapter 4, 5
Voice recording RQ 3
retrospection and 6

1.4 Outline of the thesis

This thesis is written in a “thesis by publication” format, an encouraged and preferred

approach at Macquarie University for Higher Degree Research candidates. According to
the university’s guideline (see Appendix A), a thesis by publication may include

12
relevant papers which have been published, accepted, submitted or prepared for
publication during the candidature. The papers need to be sufficient to support the
important findings from the research and presented in a logical and coherent way. Most
theses by publication have between 2 and 8 papers, each forming a chapter of the thesis.
This thesis consists of seven chapters. Chapter 1 is the introductory chapter and
Chapter 7 is the concluding chapter. Chapters 2 to 6 are self-contained journal articles.
Chapters 2, 3, and 4 have been published in peer-reviewed journals; Chapter 5 is
currently under peer review and Chapter 6 has been submitted to a peer-reviewed
journal. Relevant publication details are specified in a footnote at the beginning of each
chapter. The links between the article-based chapters are illustrated in Figure 1.4.

Figure 1.4 Linkages between the article-based chapters in this thesis by publication
13
 Chapter 1 gives an overview of the literature and theoretical underpinnings of the
study, states the scope, purpose and questions of research, and introduces the design of
the project.
Chapter 2 is a critical review of the studies that have been carried out on
consecutive note-taking. It sets the literature background for the thesis and identifies the
most important variables on the topic. Two major limitations of previous studies are
identified as a lack of process research and a lack of empirical data. Cognitive load is
established as a promising avenue for future research because it allows us to approach
the process of note-taking and interpreting while contributing ample empirical data.
Chapter 3 sets the theoretical and methodological foundation for the thesis. It
defines and operationalises the construct of cognitive load in interpreting and its
measurement. Some techniques that could be used to study cognitive processing and
cognitive load in CI are also introduced, with special emphasis put on the combination
of pen recording and eye tracking.
Chapter 4 reports the data collected in a pilot study of the PhD project. Through a
carefully-selected sample of five professional interpreters, this pilot study is able to find
some empirical evidence concerning the important variables identified in Chapter 2. Pen
recording is proven to be a powerful method to tap into the process of note-taking and
interpreting, and the collected data serve as useful indicators of cognitive load. This
article paves the way for the next stage of the project where a main study is conducted,
combining pen recording and eye tracking and involving a larger group of professional
interpreters. Findings of this study are instrumental in determining the hypotheses for
the next stage of the study.
Chapter 5 and Chapter 6 report the data collected in the main study of the PhD
project. Chapter 5 reports the pen and voice recording data collected from professional
interpreters while they perform CI with notes, with special attention paid to Phase I
where interpreters listen to the source speech and take notes. Most of the findings in the
pilot study have been successfully replicated in the main study. Pedagogical
recommendations are provided on the basis of the empirical findings.
Chapter 6 reports the eye tracking data collected from the main study. The focus is
on Phase II, in which interpreters read back their notes and produce a target speech.
This study is among the first to visualise note-reading, showing that it proceeds in a
non-linear manner and requires a high level of cognitive costs. The data provide
important indicators of cognitive processing in note-reading and interpreting. A

14
combined analysis of findings from the fourth and fifth article reveals that there is a
trade-off between the cognitive costs of Phase I and Phase II in CI.
Chapter 7 summarises the main findings of the research, explains their implications,
examines the strengths and limitations of the research, and discusses the possibilities for
future work.

1.5 References

Abuín González, M. (2012). The language of consecutive interpreters’ notes:

Differences across levels of expertise. Interpreting, 14(1), 55–72.
doi:10.1075/intp.14.1.03abu
Alexieva, B. (1994). On teaching note-taking in consecutive interpreting. In C. Dollerup
& A. Lindegaard (Eds.), Teaching translation and interpreting 2: Insights, aims,
visions (pp. 199–206). Amsterdam: John Benjamins.
Allioni, S. (1989). Towards a grammar of consecutive interpretation. In L. Gran & J. M.
Dodds (Eds.), The theoretical and practical aspects of teaching conference
interpretation (pp. 191–197). Udine: Campanotto.
Barik, H. C. (1973). Simultaneous interpretation: Temporal and quantitative data.
Language and Speech, 16(3), 237–270. doi:10.1177/002383097301600307
Becker, W. (1972). Notizentechnik [Note-taking]. Germersheim: BBK.
Cardoen, H. (2013). The effect of note-taking on target-text fluency. In G. González
Núñez, Y. Khaled, & T. Voinova (Eds.), Emerging research in translation
studies: Selected papers of the CETRA Research Summer School 2012 (pp. 1–
22). Leuven: CETRA.
Christoffels, I. K. (2004). Cognitive studies in simultaneous interpreting (Unpublished
doctoral thesis), University of Amsterdam, Amsterdam.
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
Christoffels, I. K., & De Groot, A. M. B. (2005). Simultaneous interpreting: A cognitive
perspective. In J. F. Kroll & A. M. B. De Groot (Eds.), Handbook of
bilingualism: Psycholinguistic approaches (pp. 454–479). New York: Oxford
University Press.

15
Dai, W., & Xu, H. (2007). 汉英交替传译过程中译员笔记特征实证研究——以职业
受训译员和非职业译员为例 [An empirical study of the features of interpreters'
notes in Chinese-English consecutive interpreting: The examples of
professionally trained and unprofessional interpreters]. Foreign Language
Teaching and Research, 39(2), 136–144.
Dam, H. V. (2004a). Interpreters' notes: On the choice of form and language. In G.
Hansen, K. Malmkjær, & D. Gile (Eds.), Claims, changes and challenges in
Translation Studies (pp. 251–261). Amsterdam: John Benjamins.
Dam, H. V. (2004b). Interpreters' notes: On the choice of language. Interpreting, 6(1),
3–17. doi:10.1075/intp.6.1.03dam
Dam, H. V. (2007). What makes interpreters’ notes efficient?: Features of (non-
)efficiency in interpreter's notes for consecutive. In Y. Gambier, M. Shlesinger,
& R. Stolze (Eds.), Doubts and directions in Translation Studies: Selected
contributions from the EST Congress, Lisbon 2004 (pp. 183–197). Amsterdam:
John Benjamins.
Dam, H. V. (2010). Consecutive interpreting. In Y. Gambier & L. van Doorslaer (Eds.),
Handbook of translation studies: Volume 1 (pp. 75–79). Amsterdam: John
Benjamins.
Dam, H. V., Engberg, J., & Schjoldager, A. (2005). Modelling semantic networks on
source and target texts in consecutive interpreting: A contribution to the study of
interpreters’ notes. In H. V. Dam, J. Engberg, & H. Gerzymisch-Arbogast (Eds.),
Knowledge systems and translation (Vol. 7, pp. 227–254). Berlin: Mouton de
Gruyter.
De Groot, A. M. B. (1997). The cognitive study of translation and interpretation: Three
approaches. In J. H. Danks, G. M. Shreve, S. B. Fountain, & M. K. McBeath
(Eds.), Cognitive processes in translation and interpreting (pp. 25–56). London
and New Delhi: Thousand Oaks.
Gerver, D. (1974a). The effects of noise on the performance of simultaneous
interpreters: Accuracy of performance. Acta Psychologica, 38(3), 159–167.
Gerver, D. (1974b). Simultaneous listening and speaking and retention of prose. Q J
Exp Psychol (Hove), 26(3), 337–341.

16
Gerver, D. (1976). Empirical studies of simultaneous interpretation: A review and a
model. In R. B. W. Anderson & R. W. Brislin (Eds.), Translation: Applications
and research (pp. 165–207). New York: Gardner Press.
Gile, D. (1991). Prise de notes et attention en début d’apprentissage de l’interprétation
consécutive–une expérience– démonstration de sensibilisation [Note-taking and
attention at the beginning of consecutive interpreting learning –an experience–
demonstration of awareness]. Meta, 36(2–3), 431–439.
Gile, D. (2001). The role of consecutive in interpreter training: A cognitive view.
Retrieved 26 November 2016, from AIIC http://aiic.net/page/377/
Gile, D. (2009). Basic concepts and models for interpreter and translator training
(revised edition). Amsterdam: John Benjamins.
Gillies, A. (2005). Note-taking for consecutive interpreting: A short course. Manchester
and Kinderhook: St. Jerome.
Goldman-Eisler, F. (1972). Segmentation of input in simultaneous translation. Journal
of Psycholinguistic Research, 1(2), 127–140.
Gran, L. (1982). L'annotazione grafica nell'interpretazione consecutiva [Note-taking in
consecutive interpreting]. Trieste: Universitàdegli Studi di Trieste.
Henderson, J. A. (1976). Note-taking for consecutive interpreting. Babel, 22(3), 107-
116.
Hill, J. K. (1979). Consecutive interpreting in advanced language work. Meta, 24(4),
442-450. doi:10.7202/003071ar
Ilg, G. (1988). La prise de notes en interprétation consécutive. Une orientation générale
[Note-taking in consecutive interpretation: A general approach]. Parallèles, 9,
9–13.
Kirchhoff, H. (1979). Die notationssprache als hilfsmittel des konferenzdolmetschers im
konsekutivvorgang [The language of note-taking as a tool for the conference
interpreter in consecutive interpreting]. In W. Mair & E. Sallager (Eds.),
Sprachtheorie und sprachpraxis [Language theory and language practice] (pp.
121-133). Tübingen: Gunter Narr.
Kohn, K., & Albl-Mikasa, M. (2002). Note-taking in consecutive interpreting: On the
reconstruction of an individualised language. Linguistica Antverpiensia, 1, 257-
272.

17
Köpke, B., & Signorelli, T. M. (2012). Methodological aspects of working memory
assessment in simultaneous interpreters. International Journal of Bilingualism,
16(2), 183-197. doi:10.1177/1367006911402981
Lambert, S. (1988). Information processing among conference interpreters: A test of the
depth-of-processing hypothesis. Meta, 33(3), 377-387. doi:10.7202/003380ar
Matyssek, H. (1989). Handbuch der notizentechnik für dolmetscher [Handbook of note-
taking for interpreters]. Heidelberg: Julius Groos.
Moser-Mercer, B. (1997). Beyond curiosity: Can interpreting research meet the
challenge? In J. H. Danks, G. M. Shreve, S. B. Fountain, & M. K. McBeath
(Eds.), Cognitive processes in translation and interpreting (pp. 176–195).
London and New Delhi: Thousand Oaks.
O'Donnell, C. R. D., & Eggemeier, F. T. (1986). Workload assessment methodology. In
K. R. Boff, L. Kaufman, & J. P. Thomas (Eds.), Handbook of perception and
human performance, vol. II: Cognitive processes and performance (pp. 42-41–
42-49). New York: Wiley-Interscience.
Paas, F. G. W. C., & Van Merriënboer, J. J. G. (1994). Instructional control of cognitive
load in the training of complex cognitive tasks. Educational Psychology Review,
6(4), 351–371. doi:10.1007/BF02213420
Paneth, E. (1984). Training in note-taking (For interpreting). In W. Wilss & G. Thome
(Eds.), Die theorie des übersetzens und ihr aufschlußwert für die übersetzungs-
und dolmetschdidaktik [The theory of translation and its value for translating-
interpreting dichotomy] (pp. 326–332). Tübingen: Gunter Narr.
Pöchhacker, F. (2004). Introducing interpreting studies. London and New York:
Routledge.
Rozan, J.-F. o. (1956/2002). Note-taking in consecutive interpreting (A. Gillies, Trans.).
Cracow: Tertium Society for the Promotion of Language Studies.
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
Seleskovitch, D. (1975). Langage, langues et mémoire: étude de la prise de notes en
interprétation consécutive [Speech, language, and memory: A study of note-
taking in consecutive interpreting]. Paris: Minard Lettres Modernes.

18
Shlesinger, M. (2000). Strategic allocation of working memory and other attentional
resources in simultaneous interpreting (Doctoral dissertation), Bar-Ilan
University, Israel.
Sweller, J., Ayres, P., & Kalyuga, S. (2011). Cognitive Load Theory. New York:
Springer.
Szabó, C. (2006). Language choice in note-taking for consecutive interpreting.
Interpreting, 8(2), 129–147. doi:10.1075/intp.8.2.02sza

19
An introductory note to Chapter 2

As has been elaborated in Chapter 1, note-taking as a unique characteristic of CI has

attracted the interest of researchers for over half a century. The earlier prescriptive
literature has established an array of note-taking systems, principles of which are still
useful today. The descriptive literature has inspected specific variables concerning note-
taking and CI, contributing valuable empirical data for a deeper understanding of the
topic. This rich dichotomy of literature contains valuable information on how note-
taking research has evolved over the years and what awaits future studies.
Chapter 2 therefore consists of a critical literature review for the thesis. It combs
through the prominent and influential studies in the field, with an aim to identify the
gaps and limitations in previous research and to inform how investigations should be
carried out in this study. It examines several issues that are of central importance to the
PhD thesis: the major topics in previous note-taking studies, the research methods that
have been used, and the findings, controversies, and limitations of these studies. The
chapter also pinpoints the study of cognitive processing and cognitive load as a
productive avenue for future research.

20
Chapter 2 Note-taking in consecutive interpreting: A review with
special focus on Chinese and English literature1

Abstract: Publications on note-taking in consecutive interpreting are reviewed, with

special attention being awarded to literature written in Chinese and English. The review
identifies two main streams of note-taking literature, a prescriptive stream and a
descriptive stream. Prescriptive publications are concerned with the question “How
should notes be taken?” They introduce the established note-taking systems and
principles, and discuss how to teach them to students. The second stream, consisting of
descriptive studies, tackles the question “How are notes taken?” The studies strive to
approach the topic with scientific rigor by collecting data from simulated interpreting
practices. Fruitful results have been created, but there are several limitations. The
prescriptive stream lacks systematic empirical research to support the proposals. The
descriptive stream is mostly product-oriented, lacking process research, and no study
has designed true experiments to explain the causal relationships behind the observed
phenomena. Cognitive load offers a promising perspective to approach the process of
note-taking while contributing ample empirical data. It is therefore worthwhile to
investigate cognitive load during note-taking in consecutive interpreting.
Keywords: Note-taking, consecutive interpreting, review, cognitive load.

2.1 Introduction

Note-taking is a distinctive feature of consecutive interpreting (CI), in particular ‘classic’

consecutive where systematic note-taking is used (Pöchhacker, 2004, p. 19), and
scholars’ sustained interest in the subject has generated a considerable volume of
literature. This review attempts to explore how note-taking literature has evolved for
over half a century, and what awaits future research. It is part of a larger project that
looks into the process of CI and note-taking. The aim is to identify the most productive
avenue of investigation by combing through the prominent and influential studies in the
field. The review is interested in the following questions: (1) what are the major topics
in note-taking studies, (2) what research methods have been used, (3) what are the

1
This chapter is a published journal article: Chen, S. (2016). Note-taking in consecutive interpreting: A
review with special focus on Chinese-English literature. The Journal of Specialised Translation, 26, 151–
171.

21
findings, controversies, and limitations, and (4) what could be a promising avenue for
future research.
The review is written from the perspective of an interpreting researcher and
practitioner whose interest lies in interpreting between the Chinese and English
language pair, and as a result, the review pays special attention to publications written
in Chinese and English2. The selection of texts to be reviewed followed a number of
successive steps. The inclusion criteria might seem overly restrictive, but they were
designed to identify the most relevant and representative literature, not to
comprehensively survey all the studies available.
Firstly, the key databases relevant to this review were identified: the Conference
Interpreting Research Information Network Bulletin (CIRIN Bulletin), Translation
Studies Abstracts Online (TSA Online), and the Chinese Social Sciences Citation Index
(CSSCI). The reason why an additional China-based database is necessary for the
search of relevant literature is that, according to Zhou, Su, and Leydesdorff (2010, p.
1362), “the Chinese and international communication systems in the social sciences are
almost completely uncoupled in terms of the coverage in the databases.”
Secondly, the selected databases were searched for relevant publications. The
CIRIN Bulletin did not provide keywords of its collections, so a keyword-based search
was not possible. The database was searched for titles that contained any of the
following words: ‘note,’ ‘notation’ and ‘note-taking.’ TSA Online and the CSSCI were
searched by looking for titles and keywords that contained ‘consecutive interpreting’3
and any of the above mentioned words.
Thirdly, the references of the retrieved items were scanned for relevant publications.
This was repeated until no more relevant publications came to light.
A further step was to finalise the list using the following criteria: (1) The
publication addressed note-taking in CI as a subject in its own right 4 . (2) The
publication was written in Chinese or English. For publications written in other
languages, only those that have been most referenced were included in the review. (3)

2
It has to be admitted that the author does not have the linguistic prerequisites to read all that she has
collected, and that what she has collected represents a comprehensive yet not complete list of the relevant
publications.
3
This is not necessary for the CIRIN Bulletin because the database collects only interpreting studies.
4
Many authors include note-taking as a part of their discussions on interpreter training or education.
These publications are not included in this review.

22
The study was peer-reviewed (published as a book or book chapter; journal article; in
refereed conference proceedings)5.
The selected publications are grouped according to topic: note-taking systems and
principles, didactics, cognitive and linguistic aspects of note-taking, choice of form and
language, and the relationship between note-taking and interpreting quality. The
boundaries between the groups are by no means clear-cut. The classification has been
made on the basis of the most prominent focus of each. Two main streams of literature
have been identified: a prescriptive stream and a descriptive stream. Prescriptive
publications on note-taking are concerned with the question “How should notes be
taken?” Usually starting from the authors’ experience in the profession and/or in
teaching, this first stream aims at introducing established note-taking systems and
principles, and discusses how to teach them. The second stream, consists of descriptive
studies, tackles the question “How are notes taken?” It strives to approach the topic with
scientific rigor by collecting data from simulated interpreting practices, using either
students or professional interpreters as subjects. As will be made clear in the review,
there is a shift in note-taking literature from prescriptive to descriptive along this
continuum.

2.2 Note-taking systems and principles: a prescriptive starting point

Among the first publications on note-taking are a number of books and articles that
introduce the well-known note-taking systems and principles. These publications adopt
a prescriptive stance, and propose the ways that notes should be taken. More often than
not, the prescriptions made are based on the author’s experience as professional
interpreters and/or teachers. For example, Rozan (2002, p. 11) mentioned in the
introduction of his book, that “This system is the product of 10 years as a practising
interpreter and 4 years teaching the profession.” Little if any empirical data has been
collected in this stream of literature. Nevertheless, the contributions are obvious: they
have offered important experience and knowledge of the profession, and are therefore
fundamental to note-taking research.

5
A number of relevant studies are in the form of unpublished master’s and doctoral theses, but the scope
of this article does not allow the inclusion of those studies. Interested readers can find the studies reported
in various issues of the CIRIN Bulletin at www.cirinandgile.com.

23
 The earliest note-taking system was proposed by Rozan in 1956. The influence of
the work is far-reaching. When it was translated into English and Polish in 2002, the
editors commented that “it would be hard to find an interpreter in Western Europe
whose note-taking style owes nothing to Rozan” (Rozan, 2002, p. 7). Following Rozan,
many books and articles on note-taking systems and principles were published in
different languages, each generating a profound influence in its own country and some
even reached beyond. Some outstanding examples would be Allioni (1989), Becker
(1972), Gillies (2005), Gran (1982), Ilg (1988), Kirchhoff (1979), and Matyssek (1989).
When new systems are introduced, authors usually build on the wealth of the
previously-established ones, adapting the existing rules as they see fit. To avoid
repetition, this part of the review starts from Rozan’s system, uses it as a reference, and
discusses some of the best-known principles of note-taking. These principles can be
found in most of the existing note-taking systems, and different authors have
contributed to them.
The first principle is at the core of almost all note-taking systems: noting the idea
and not the word. It has been variously expressed as “comprehension” before note-
taking (e.g. Deng, 1991, p. 285; Jia, 1995, pp. 77-78) or “analysis” before note-taking
(e.g. Alexieva, 1994, p. 206; Chuang, 2008, p. 95; Han, 2002, pp. 25-26; Mu & Lei,
1998, pp. 82-83). This principle emphasises that what is important in note-taking is the
idea or “concept” (Gillies, 2005, p. 53) that lies under the actual words used. When
taking notes, interpreters should arrive at the underlying meaning through analysis and
comprehension of the source speech.
Rozan’s second principle consists of the rules of abbreviation. The most important
rule, according to Rozan Rozan (2002, p. 16), is that long words (more than 4 to 5
letters) should not be written in full. It is generally suggested that the first and last
letters should be used to abbreviate the word, with the latter written as superscript
(Gillies, 2005, p. 130; Matyssek, 1989, p. 115; Rozan, 2002, p. 17; Schweda-Nicholson,
1993, p. 200). Using the first letters to abbreviate is also recommended (Becker 1972:
30). Other rules of abbreviation include: using abbreviations to indicate gender, tense
and register (Rozan, 2002, pp. 17-18); borrowing commonly known abbreviations from
daily life (Matyssek, 1989, p. 113; Wu, 2008, p. 8); using international suffixes such as
“-tion” (Gillies, 2005, p. 130; Matyssek, 1989, p. 117); and using phonetic spelling and
misspelling (Gillies, 2005, p. 131; Han, 2002, p. 26). It is common to put abbreviations
at a prominent place when discussing note-taking in CI between European languages,

24
but the case with Chinese is different. Some of the rules which are largely based on
European languages are difficult for native speakers of Chinese, and some rules are
even rendered useless because of the differences in languages (M. Liu, 2008, p. 65f).
That being said, part of the rules are still applicable, especially when the task is
interpreting from Chinese into English. No matter how abbreviations are used, they
have to meet certain conditions: they should be unambiguous (Henderson, 1976, p. 110;
Matyssek, 1989, p. 115), easy to write (Alexieva, 1994, p. 204), and should not sacrifice
accuracy (Schweda-Nicholson, 1990, p. 140).
The third principle concerns the noting of links. Links are believed to be
indispensable in note-taking (Matyssek, 1989, p. 53; Wu, 2008, p. 17) because “an idea
can be distorted completely if its relation to the previous idea is not clearly indicated”
(Rozan, 2002, p. 18). Many authors (e.g. Gillies, 2005; Matyssek, 1989; Wu, 2008)
have identified the main types of linking words and expressions, including additive,
adversative, and causal (cause, purpose and consequence) links, and recommended the
use of only one abbreviation, short word, or symbol to represent the whole family.
Gillies (2005, pp. 147, 149) also points out the importance of adding implicit links and
dropping link words that are not links.
Rozan’s fourth and fifth principles refer to the noting of negation and emphasis.
Negation is usually achieved by crossing out, and emphasis by underlining (Gillies,
2005, p. 106; Matyssek, 1989, pp. 107-110; Rozan, 2002, p. 19; Schweda-Nicholson,
1993, pp. 201-202). Emphasis could also be achieved by shifting, i.e. moving notes
further to the left or right on the notepad (Gillies, 2005, p. 83).
The last two on Rozan’s list are the principles of verticality and shift, the
“backbone” of his system (Rozan, 2002, p. 20). These two are fundamentally principles
on the layout of the notes, and have been given different names by other authors, such
as the use of space (M. Liu, 2008, p. 52) and diagonal layout (Jones, 1998, p. 44; Özben,
1993, p. 42). According to the principles, notes should be structured in a “vertical,
indented and terraced way” (Kohn & Albl-Mikasa, 2002, p. 262), so that the units of
meaning are easy to identify when reading back notes. A mind-mapping note-taking
technique which starts from the centre of the page is also proposed by Torres Díaz
(1997).
Another important part in any existing note-taking system is the use of symbols.
Symbols are used because they are easy to write and read, and represent concepts not
words, thus avoiding source language influence (Gillies, 2005, p. 99). Distributed

25
towards the two ends of the minimalist-maximalist continuum of symbols are Rozan
and Matyssek. The former recommended a total of 20 symbols, of which “only 10 were
indispensable” (Rozan, 2002, p. 25), while the latter used a whole book volume to
introduce a detailed code of drawings and symbols. Although Matyssek’s system was
sometimes criticised as running the risk of becoming an “interpreter’s shorthand” (Ilg &
Lambert, 1996, p. 72), he emphasised that the symbols were suggestions rather than
obligatory requirements (Matyssek, 1989, p. 233). Moreover, an in-depth analysis into
the two systems by Ahrens indicated that they “do not differ at all as far as the basic
principles of note-taking are concerned” (Ahrens, 2005, p. 13). Other authors are more
or less distributed along the continuum, suggesting more symbols than Rozan, but
rejecting the idea of using a symbol-based note-taking system. Generally speaking,
symbols are believed to be very helpful when they are simple, unambiguous, and fully
mastered by the interpreter. It is also pointed out by many authors that it should be
possible to combine symbols to create new symbols (Allioni, 1989; Gillies, 2005;
Matyssek, 1989; Wu, 2008).
So far, it would seem that the principles of note-taking are well-developed, and
once the students are made aware of them and practice accordingly, note-taking should
not be a problem at all. However, when it comes to teaching and learning these
principles, both the students and teachers find it challenging.

2.3 Note-taking didactics: the beginning of a shift from prescriptive to

descriptive

With effective note-taking principles having been worked out and applied by eminent
professionals, two problems now arise: the first is whether these principles and systems
can be taught to students; and if so, the second is how note-taking can be taught
systematically.
The individuality of any note-taking system is emphasised by all who have written
on the topic. This is why some authors do not believe in the systematic teaching of note-
taking. The case in France is typical of this attitude. As Ilg and Lambert pointed out,
“The École Supérieure d'Interprètes et de Traducteurs (ESIT, Paris) never thought much
of note-taking as an underpinning of CI”, and the publications “were sketchy as far as
the techniques of CI are concerned” (Ilg & Lambert, 1996, p. 71). Thiéry (1981) was an
example of this sceptical attitude towards teaching note-taking systematically. He

26
argued that instructions on note-taking should be limited only to essentials, and that
systematic note-taking as a creative and individual activity, could not be taught.
Nevertheless, many authors believe note-taking should be taught systematically to
students, and they have made great efforts to operationalise their didactic proposals. The
discussions target three different student groups: post-graduate level interpreting
students, undergraduate language students, and community interpreters.
The discussions begin with note-taking training for potential candidates of the
profession, usually at post-graduate levels. In fact most of the above-mentioned
literature on note-taking systems and principles fall into this category. Apart from the
publications that focus exclusively on note-taking and treating it as a subject in its own
right, there are also a large quantity of literature that has addressed note-taking as part
of the discussions on interpreter training. Those discussions however, go beyond the
scope of this article. Interested readers are referred to such authors as Bowen and
Bowen (1980), Ilg and Lambert (1996), Jones (1998), Kunihiro et al. (1969),
Seleskovitch and Lederer (1989/1995), Schweda-Nicholson (1985), van Hoof (1962)
and Zhong (1999).
With interpreting being taught to more and more undergraduate language students
as a language reinforcement activity, many teachers have detected the differences in this
new group (e.g. no aptitude testing before entering the classes and great student
numbers), and discussed how to make adaptations accordingly (e.g. Weiwei Dai &
Xiang, 2008; Henderson, 1976; Her, 2001; Paneth, 1984).
Teaching note-taking to community interpreters is uniquely addressed by Schweda-
Nicholson (1990), who is interested in those natural bilinguals without much specialised
training. The goal was to enable community interpreters to benefit from note-taking by
teaching them the basic techniques.
Differences in the type of students lead to differences in the teaching objectives and
choice of materials (e.g. Henderson, 1976; Her, 2001), but the fundamental training
rationales are quite similar. Teachers are well aware that note-taking could take away
attentional resources from other activities in the interpreting process and cause problems.
They usually advise the students against taking notes in the beginning stage of training.
Instead, much attention is devoted to a series of other exercises such as speech analysis,
summarising exercises, and memory training. Actual note-taking is only introduced
after a period of those trainings, and students begin practice with easy materials so that
they are not overwhelmed by the multi-tasking. Gillies (2005) even suggested practising

27
with written materials (transcripts of speeches) rather than spoken ones in the initial
stage.
However, despite the awareness of the difficulties and the precautions taken, both
the teachers and students still find it challenging to teach/learn note-taking. Studies that
describe the difficulties met by students in classes represent the beginning of a shift
from prescriptive to descriptive stream in note-taking research.
Gile (1991) divided 14 students evenly into two groups for CI exercises containing
proper nouns. One group was instructed to take notes and the other was refrained from
doing so except for names and figures. He found that the note-taking group heard the
names worse, and explained that it was because note-taking diverted attention from
listening and led to a degradation of listening quality.
A longitudinal study by Alexieva showed that the instruction in note-taking systems
and principles “brings about a trough in students’ performance, which remains
consistently low for a comparatively long period” (1994, p. 200). The same
phenomenon was found by Her (2001, p. 62). Alexieva (Alexieva, 1994, p. 200)
inferred that at this stage, note-taking learning was characterised by “a weaker memory
operational capacity,” because most of the students’ energy was spent on deciding what
symbols to use, recalling the symbols, and deciding what to put in notes and what to put
in memory.
To see how difficulties were perceived by students in note-taking, Xu and Chai
(2008) used stimulated recall and post-task interviews to investigate the issue. The
major difficulties reported include: insufficient memory, inadequate recall when using
notes as cues, improper form of notes, and overdependence on notes without proper
processing of source information.
Chmiel was interested in the effectiveness of note-taking teaching, and put students
to a test after a note-taking course. The overall results were “less encouraging than
expected” (Chmiel, 2010, p. 248), with the techniques taught in the course being applied
in only 57% of the cases. She also found that layout and visualisation techniques were
more readily transferable than symbols to students’ individual note-taking systems.
Also interested in evaluating learning outcomes, Orlando (2010) made a
technological contribution to the didactic advancement. He pointed out the deficit in the
product-oriented evaluation method, and suggested the application of digital pens, a
technology that allows easy recording of the process of note-taking. The questionnaire

28
results he collected from students showed encouraging potentials of the technology in
classes.
The studies reviewed in this section represent an early descriptive stance taken by
researchers. Instead of simply prescribing how notes should be taken, the authors set out
to observe and describe how notes are actually taken by students. This shift from
prescriptive to descriptive research is strengthened by scholars who approach the topic
of note-taking from linguistic and cognitive perspectives.

2.4 Cognitive and linguistic aspects of note-taking: a theoretical drive

Investigations on the cognitive and linguistic aspects of note-taking are mainly

motivated by an attempt to theorise note-taking and CI. The two pioneering authors and
their investigations (Kirchhoff, 1979; Seleskovitch, 1975) were certainly ahead of their
time.
Seleskovitch (1975) set out to develop a theory to systematise the ESIT’s training
methods. She conducted an experiment in which she collected and analysed the notes
taken by 12 professional interpreters. She found that the notes included few of the
words in the source speech and many outside the speech, that the renditions expressed
much more than the notes, and that some items appeared in different forms. Based on
the findings, she inferred the formal independence of the source speech, notes, and
target speech, pointing to an intermediate stage of “deverbalisation.” Her cognitive
model of interpreting assigned linguistic and cognitive processing to different kinds of
memory, and pointed out that notes functioned as minimal memory triggers, rather than
“an exhaustive code” (Setton, 2002, p. 119).
Standing in contrast to Seleskovitch’s deverbalised view towards note-taking,
Kirchhoff (1979) was concerned about the linguistic surface structures of the notes. She
saw notes as a kind of physical storage as opposed to the cognitive storage of memory.
Note-taking was believed to be a primarily linguistic process, based on the
microstructures of the source text. Her view of notes as a type of language was
supported and followed by Albl-Mikasa, who looked into the language and discourse
dimensions of consecutive notes (Kohn & Albl-Mikasa, 2002), the reduction and
expansion processes in note-taking and note-reading (Albl-Mikasa, 2006) and how
interpreters worked closely along micro-propositional lines when processing the source,
notation and target texts (Albl-Mikasa, 2008). The authors believe that, although the

29
fundamental principle of note-taking is noting the idea and not the word, note-taking
usually operates on a micro-level that stays close to the source text.
Despite the difference in stress (in sense or in linguistic surface structure), the
scholars have consistently pointed out a concurrent storage of information in memory
and in notes, as well as a competition for cognitive resources between note-taking and
other activities in CI, an issue at the core of Gile’s (2009) Effort Models of interpreting.
Giles’s Effort Model of consecutive interpreting conceptualises the interpreting
process in two phases: a comprehension (or listening and note-taking) phase, and a
speech production (or reformulation) phase. The model assumes four processing
capacity demands, or “Efforts” (2009, p. 160) in the first phase, each relating to a
specific activity in the process: Listening and analysis, Note-taking, Short-term memory
operations, and Coordination. In the second phase there are three Efforts: Remembering,
Note-reading, and Production. The Efforts are competing and processing capacity is
limited. In order for interpreting to proceed smoothly, the total processing capacity
demands should not exceed the available capacity, and each Effort should not exceed
the available capacity for each activity. Gile believes note-taking is critical for CI in
terms of cognitive capacity, and the key lies in “how to reduce processing capacity and
time requirements of note-taking while maintaining the efficiency of notes as memory
reinforcers” (2009, p. 178).
Gile’s model, though originally developed to inform teaching, is found useful by
many scholars in academic research. It is mentioned in various explorations on the
prominent features of note-taking.

2.5 Exploring the key note-taking features: descriptive studies on notes and
quality

Unlike the early empirical investigations which have a general interest in what real
notes look like, and set out to discover some overall trends, studies reviewed in this
section have more specific targets. They usually focus on certain note-taking features,
and conduct experiments to closely investigate the features of interest. They have
contributed the largest quantity of empirical data on the topic to date. The most
important variables explored are: the choice of form, the choice of language, and the
relationship between note-taking and interpreting quality.

30
2.5.1 The choice of form and language in note-taking

The choice of form in note-taking refers to the choice between language and symbol,
and the choice between abbreviation and full word; while the choice of language refers
to the choice between source and target language, and the choice between A and B
language6.
A rare and detailed video documentation of note-taking was compiled by Andres
(2002). She recorded the note-taking processes of 14 professionals and 14 students
interpreting from French to German. The notes of the two groups were compared, and
Andres found that despite a source language preference in both groups, the professional
group wrote more target language units than the student group. She also used the time-
coded videos to study time lags in note-taking. According to her findings, the time lag
between listening and note-taking was three to six seconds for professionals, while
reaching as much as ten seconds for students. Her findings provided abundant evidence
of processing overload in students during the first phase of interpreting.
The most comprehensive series of studies to date on note-taking features were
conducted by Dam and her colleagues (Dam, 2004a, 2004b, 2007; Dam, Engberg, &
Schjoldager, 2005). Dam’s study (2004b) with notes taken by four students shows that
the choice of language in note-taking is largely governed by the A/B language status,
rather than the source/target one, with all participants preferring A language regardless
of the direction of interpreting. Her study with five professionals revealed that the
participants’ preferences for the form of note-taking were: symbols (41% of all note
units), followed by full words (35%) and abbreviations (25%) (Dam, 2004a, p. 254).
Again, all participants showed a clear preference for target language, their A language.
She also found that more notes were taken in the source language when the source text
was more difficult.
Dam’s studies were based on CI between Danish and Spanish, and that raises
questions about the generalisability of her results to other language pairs. Following
Dam, other scholars have experimented with different language pairs. Some
representative examples are: Lung (2003), Dai and Xu (2007), Liu (2010), and Wang,
Zhou, and Wang (2010) with Chinese and English; Lim (2006) with Korean and
English; Szabó (2006) with Hungarian and English; and González (2012) with Spanish
and English.

6
In this article, A language refers to the native language while B language refers to the foreign language.

31
 Lung (2003) studied the notes of 21 students interpreting from English to Chinese,
and found that the students made little use of either abbreviations or symbols, and that
the notes consisted mainly of source and B language. Dai and Xu (2007) looked at the
notes taken by 12 students interpreting from Chinese to English, and found that the
notes were source and A language dominated. The 120 students in Liu’s (2010)
experiment on the whole showed a preference for language over symbol, and full word
over abbreviation. Wang et al. (2010) experimented with 12 students, and the notes
were predominantly source language with few symbols used, and abbreviations were
used more than full words. Szabó (2006) looked at the notes taken by eight
professionals interpreting between Hungarian and English, and discovered that her
subjects showed a clear preference for English, their B language, regardless of the
direction of interpreting. The results suggested that the language combination itself
played an important role in the choice of language. Abuín González (2012) compared
the notes taken by three groups of subjects with varying levels of experience (beginner
students, advanced students and interpreters) when interpreting from English to Spanish.
The results showed a shift in language preference from source to target with an
increasing level of expertise.
The details of the studies are summarised in Table 2.1. It is easy to see how they
vary greatly in terms of the design (e.g. type of participants, language pair, interpreting
direction). Moreover, many studies did not specify the details of the tasks used in the
experiment, making it even more difficult to compare the results.

32
Table 2.1 A summary of studies on key note-taking features

Neverthless the author believes it could be beneficial to try and compare the
findings on each note-taking feature (i.e. the choice of form and the choice of language),
and see if some general trends could be detected. Results on the choice of form, as
presented in Table 2.2, point to a dominance of language over symbol, and a slight
tendency to use more full words than abbreviations. Results on the choice of language,
however, yileds much more inconsistent findings.

Table 2.2 Findings on the choice of form in note-taking

To reveal the trends in the choice of language in a clearer way, Table 2.3 organises
the studies according to the type of participants and interpreting direction. While the
language choices of professionals still appear greatly varied, the choices made by
students are obviously source-language dominated. This could be explained using Gile’s
Effort Model. The skills of students are not fully developed, so note-taking consumes a
considerable amount of processing capacity, leaving less available for producing target-

33
language equivalents during the note-taking phase. As a result, students opt for source
language notes to avoid saturation during the first phase. In the second phase, since it is
self-paced, the students have extra time and processing capacity to deal with the
translation.

Table 2.3 Findings on the choice of language in note-taking

What is also made clear in Table 2.3 is that, despite the efforts to describe how
notes are acually taken, there is a lack of research done with professional interpreters.
However, in order to observe know how notes are acually taken in consecutive
interpreting, it is necessary to observe the behaviours of practicing interpreters, rather
than students who have not fully mastered the technique. The same weakness could be
detected in studies on the relationship between note-taking and interpreting quality.

2.5.2 The relationship between note-taking and interpreting quality

Having observed the greatly varied features of note-taking, some researchers begin to
empirically investigate the relationship between these features and the quality of
interpreting performance. Most of the studies use student interpreters as participants,
because quality is an issue at the core of the teaching of interpreting.

34
 Dam et al. (2005) generated hypotheses about features of efficiency and non-
efficiency in notes, based on their proposal to judge the accuracy of the target text
through analysing the semantic network. The hypotheses were later tested by Dam
(2007) with notes taken by five professionals interpreting from Spanish to Danish. She
found evidence for two of the hypotheses: “the more notes, the better the target text –
and vice versa,” and “the more abbreviations/the fewer full words, the better the target
text – and vice versa,” but the data failed to support the third hypothesis: “the more
notes in the source language/the fewer in the target language, the better the target text”
(2007, p. 194).
Experimenting on the language pair of Chinese and English, Her (2001) analysed
the notes taken by undergraduate students interpreting between Chinese and English.
She found that there was a general positive relationship between the quality of notes and
the quality of interpreting, although good notes did not necessarily yield good
performance. Dai and Xu (2007) were unable to find evidence for Dam’s (2007)
hypotheses. Their data showed that an increase in the quantity of notes did not
necessarily mean better target text. Similar conclusions were reached by Liu (2010),
who found no significant difference in the quantity or language of the notes taken by
high- and low-score groups. But he was able to observe that the high-score group used
more symbols than the low-score group. Wang et al. (2010) also found no significant
relationship between interpreting quality and the quantity, form or language of note-
taking. The fact that Dam’s findings were not replicated in the above studies might
partly be explained by the participants used: Dam used professional interpreters, while
the others used students.
Also using students as participants, a study by Cardoen (2013) found relationships
that were opposite to Dam’s findings. Three participants interpreted from Spanish to
Dutch, and Cardoen found that fluent chunks contained fewer notes, more full words
and fewer abbreviations when compared with disfluent chunks.
Studies reviewed in this section are summarised in Table 2.4. They have used
different types of participants and tasks, and they do not always specify the details of
their design. Based on what has been collected so far, it would seem that the interactions
between note-taking and interpreting quality are more complex than researchers have
imagined.

35
Table 2.4 A summary of studies on the relationship between note-taking and
interpreting quality

The empirical studies reviewed in this section vary greatly in terms of their design
(as made evident in Table 2.1 and Table 2.4), and are therefore difficult to compare.
Although some general trends can be found, such as a source language dominance in the
notes taken by students, and more target language in professional interpreters’ notes
compared with students, there are also vast inconsistencies. These inconsistencies are a
great place to start with for future studies.

2.6 Limitations of previous studies

There is no doubt that fruitful results have been created during the past decades, but it is
necessary to point out the limitations in order to inform future research endeavours.
In the prescriptive stream, a common limitation is a lack of systematic and rigorous
empirical research to support the proposals. It is therefore gratifying to see a shift from
prescriptive to descriptive research, with an increase in the quantity of empirical studies.
Also, a variety of research methods have been used, such as simulation, case study,
questionnaire survey, stimulated recall, and interview. However, a few limitations still
exist. First, most of the descriptive studies are product-oriented, but product analysis
only allows speculations about the underlying processes based on data collected
afterwards. Besides due to the highly individualised nature of interpreting notes, it is
often difficult to observe any uniformity in their surface structures. Second, most of the
studies use students instead of professional interpreters as participants, and data is
collected under simulated rather than real life contexts. But in order to get a better

36
picture of how notes are taken, it is necessary to observe the behaviours of professionals
in field interpreting. Third, no study has pushed the shift forward to an explanatory
stream. The researchers usually stop at describing what notes look like, but no one has
designed true experiments to explain the causal relationships behind the phenomena
observed.
In order to initiate a shift from descriptive to explanatory research, an overarching
framework is needed to cohesively pull together all the efforts in note-taking studies. It
is the belief of the author that a cognitive load perspective towards note-taking has great
potentials in that regard.

2.7 Cognitive load: a promising avenue for investigation

Interpreting is deemed a cognitively demanding task by different scholars, many of

whom have pioneered the investigation of cognitive load in simultaneous interpreting
(e.g. Gile, 2008; Hyönä, Tommola, & Alaja, 1995; Seeber, 2011, 2013; Seeber &
Kerzel, 2012; Tommola & Hyönä, 1990). Compared to that, research on cognitive load
in CI and note-taking seems to be scarce. However, as Gile (2009, p. 178) points out,
“note-taking is an area in which the concept of processing capacity can be useful.”
Cognitive load is defined by Seeber (2013, p. 19) as “the amount of capacity the
performance of a cognitive task occupies in an inherently capacity-limited system.”
Starting from a cognitive load perspective, all discussions on note-taking boil down to
one fundamental question: how to reduce the cognitive load of note-taking while
maintaining the efficiency of notes.
If cognitive load can be measured while interpreters take notes and interpret, some
fundamental principles underlying the note-taking choices might be unveiled. For
example, it is possible that no matter what choices an interpreter makes (e.g. writing
notes in the source or target language), the result is always a lower level of cognitive
load for that particular interpreter in that particular task. That is to say, the differences
observed in the note-taking behaviours in previous studies might not be controversies,
but rather converging evidence in proving that interpreters make choices according to
their own situations to reduce cognitive load.
However, measuring cognitive load is no easy task. The construct is generally
believed to be multi-dimensional and therefore difficult to measure. Scholars working
on the Cognitive Load Theory (e.g. Paas, Tuovinen, Tabbers, & Van Gerven, 2003;

37
Paas & Van Merriënboer, 1994) specified two dimensions of cognitive load: a causal
dimension reflecting the factors that affect cognitive load, and an assessment dimension
corresponding to factors that are affected by cognitive load. The assessment factors,
including mental load, mental effort, and performance, are indicative of cognitive load,
and are therefore used for its measurement. A detailed discussion into the assessment
factors and the related measures goes beyond the scope of this article. Interested readers
are referred to such works as Paas et al. (2003) and Plass, Moreno, and Brünken (2010)
for a starting point.
The measurement of cognitive load is not new to the field of interpreting. Many of
the studies are overviewed in Seeber (2013). The pioneering studies have laid the
groundwork by reviewing important theories, building useful models, discussing
methods of measurement, and providing empirical findings. Although the studies have
only investigated simultaneous interpreting, much of what has been discussed is also
meaningful for CI and note-taking. Hopefully, note-taking research would be able to
build on the wealth of those studies and studies in such fields as Cognitive Load Theory,
to overcome the limitations faced by previous studies, and to move forward to an
explanatory stream of note-taking research.

2.8 Bibliography

Abuín González, M. (2012). The language of consecutive interpreters’ notes:

Differences across levels of expertise. Interpreting, 14(1), 55–72.
doi:10.1075/intp.14.1.03abu
Ahrens, B. (2005). Rozan and Matyssek: Are they really that different? A comparative
synopsis of two classic note-taking schools. Forum, 3(2), 1–15.
Albl-Mikasa, M. (2006). Reduction and expansion in notation texts. In C. Heine, K.
Schubert, & H. Gerzymisch-Arbogast (Eds.), Text and translation: Theory and
methodology of translation (pp. 195–214). Tübingen: Narr.
Albl-Mikasa, M. (2008). (Non-) sense in note-taking for consecutive interpreting.
Interpreting, 10(2), 197–231. doi:10.1075/intp.10.2.03alb
Alexieva, B. (1994). On teaching note-taking in consecutive interpreting. In C. Dollerup
& A. Lindegaard (Eds.), Teaching translation and interpreting 2: Insights, aims,
visions (pp. 199–206). Amsterdam: John Benjamins.

38
Allioni, S. (1989). Towards a grammar of consecutive interpretation. In L. Gran & J. M.
Dodds (Eds.), The theoretical and practical aspects of teaching conference
interpretation (pp. 191–197). Udine: Campanotto.
Andres, D. (2002). Konsekutivdolmetschen und notation [Consecutive interpreting and
note-taking]. Frankfurt: Peter Lang.
Becker, W. (1972). Notizentechnik [Note-taking]. Germersheim: BBK.
Bowen, D., & Bowen, M. (1980). Steps to consecutive interpretation. Washington, DC:
Pen and Booth.
Cardoen, H. (2013). The Effect of Note-taking on Target-text Fluency. In G. González
Núñez, Y. Khaled, & T. Voinova (Eds.), Emerging Research in Translation
Studies: Selected Papers of the CETRA Research Summer School 2012 (pp. 1-
22). Leuven: CETRA.
Chmiel, A. (2010). How effective is teaching note-taking to trainee interpreters? The
Interpreter and Translator Trainer, 4(2), 233–250.
Chuang, L.-l. (2008). Note-taking know-how: A processing perspective on consecutive
interpreting. Spectrum: Studies in Language, Literature, Translation, and
Interpretation, 2, 93–101.
Dai, W., & Xiang, C. (2008). 口译教学中记忆、演讲和笔记能力的培养 'The training
of memory, speaking and note-taking ability in interpreting teaching'. Journal of
Southwest University for Nationalities, S3, 101-102.
Dai, W., & Xu, H. (2007). 汉英交替传译过程中译员笔记特征实证研究——以职业
受训译员和非职业译员为例 [An empirical study of the features of interpreters'
notes in Chinese-English consecutive interpreting: The examples of
professionally trained and unprofessional interpreters]. Foreign Language
Teaching and Research, 39(2), 136–144.
Dam, H. V. (2004a). Interpreters' notes: On the choice of form and language. In G.
Hansen, K. Malmkjær, & D. Gile (Eds.), Claims, changes and challenges in
Translation Studies (pp. 251–261). Amsterdam: John Benjamins.
Dam, H. V. (2004b). Interpreters' notes: On the choice of language. Interpreting, 6(1),
3–17. doi:10.1075/intp.6.1.03dam
Dam, H. V. (2007). What makes interpreters’ notes efficient?: Features of (non-
)efficiency in interpreter's notes for consecutive. In Y. Gambier, M. Shlesinger,
& R. Stolze (Eds.), Doubts and directions in Translation Studies: Selected

39
 contributions from the EST Congress, Lisbon 2004 (pp. 183–197). Amsterdam:
John Benjamins.
Dam, H. V., Engberg, J., & Schjoldager, A. (2005). Modelling semantic networks on
source and target texts in consecutive interpreting: A contribution to the study of
interpreters’ notes. In H. V. Dam, J. Engberg, & H. Gerzymisch-Arbogast (Eds.),
Knowledge systems and translation (Vol. 7, pp. 227–254). Berlin: Mouton de
Gruyter.
Deng, Y. (1991). Consecutive notes: An artist's tool. In C.-c. Liu (Ed.), Collected essays
on translation (pp. 284–292). Hong Kong: Commercial Press.
Gile, D. (1991). Prise de notes et attention en début d’apprentissage de l’interprétation
consécutive–une expérience– démonstration de sensibilisation [Note-taking and
attention at the beginning of consecutive interpreting learning –an experience–
demonstration of awareness]. Meta, 36(2–3), 431–439.
Gile, D. (2008). Local cognitive load in simultaneous interpreting and its implications
for empirical research. Forum, 6(2), 59–77.
Gile, D. (2009). Basic concepts and models for interpreter and translator training
(revised edition). Amsterdam: John Benjamins.
Gillies, A. (2005). Note-taking for consecutive interpreting: A short course. Manchester
and Kinderhook: St. Jerome.
Gran, L. (1982). L'annotazione grafica nell'interpretazione consecutiva [Note-taking in
consecutive interpreting]. Trieste: Universitàdegli Studi di Trieste.
Han, Z. (2002). 笔记在口译中的作用 [The function of note-taking in interpreting].
Chinese Science and Technology Translators Journal, 15(2), 25-26, 33.
Henderson, J. A. (1976). Note-taking for consecutive interpreting. Babel, 22(3), 107-
116.
Her, H. (2001). Notetaking in basic interpretation class: An initial investigation. Studies
of Translation and Interpretation, 6, 53-77.
Hyönä, J., Tommola, J., & Alaja, A.-M. (1995). Pupil dilation as a measure of
processing load in simultaneous interpretation and other language tasks. The
Quarterly Journal of Experimental Psychology, 48(3), 598-612.
Ilg, G. (1988). La prise de notes en interprétation consécutive. Une orientation générale
[Note-taking in consecutive interpretation: A general approach]. Parallèles, 9,
9–13.

40
Ilg, G., & Lambert, S. (1996). Teaching consecutive interpreting. Interpreting, 1(1), 69-
99.
Jia, Y. (1995). 口译 笔记 技 巧之 我 见 [My understanding of note-taking skills in
interpreting]. Foreign Language Education, 1, 77-80.
Jones, R. (1998). Conference interpreting explained. Manchester: St. Jerome.
Kirchhoff, H. (1979). Die notationssprache als hilfsmittel des konferenzdolmetschers im
konsekutivvorgang [The language of note-taking as a tool for the conference
interpreter in consecutive interpreting]. In W. Mair & E. Sallager (Eds.),
Sprachtheorie und sprachpraxis [Language theory and language practice] (pp.
121-133). Tübingen: Gunter Narr.
Kohn, K., & Albl-Mikasa, M. (2002). Note-taking in consecutive interpreting: On the
reconstruction of an individualised language. Linguistica Antverpiensia, 1, 257-
272.
Kunihiro, M., Nishiyama, S., & Kanayama, N. (1969). Tsuuyaku: Eikaiwakara
doujitsuuyakumade [Interpreting: from English conversation to simultaneous
interpreting]. Tokyo: Nihonhousoshuppankyoukai.
Lim, H.-O. (2006). A post-mortem of note-taking. Forum, 4(2), 89–112.
Liu, J. (2010). 英语专业本科学生汉英交传笔记特征——一项基于学生交传笔记的
实证研究 [Note-taking characteristics of English majored undergraduates in
Chinese-English consecutive interpreting: an empirical study based on students'
consecutive interpreting notes]. Foreign Language World, 2, 47–53.
Liu, M. (2008). 逐步口譯與筆記 [Consecutive interpretation and note-taking]. Taipei:
Bookman.
Lung, R. (2003). Taking "notes" seriously in the interpretation classroom. In Á.
Collados Aís, M. M. Fernández Sánchez, & D. Gile (Eds.), La evaluación de la
calidad en interpretación: investigación [The evaluation of quality in
interpretation: Research] (pp. 199–205). Granada: Comares.
Matyssek, H. (1989). Handbuch der notizentechnik für dolmetscher [Handbook of note-
taking for interpreters]. Heidelberg: Julius Groos.
Mu, D., & Lei, R. (1998). 试论口译笔记训练中的理解和记忆 [Comprehension and
memory in note-taking training]. Foreign Language Education, 3, 82–84.

41
Orlando, M. (2010). Digital pen technology and consecutive interpreting: Another
dimension in note-taking training and assessment. The Interpreters’ Newsletter,
15, 71–86.
Özben, R. T. (1993). Considerations on the note-taking process during consecutive
interpretation from Italian into Turkish. The Interpreters' Newsletter, 5, 42–48.
Paas, F. G. W. C., Tuovinen, J. E., Tabbers, H., & 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). Instructional control of cognitive
load in the training of complex cognitive tasks. Educational Psychology Review,
6(4), 351–371. doi:10.1007/BF02213420
Paneth, E. (1984). Training in note-taking (For interpreting). In W. Wilss & G. Thome
(Eds.), Die theorie des übersetzens und ihr aufschlußwert für die übersetzungs-
und dolmetschdidaktik [The theory of translation and its value for translating-
interpreting dichotomy] (pp. 326–332). Tübingen: Gunter Narr.
Plass, J. L., Moreno, R., & Brünken, R. (Eds.). (2010). Cognitive Load Theory. New
York: Cambridge University Press.
Pöchhacker, F. (2004). Introducing interpreting studies. London and New York:
Routledge.
Rozan, J.-F. o. (2002). Note-taking in consecutive interpreting (A. Gillies, Trans.).
Cracow: Tertium Society for the Promotion of Language Studies.
Schweda-Nicholson, N. (1985). Consecutive interpretation training: Videotapes in the
classroom. Meta, 30(2), 148. doi:10.7202/003731ar
Schweda-Nicholson, N. (1990). Consecutive note-taking for community interpretation.
In D. Bowen & M. Bowen (Eds.), Interpreting: Yesterday, today, and tomorrow
(pp. 136–145). Amsterdam and Philadelphia: John Benjamins.
Schweda-Nicholson, N. (1993). An introduction to basic note-taking skills for
consecutive interpretation. In E. Losa (Ed.), Keystones of communication:
Proceedings of the 34th Annual Conference of the American Translators
Association (pp. 197–204). Medford: Learned Information.
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

42
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
Seleskovitch, D. (1975). Langage, langues et mémoire: étude de la prise de notes en
interprétation consécutive [Speech, language, and memory: A study of note-
taking in consecutive interpreting]. Paris: Minard Lettres Modernes.
Seleskovitch, D., & Lederer, M. (1989/1995). A systematic approach to teaching
interpretation (J. Harmer, Trans.). Silver Spring, MD: Registry of Interpreters
for the Deaf.
Setton, R. (2002). Revisiting the classics. Seleskovitch: A radical pragmatist before her
time. The Translator, 8(1), 117–124.
Szabó, C. (2006). Language choice in note-taking for consecutive interpreting.
Interpreting, 8(2), 129–147. doi:10.1075/intp.8.2.02sza
Thiéry, C. (1981). L’enseignement de la prise de notes en interprétation consecutive: un
faux problème? [The teaching of note-taking in consecutive interpretation: A
false problem?]. In J. Delisle (Ed.), L’enseignement de l’interprétation et de la
traduction: de la théorie a la pedagogie (pp. 99–112). Ottawa: University of
Ottawa Press.
Tommola, J., & Hyönä, Jukka. (1990). Mental load in listening, speech shadowing and
simultaneous interpreting: A pupillometric study. In J. Tommola (Ed.), Foreign
language comprehension and production (pp. 179–188). Turku: AfinLA.
Torres Díaz, M. G. (1997). Why consecutive note-taking is not tantamount to shorthand
writing. In K. Klaudy & J. Kohn (Eds.), Transferre necesse est. Proceedings of
the Second International Conference on Current Trends in Studies of
Translation and Interpreting 5-7 September 1996, Budapest, Hungary (pp. 213–
216). Budapest: Scholastica.
van Hoof, H. (1962). Théorie et pratique de l'interprétation, avec application
particulière àl'anglais et au français [Theory and practice of interpreting, with
particular application to English and French]. München: Max Hueber.
Wang, W., Zhou, D., & Wang, L. (2010). 口译笔记特征与口译产出质量实证研究
[An empirical study of note-taking characteristics and output quality in
interpreting]. Foreign Language World, 4, 9–18.

43
Wu, Z. (2008). 英语口译笔记法实战指导 [Practical guide of English interpretation
note-taking]. Wuhan: Wuhan University Press.
Xu, H., & Chai, M. (2008). 汉英交替传译活动中译员笔记困难及其原因的实证研究
——以国际会议职业受训译员和非职业译员为例 [Difficulties perceived by
professional trainee interpreters and non-professional interpreters in note-taking
when doing consecutive interpreting: An empirical inquiry through stimulated
recall]. Foreign Language Research, 1, 122–127.
Zhong, S. (1999). A practical handbook of interpretation (enlarged and revised edition).
Beijing: China Translation and Publishing Corporation.
Zhou, P., Su, X., & Leydesdorff, L. (2010). A comparative study on communication
structures of Chinese journals in the social sciences. Journal of the American
society for Information science and Technology, 61(7), 1360–1376.

44
An introductory note to Chapter 3

As has been made clear in Chapter 2, cognitive load is a core issue in consecutive note-
taking. If cognitive load can be measured while interpreters take notes and interpret, the
fundamental principles underlying the note-taking choices might be unveiled.
However, the measurement of cognitive load in interpreting is not an easy
undertaking. The definition of the construct is somewhat under-specified in interpreting
research, accompanied by a lack of systematic discussion on its measurement.
Fortunately, the construct of cognitive load has already received ample attention in
other fields of research before being introduced to the field of interpreting. Findings in
these adjacent fields can inform how cognitive load should be defined and measured in
interpreting studies.
Chapter 3 identifies two of the most relevant constructs to the current research:
mental workload in human factors research and cognitive load in Cognitive Load
Theory research. Building on these two fields, Chapter 3 establishes the theoretical and
methodological framework for the PhD study. It provides the validation of an
operational definition of the construct of cognitive load in interpreting and discusses its
measurement. In particular, this chapter proposes some techniques that are potentially
useful for measuring cognitive load in note-taking and CI.

45
Chapter 3 The construct of cognitive load in interpreting and its
measurement1

Abstract: Interpreting is a cognitively demanding task, and cognitive load in

interpreting is an intriguing topic of research. It is consequently somewhat surprising
that relatively little research has been devoted to the topic to date. This article attempts
to contribute to that effort by presenting an in-depth discussion on the construct of
cognitive load in interpreting and its measurement. Borrowing from mental workload
and Cognitive Load Theory research, cognitive load in interpreting is defined as the
portion of an interpreter’s limited cognitive capacity devoted to performing an
interpreting task in a certain environment. The article then presents a methodological
discussion on how to measure cognitive load, focusing on the major categories of
cognitive load measures and a series of selection criteria. Considering that existing
studies only focus on simultaneous interpreting, the article also introduces some
techniques that are potentially useful for measuring cognitive load in consecutive
interpreting, including the NASA Task Load Index (NASA-TLX), pen recording and
eye tracking.
Keywords: cognitive load; interpreting; measurement; NASA-TLX; pen recording; eye
tracking

3.1 Introduction

Interpreting is seen by many scholars as a cognitively demanding task, and many have
approached the topic from a cognitive perspective. Such an interdisciplinary effort
comes both from outside and within the field. On the one hand, researchers from
psychological and psycholinguistic backgrounds (e.g. Barik, 1973; Christoffels, 2004;
Christoffels & De Groot, 2004, 2005; De Groot, 1997; Gerver, 1976; Köpke &
Signorelli, 2012) are interested in interpreting because it is a special and complex
language processing task. They hope that investigating the cognitive processes in
interpreting will shed light on how the human mind processes language under severe
challenges and in the presence of a high level of multi-tasking. On the other hand,
researchers with a background in interpreting (e.g. Moser-Mercer, 1997; Seeber, 2011,

1
This chapter is a published (on-line first publication) journal article: Chen, S. (2017). The construct of
cognitive load in interpreting and its measurement. Perspectives. doi:10.1080/0907676X.2016.1278026

46
2013; Seeber & Kerzel, 2012; Shlesinger, 2000) wish to approach the topic from an
interdisciplinary perspective. They believe that interpreting research could benefit from
the theories and empirical findings in cognitive sciences.
The measurement of cognitive load has received some attention in studies where
interpreting is investigated from a cognitive perspective. Most of these studies highlight
one particular technique for measuring cognitive load, but methodological discussions
on a cohort of measures are scant (an exception to this is Seeber (2013), whose study
will be discussed in the following section). However, little effort has been devoted to
illustrating the nature of the construct of cognitive load. Moreover, all existing research
efforts cater only for simultaneous interpreting, leaving consecutive interpreting largely
ignored. It is against this backdrop that the article sets out to initiate an in-depth
discussion of the construct of cognitive load and its measurement in interpreting. It is
hoped that such a discussion on a theoretical and methodological framework can attract
more research interest to the topic and reveal potential directions for future studies.
To that end, the article firstly investigates how cognitive load has been studied in
the field of interpreting. Secondly, against the background of the lack of discussion on
the nature of the construct of cognitive load, an overview is provided of how similar
constructs, i.e. mental workload in human factors research and cognitive load in
Cognitive Load Theory, have been defined and investigated, and what can be borrowed
from these adjacent fields. The mental workload model by Meshkati (1988) is identified
as the most important reference for the current study. Thirdly, the article presents a
detailed illustration of cognitive load in interpreting based on the mental workload
model. Meshkati’s model is adapted to the specific case of interpreting. Fourthly, a
methodological discussion on the main categories of cognitive load measures is
presented. The theoretical status of the measures is demonstrated, and a series of
selection criteria are discussed. Given the lack of research on cognitive load in
consecutive interpreting, an attempt is finally made to introduce a number of techniques
that are potentially useful for measuring cognitive load in consecutive interpreting.
The article therefore aims to initiate a discussion on the theoretical and
methodological foundation for further empirical research to be carried out on cognitive
load in interpreting. The comprehensive illustration of the construct provided in this
article will help to standardise the definition and provide an important basis for
cognitive load measurement in the field of interpreting.

47
3.2 Investigating cognitive load in interpreting

The interest in cognitive load in interpreting comes both from an attempt to capture and
understand the difficult and demanding nature of the task, and from a desire to find out
how interpreters deal with the challenges. Few other tasks result in a level of cognitive
load similar to that imposed by interpreting, where ‘no physical activity is involved or
need be accomplished, no instruments can be of help, everything goes on in the mind’
(Riccardi, Marinuzzi, & Zecchin, 1998, p. 97). Nevertheless, wielding the power of
their cognitive systems, interpreters usually succeed in accomplishing this challenging
task. Understanding the nature of a construct such as cognitive load in interpreting is
important for several reasons. Firstly, it sheds light on the multi-lingual processing that
takes place in the human mind under challenging conditions. Secondly, the skills and
strategies used by interpreters to cope with the high load are a central component of
interpreting competence (Kalina, 2000). Thirdly, cognitive load is indicative of the
difficulty of interpreting tasks, which is a key concern in interpreter education and
testing, where tasks need to be carefully selected to meet the varied instructional and
testing demands (Liu & Chiu, 2009).
Although many researchers have investigated interpreting from a cognitive
perspective, cognitive load is studied by only a few, and the main focus is on its
measurement. The studies usually apply one particular technique to explicate the
cognitive processes in (simultaneous) interpreting or to measure the cognitive load. For
example, Petsche, Etlinger, and Filz (1993) found that the electroencephalography
(EEG) coherence measure is useful in identifying task-specific cognitive processing.
Their experiment with professional interpreters showed that when the verbal task is
more difficult, the incidence of coherence increases is higher. In Rinne et al. (2000),
brain activation in professional interpreters during simultaneous interpreting was
measured by positron emission tomography (PET). It was found that brain activation
patterns were modulated by the direction of interpreting. During interpreting into the
non-native language, which was deemed the more demanding task, the activation was
more extensive. Task-evoked pupillary response (TEPR) is used by Hyönä, Tommola,
and Alaja (1995), Seeber and Kerzel (2012), Tommola and Hyönä (1990), and
Tommola and Niemi (1986) to measure cognitive load in simultaneous interpreting. In
these studies, pupil dilation is found to be indicative of both inter- and intra-task load

48
variations. Furthermore, Gile (2009) and Seeber (2011) use models to assign a priori
estimates of cognitive load in interpreting via task analysis.
Despite these pioneering efforts, research conducted on cognitive load in
interpreting has been rather limited to date. In particular, the nature of the construct is
under-researched. It has been termed variously as mental load, processing load and
cognitive load, with these terms used interchangeably in most cases without a formal
definition. When it comes to the measurement of cognitive load, the methodological
discussions are often insufficient in that they tend to advocate the unique applicability
of the one technique being used, but fail to provide a comprehensive picture of the
cohort of measures, their respective theoretical statuses, the selection criteria, and how
they could be combined.
An exception is Seeber (2013), who has contributed a discussion on all the common
measures and methods, their applications in interpreting research, and their respective
advantages and disadvantages. It is also the only study in the field to formally define the
construct. Seeber (2013) defines cognitive load as ‘the amount of capacity the
performance of a cognitive task occupies in an inherently capacity-limited system’. This
is an exciting step towards a systematic investigation of the topic. However, research on
cognitive load in the field is still very limited to date. More efforts are required to get an
in-depth understanding of what cognitive load in interpreting is, and to carry out the
measurement on a solid foundation.
Fortunately, the construct of cognitive load was not new when it was introduced to
interpreting research. Similar constructs have already been investigated, with two of the
most relevant ones being mental workload in human factors research, and cognitive load
in Cognitive Load Theory. In the following section, we will briefly review how these
two constructs are defined and studied in their respective fields, and what can be
borrowed to illustrate cognitive load in interpreting.

3.3 Mental workload in human factors research and cognitive load in Cognitive
Load Theory

Mental workload is a key concern in human factors research, which focuses on how
humans accomplish tasks in the context of human-machine system operation, and how
different variables affect that accomplishment (Meister, 1989). Investigating the relation
between mental workload and human performance is a central research focus, and there
has been a long-standing interest in defining and measuring the workload of human
49
operators. Existing definitions of mental workload usually feature slightly different
terms, mainly due to the research needs associated with different task types. For
example, O’Donnell and Eggemeier (1986) define the construct as ‘that portion of the
operator’s limited capacity actually required to perform a particular task’. Curry, Jex,
Levison, and Stassen (1979, p. 236) define the construct as ‘the mental effort that the
human operator devotes to control and/or supervision relative to his capacity to expend
mental effort’.
Cognitive Load Theory, as the name suggests, is a theory that puts cognitive load in
the centre. It attempts to explore the effects of instruction on cognitive load, and in turn
on learning. Researchers are interested in the cognitive load brought by different
instructional methods. The construct has been defined by Paas and van Merriënboer
(1994, p. 353) as ‘the load that performing a particular task imposes on the cognitive
system of a learner’.
What is important to note is that the two constructs both capture the interactions
between a particular task and a human with limited cognitive capacity. In that sense,
both constructs are relevant to interpreting, where the concept of cognitive load should
be able to capture the interactions between an interpreting task (in a certain environment)
and an interpreter (with limited cognitive capacity). Therefore, both fields serve as good
references for our definition and illustration of cognitive load in interpreting. In this
study, we will mainly draw on mental workload studies, while referring to Cognitive
Load Theory research where applicable. This is mostly because the latter puts less
emphasis on individual characteristics (one admitted limitation of the theory (Bannert,
2002; Moreno & Park, 2010)), whereas interpreter characteristics could potentially
affect cognitive load in interpreting in important ways (see Section 4.2).
A representative and influential mental workload model is given by Meshkati (to
see a full graphic model please refer to Meshkati (1988, p. 307)). The model consists of
a causal section and an effect section, each consisting of two groups of variables. On the
causal section are variables that cause mental workload, including ‘task and
environmental variables’ and ‘operator’s characteristics and moderating variables’; on
the effect section are variables affected by mental workload, including ‘difficulty,
responses and performance’ and ‘mental workload measures’ (Meshkati, 1988, pp. 306-
308). Based on the mental workload model, we will now try to define the construct of
cognitive load in interpreting, and illustrate the important variables in detail.

50
3.4 Cognitive load in interpreting: an illustration of the construct

Based on the definitions of similar constructs in adjacent fields, we now propose to

define cognitive load in interpreting as that portion of an interpreter’s limited cognitive
capacity devoted to performing an interpreting task in a certain environment. It is a
multi-dimensional construct that reflects the interaction of two main groups of variables:
task and environmental characteristics on the one hand, and interpreter characteristics
on the other hand (Figure 3.1). Task and environmental characteristics determine the
amount of mental work to be done in a certain task under certain circumstances. This
dimension of cognitive load is sometimes called the “input load” (Johannsen, 1979, p.
4), or “mental load” (Paas & Van Merriënboer, 1994, p. 353). Interpreter characteristics
are closely related to the effort that is exerted and experienced by a particular interpreter.
This dimension of cognitive load is sometimes called the “operator effort” (Johannsen,
1979, p. 4) or “mental effort” (Paas & Van Merriënboer, 1994, p. 353).

Figure 3.1 A graphical illustration of the construct of cognitive load in interpreting

3.4.1 Task and environmental characteristics

3.4.1.1 Task characteristics

Task characteristics is a topic of interest for many interpreting researchers. Some study
them from the perspective of testing (e.g. Chao, 2015; Chen, 2009), aiming to build a

51
framework of task characteristics for designing authentic interpreting tests. Others start
from the perspective of quality (e.g. Kalina, 2002), looking at factors that could
determine the potential quality of a given interpretation. Drawing on those studies, the
task characteristics included here are: interpreting mode (e.g. simultaneous or
consecutive interpreting), language pair (e.g. some difficulties in interpreting are
specific to certain language pairs), interpreting direction (e.g. from or into the native
language), features of the speech (formal features such as length, speed, and
scripted/spontaneous speech, and content features such as topic, lexical and syntactic
complexity), features of the speaker (e.g. native or non-native speaker, accent, speaking
competence), expected response (e.g. accuracy, language quality, delivery), time on task
(e.g. total hours of working, duration of one turn), preparation (e.g. availability of
background material and advance text), task criticality (the level of harm associated
with poor task performance), and task novelty (how novel an interpreting task is to an
interpreter).
Existing studies on these task characteristics give us a glimpse of the extent to
which they could potentially affect cognitive effort in interpreting. For example, Seeber
and Kerzel (2012) studied the influence of morphosyntactic asymmetry between source
and target language on cognitive load during simultaneous interpreting. They found that
when verb-final and verb-initial constructions were interpreted into a verb-initial
language like English, the former induced larger pupil dilation, suggesting higher
cognitive load. The International Association of Conference Interpreters (AIIC)
commissioned a workload study on professional interpreters (AIIC, 2002). According to
the interpreters surveyed, many of the factors impacting on the level of burnout (‘a
combination of physical fatigue, emotional exhaustion and cognitive weariness’ (AIIC,
2002, p. 12)) were intrinsic task characteristics, such as long hours in the booth, high-
speed speech, and strong accents of the speaker. Moreover, interpreters were asked to
rate the frequency of potential difficulties in their work. According to the results, 60%
of the interpreters rated the frequency of ‘not receiving background material’ high, and
44% rated the frequency of ‘not having enough time to prepare’ high (AIIC, 2002, p.
27). Moser-Mercer, Künzli, and Korac (1998) looked at how increased time on task
affected interpreting performance. The study showed that with task speed held steady,
stress level increased with time during the first 30 minutes of an interpreting task, and
with further time on task (30–60 minutes), performance quality dropped significantly,
indicating cognitive overload (see Section 5.2).

52
 Task criticality and task novelty are not often discussed in interpreting research.
Task criticality is ‘the level of harm associated with performing the task poorly’ (Bunch,
2001, p. 1). For example, interpreting for a national leader giving a welcome speech in a
diplomatic context is usually considered more critical than interpreting a welcome
speech at a business banquet or academic conference. Given the same intrinsic task
characteristics, the level of cognitive load associated with the former task is likely to be
higher than the latter. Task novelty refers to how novel an interpreting task is to an
interpreter. Different tasks vary in novelty, and even the same task could present
different levels of novelty to different interpreters, especially considering that many
interpreters have their areas of specialisation. For example, interpreters who specialise
in business negotiations would find engineering conferences present a higher level of
novelty.

3.4.1.2 Environmental characteristics

Environmental characteristics can potentially play important roles in affecting cognitive

load in interpreting (a typical example would be the detrimental effect of noise on
interpreting performance). A list of these characteristics can also provide a guideline in
terms of which variables to control for in order to create better working conditions for
interpreters, and to obtain more valid experimental results in research.
Environmental variables that are important to the discussion on cognitive load in
interpreting include: physical environment conditions (e.g. location, seating condition,
noise, lighting, temperature, air circulation and quality), visibility of the speaker and/or
audience, and the equipment used. Poor conditions in the physical environment could
lead to an increase in load, as is clearly evident from the results obtained by Parsons
(1978). Visibility of the speaker and/or audience is a factor quite unique to the task of
interpreting. Its importance in facilitating interpreting has long been asserted by
professional interpreters (e.g. Gile, 1990; Rennert, 2008). Although some studies failed
to detect a significant difference in performance between interpreting from visual and
auditory speeches (Jesse, Vrignaud, Cohen, & Massaro, 2000), it was admitted that the
reason could be the presentation of the auditory signal without noise. Moreover,
visibility can become especially important ‘when the verbal message refers to
something visible to the audience or when the nonverbal adds information not present in
the verbal message’ (Rennert, 2008, p. 204). Equipment used in interpreting is typically
not complicated. In simultaneous interpreting, there is usually a control panel, together

53
with earphones and microphones. In consecutive interpreting, the interpreter sometimes
relies exclusively on pen and paper. However, there are some situations where slightly
more complicated equipment could be used. Typical examples would be various types
of remote interpreting, such as telephone and videoconference interpreting. There is also
a new form of interpreting called ‘simultaneous consecutive interpreting’ (Hamidi &
Pöchhacker, 2007), where the source speech is first recorded and then played back to
the interpreter via earphones, and rendered in the simultaneous mode. Under such
circumstances, the influence of equipment on cognitive load could be more significant.

3.4.2 Interpreter characteristics

Interpreter characteristics that affect cognitive load most significantly include the
cognitive abilities, motivation, experience, and state of arousal or activation level of the
interpreter.
Cognitive abilities are at the heart of discussions on such topics as interpreter
competence, expertise and aptitude (e.g. Hoffman, 1997; Macnamara, 2012). Cognitive
abilities that have a strong influence on cognitive load in interpreting include not only
general abilities such as intellect, knowledge (both general knowledge and topical
knowledge), language proficiency, cultural competence, and memory (especially
working memory), but also skills that are specific to certain types of interpreting, such
as note-taking in consecutive interpreting.
Motivation is closely related to an interpreter’s goals and attitude towards a task,
and can affect the focus and level of effort expended on the task. Although there is a
lack of research on its impact on cognitive load, the importance of motivation as a
determinant of interpreting performance has been discussed in multiple studies,
especially from the perspectives of skill acquisition and training (e.g. Moser-Mercer,
2008; Timarová& Salaets, 2011).
The experience of an interpreter comes from both training and working. Experience
is a known variable to influence mental workload (O'Donnell & Eggemeier, 1986;
Young & Stanton, 2001). However, the mechanism of how experience and cognitive
load interact in interpreting remains unclear due to a lack of research. To investigate the
issue, a good starting point might be a comparison between the performance of
interpreters with varied levels of experience (e.g. Köpke & Nespoulous, 2006; Liu,
Schallert, & Carroll, 2004).

54
 State of arousal refers to the physical activation level of an interpreter. To achieve
an optimum level of task performance, it is necessary to have a certain level of
stimulation or arousal. This level of stimulation or arousal varies from interpreter to
interpreter. One simple example would be: some interpreters might find themselves
functioning better in the morning (attentional mechanisms more active), while others
might be more efficient in the afternoon.

3.4.3 Interactions

Interactions (marked by dotted lines in Figure 3.1) could happen both between and
within the two groups of variables in the causal dimension. Between the groups, firstly,
the state of arousal is affected by task criticality, task novelty, and environmental factors.
Higher levels of task criticality and novelty could lead to a higher level of arousal.
Environmental hazards could also lead to high arousal levels which are harmful to task
performance, increasing the risk of cognitive overload. Secondly, task criticality affects
motivation. Increased task criticality could motivate an interpreter, putting the
interpreter in a better state to marshal cognitive resources.
Interactions within the group of interpreter characteristics are shown in that both
motivation and experience affect an interpreter’s arousal state. Motivation usually leads
to higher arousal levels. A lack of experience could also lead to higher arousal,
especially compared to situations when an interpreter is too familiar with a task and
even feels bored.
So far, cognitive load in interpreting has been conceptualised as a multi-
dimensional construct reflecting the interactions between task and environmental
characteristics and interpreter characteristics. On this basis, the following section
presents a methodological discussion on how cognitive load in interpreting could be
measured by discussing the major categories of measures and introducing a series of
selection criteria.

3.5 Measuring cognitive load in interpreting: a methodological discussion

3.5.1 Cognitive load measures

Since cognitive load is essentially a theoretical construct, it cannot be observed and

measured directly. What we can do is to rely on observable and measurable surrogates
that are indicative of cognitive load. A seemingly obvious indicator is the interpreter’s
subjective feeling of effort. The assumption is that, with increased capacity expenditure,

55
the interpreter would feel effort or exertion, which could be self-evaluated with a rating
scale. A second indicator is the interpreting performance. The rationale is that a
decrease in the quality of performance (evident from an increase in elements such as
errors, omissions, and pauses) is likely to be associated with an increase in cognitive
load. A potential third indicator is the physiological arousal of the interpreter. The
assumption is that effort, a major determinant of cognitive load (Paas, Tuovinen,
Tabbers, & Van Gerven, 2003), is quantifiable through measuring the activation level of
the human body. A fourth possible indicator is the interpreting task characteristics. This
is an a priori estimate of cognitive load by analysing task complexity.
The four types of indicators are associated with four categories of cognitive load
measures: subjective measures, performance measures, physiological measures, and
analytical measures (Figure 3.2). Subjective measures are usually produced using
psychometric rating scales. The scales can be either unidimensional or multidimensional.
Unidimensional scales treat cognitive load as a unitary construct and the subject must
assign a single rating to characterise the exerted effort. Multidimensional scales reflect
several factors that contribute to the subjective feelings of effort expenditure and allow
separate ratings on each factor. Performance measures include two types: primary task
measures and secondary task measures. Primary task measures use the interpreting
performance to indicate cognitive load changes. Secondary task measures are produced
through the concurrent performance of an interpreting task and an additional task (the
secondary task). Changes in the performance of the secondary task are evaluated as
evidence of the available spare capacity. Physiological measures approach cognitive
load by observing functions of different body parts such as brain, eye, cardiac system
and muscle. Analytical measures are usually provided by experts or derived from
models or task analysis, based on current knowledge about the task (Paas et al., 2003).

56
Figure 3.2 A graphical illustration of the measurement of cognitive load in interpreting

The cognitive load measures and their application examples both outside and within
the field of interpreting are summarised in Table 3.1.

Table 3.1 Major categories of cognitive load measures and their applications

Representative examples Pioneering studies in the field of

Measures
in adjacent fields interpreting
Subjective Subjective
Hart and Staveland (1988)
measures rating scales
Primary task
Paas et al. (2003) Moser-Mercer et al. (1998)
Performance measures
measures Secondary task Brünken, Steinbacher,
Hu (2008)
measures Plass, and Leutner (2002)
Brain measures Anderson et al. (2011) Petsche et al. (1993); Rinne et al. (2000)
Tommola and Niemi (1986); Tommola
Eye measures Beatty (1982) and Hyönä(1990); Hyönäet al. (1995);
Physiological Seeber and Kerzel (2012)
measures Cardiac system Gunn, Wolf, Block, and
Klonowicz (1994)
measures Person (1972)
Muscle Leyman, Mirka, Kaber,
measures and Sommerich (2004)
Expert opinion Kuperman (1985)
Analytical
Models Wickens (2002) Gile (1995/2009); Seeber (2011)
measures
Task analysis Sweller (1988)

57
 Because cognitive load is a multi-dimensional construct, a single measure cannot
provide a comprehensive picture. Moreover, the measures differ in their granularity and
their relation to the interpreting event in time (i.e. real-time vs. post hoc). Subjective
measures can only provide a post hoc and overall indication of cognitive load.
Performance measures can offer real-time indicators of cognitive load, but they are only
sensitive when the level of load begins to exceed the capacity of the interpreter to
compensate. Physiological measures are both real-time and objective, but their accurate
interpretation is usually reliant on additional subjective measures. Analytical measures
can only be used to estimate the input dimension of cognitive load. Given that the
measures have different strengths and weaknesses, researchers need to consider the
criteria for selecting the appropriate measures for different research situations and
purposes.

3.5.2 Selection criteria

A number of studies have discussed the criteria for selecting mental workload measures.
Two representative examples are O’Donnell and Eggemeier (1986) and Wickens and
Hollands (1999). Among the criteria that have been proposed, the following are
especially relevant to interpreting research: sensitivity, diagnosticity, and intrusiveness.
Sensitivity refers to the potential of a measure to discriminate between changes in
cognitive load. To determine the sensitivity of a measure, we will need to refer to the
theoretical relationship between performance and cognitive load (Figure 3.3, adapted
from O’Donnell and Eggemeier (1986)). The curve specifies three regions according to
the level of cognitive load. In region A, the interpreter has sufficient capacity to cope
with the increasing cognitive load without sacrificing the quality of interpreting
performance. Since no changes could be observed in performance in this region,
primary task measures would be insensitive, while subjective and physiological
measures are more suitable. In region B, cognitive load begins to exceed the capacity of
the interpreter to compensate, and performance decreases and becomes sensitive to load
changes. This is where cognitive load can be, and usually is, measured by primary task
performance. In region C, the level of load is too high, and performance drops to a
catastrophic level. All measures would indicate high cognitive load, but it would be
difficult to differentiate the levels of load. In practice, research is usually centred on
measuring cognitive load in regions A and B.

58
 Diagnosticity refers to the potential of a measure to identify the specific cause of
cognitive load from a variety of sources in the cognitive system. It is based on
Wickens’s (2002) Multiple Resources Theory, which assumes that there are more than
one reservoirs of resources within the human processing system. Sometimes it is
necessary to identify which resources are utilised or to differentiate the level of
demands on certain resources. Some measures (e.g. pupillometry and subjective
measures) only indicate the overall load level and are therefore not considered
diagnostic. Other more diagnostic measures (e.g. secondary task measures and event-
related brain measures) could be used to measure demands on certain resources, but
they are insensitive to other resource types.
Intrusiveness is determined by the extent to which a measure interrupts the
performance of the interpreting task. This is an important concern for interpreting
research because the cognitive system of interpreters is subject to heavy load during the
task. Measures that pose minimum intrusion on the interpreting performance, such as
subjective rating scales, primary task measures, some low-invasive physiological
measures, and analytical measures, are more suitable for interpreting research.
Secondary task measures, however, should be treated with extra caution because of their
intrusiveness.
These criteria should be carefully considered when choosing techniques to measure
cognitive load in interpreting. They are related not only to specific research purposes
but also to the type of interpreting task being investigated. For simultaneous interpreting,
the usefulness of certain techniques has been demonstrated (see Section 2). For
consecutive interpreting, however, discussions on measuring cognitive load are very
rare. Although some techniques could be used for both modes of interpreting, the
unique features of consecutive interpreting present new challenges as well as potentials.
Based on our methodological discussion, the remaining part of this article will be
dedicated to proposing some techniques that are potentially useful for measuring
cognitive load in consecutive interpreting.

3.6 Techniques for cognitive load measurement in consecutive interpreting

3.6.1 Subjective rating scales

There is little evidence that subjective rating scales have been employed to measure
cognitive load in interpreting, but they have been used widely in mental workload and

59
Cognitive Load Theory research, and have been proven valid, non-intrusive, and easy to
implement. It would, therefore, be meaningful to determine the usefulness of some
established rating scales from these areas to interpreting research.
The NASA Task Load Index (NASA-TLX) (Hart & Staveland, 1988) is one of the
most widely used scales for measuring mental load. It is a multi-dimensional rating
procedure that provides an overall load score based on a weighted average of ratings on
six subscales: mental demands, physical demands, temporal demands, performance,
effort, and frustration. Participants need to complete pair-wise comparisons on all
subscales and indicate which is more relevant to their personal load definition. The
number of times a subscale is chosen is the weight. This weighting scheme is used to
take individual differences into account. Participants also need to rate on each subscale
by giving a score that best represents the load experienced during task performance. The
NASA-TLX could be used for both simultaneous and consecutive interpreting. The
rating should be done immediately after an interpreting task.
The NASA-TLX is traditionally done with pen and paper, but now there are also a
few computerised versions available (e.g. Cao, Chintamani, Pandya, & Ellis, 2009). The
digital versions are much faster and more convenient to use. In an exploratory study run
by the author with five professional interpreters, the average time needed to complete
the computerised NASA-TLX2 is under ten minutes. Some studies propose and apply
modified versions of the original scale by adding, deleting or redefining the existing
subscales to improve their relevance to specific tasks or experimental questions (Hart,
2006). This strategy could also be used to better fit the NASA-TLX to interpreting
research, but the validity, sensitivity, and reliability of the modified instrument need to
be tested with empirical data.

3.6.2 Primary task performance, pen recording, and eye tracking

Performance has been a prime concern of interpreting research over the years.
According to the selection criterion of sensitivity (see Section 5.2), primary task
performance can serve as an objective indicator of cognitive load when load begins to
exceed the capacity of the interpreter to compensate, and performance decreases and
becomes sensitive to load changes (region B in Figure 3.3). Under such circumstances,
a performance score could indicate the overall cognitive load induced by a certain task.
Performance scores are therefore sometimes used to determine the difficulty of
2
A link to this computerised version of the NASA-TLX can be found at http://www.nasatlx.com/.

60
interpreting tasks (usually in combination with subjective reports and/or expert opinion)
in interpreter education and testing (e.g. Liu & Chiu, 2009).

Figure 3.3 A hypothetical relationship between cognitive load and interpreting

performance

Other performance measures such as the target speech features of span and errors
have also been studied by many. For example, pauses in consecutive interpreting have
been studied by Mead (2000, 2002, 2005), while ear-voice-span and errors in
simultaneous interpreting have been studied by Altman (1994), Barik (1973), and
Bendazzoli, Sandrelli, and Russo (2011). Although the target speech features are not
usually viewed from a cognitive load perspective, those initial investigations have laid a
useful foundation by presenting methods to analyse and quantify interpreting
performance. Now with physiological measures (such as eye tracking) gaining
popularity, performance indicators could be collected in ways that have not been
possible before, providing more powerful tools to reveal information on cognitive load.
Two potential indicators that can be investigated by combining performance and
physiological techniques, one for each of the two phases of consecutive interpreting,
deserve particular attention.
Phase I of consecutive interpreting is the ‘listening and note-taking phase’ (Gile,
2009, p. 175). The indicator that can be investigated in this phase is the ear-pen span,
which is the interval between the moment a speech segment is heard and the moment it

61
is noted down. More specifically, it is calculated from the offset of voice to the onset of
pen stroke. The hypothesis is that longer ear-pen span indicates higher cognitive load.
When a source speech unit is difficult and the cognitive load is high, it takes longer to
process that unit, to make the decision, and to put that unit into written notes. But a long
span could also indicate high cognitive load in processing units other than the one that
is noted down. A more fine-grained analysis could be reached by combining the span
data with retrospection. The recorded phase I can be played back to the interpreter,
providing retrieval cues to ensure accurate and comprehensive retrospection.
Using a piece of software called the Eye and Pen3, the note-taking process taking
place on a digital tablet can be recorded. The tablet works together with a digital pen,
and transmits the spatial, temporal and pressure data to the software as the pen moves
across its surface. The software can then analyse and reconstruct the writing process,
giving real-time data such as the distance, duration, speed, and pressure of the pen, as
well as the pauses between pen strokes. The source speech and the pen data share the
same timeline, so it is possible to accurately pinpoint the offset of any source speech
unit and the onset of any pen stroke, allowing the calculation of the ear-pen span.
Phase II of consecutive interpreting is the ‘speech production phase’ (Gile, 2009, p.
176). The indicator that can be investigated in this phase is the eye-voice span4, which is
the interval between the moment a note unit is read and the moment it is produced in the
target speech. More specifically, it is the interval between the onset of a note’s fixation
and the onset of its articulation. The hypothesis is that longer eye-voice span indicates
higher cognitive load. If a note unit is difficult to process (for example when a highly
abstract symbol is used or the handwriting is illegible) and the associated cognitive load
is high, it takes longer before the interpreter can produce an equivalent in the target
speech. Again, a retrospection of the production phase (possibly stimulated by the scan
path video of eye movements) can help with a more fine-grained analysis of the span
data. For example, it can help distinguish whether a long eye-voice span is caused by
the difficulty in processing a certain note unit, or by recalling from memory information
that’s not relevant to the note unit.

3
The website of the software is http://www.eyeandpen.net/?lng=en.
4
This is different from the eye-voice span widely discussed in reading research. The eye-voice span in
reading research is measured as the distance between the fixated item and the pronounced item during
reading aloud.

62
 Using eye tracking, the production phase can be recorded in detail. Eye tracking
data reveals information on when and for how long a note unit is being fixated. Some
eye trackers (such as the SMI Eye Tracking Glasses and the Tobii Pro Glasses) could
synchronously collect eye data while recording a scene video with sounds. The eye data
and the voice data share the same timeline, making it possible to calculate the eye-voice
span.
Of course both pen recording and eye tracking yield other potential indicators of
cognitive load, such as pen orientation and pressure (e.g. Yu, Epps, & Chen, 2011a; Yu,
Epps, & Chen, 2011b) and number and duration of eye fixations (e.g. Buettner, 2013;
Just & Carpenter, 1976, 1980; Van Orden, Limbert, & Makeig, 2001). Different sources
of data could be triangulated to present a comprehensive picture of cognitive load in
consecutive interpreting.

3.6.3 Models

The analytical measures of expert opinion and task analysis are used mostly in
interpreting education and testing for the purpose of gauging task difficulty. Teachers
and test developers often have to deal with tasks that vary in different aspects. They
make judgements about the complexity, or difficulty, of the tasks based on those
different aspects in order to choose appropriate practising materials and test tasks to
match different training and testing objectives (Liu & Chiu, 2009). Models, compared
with the previous two types of analytical measures, are used more for research purposes.
Two outstanding examples are Gile’s (2009) Effort Models, and Seeber’s (2011)
Cognitive Load Models. The latter is especially suitable for making predictions on local
cognitive load.
The Cognitive Load Models developed by Seeber are based on Wickens’s (2002)
Multiple Resources Theory (see Section 5.2). They are able to quantify cognitive load
relying principally on Wickens’s demand vectors and conflict coefficients. Local
cognitive load changes are reflected by analysing the trade-offs between time-sharing
activities in the interpreting process. Although the models are designed for simultaneous
interpreting, they can be adapted to cater to the situations in consecutive interpreting.
Using the same principles, the first phase of consecutive interpreting can be considered
as a real-time combination of a listening and a note-taking task, while the second phase
consists of a note-reading and a speech production task. Both tasks can be broken down
into their resource demand vectors and analysed using the Multiple Resources Theory.

63
The adapted models will be able to make specific predictions on the changes in
cognitive load brought by different strategies and serve as a useful analytical technique
for estimating cognitive load in consecutive interpreting.
It exceeds the scope of this article to expand the introduced techniques in full or to
present detailed application procedures. This article is only an initial step in a more
comprehensive research project on cognitive processing and cognitive load in
consecutive interpreting. However, it is hoped that the brief introductions made here of
the potential techniques set the agenda for more comprehensive and meticulous
experimental research on cognitive load during consecutive interpreting.

3.7 Conclusions

Building on previous studies in the field of interpreting, and borrowing from adjacent
fields such as mental workload and Cognitive Load Theory research, this article
attempts to formally define and illustrate the construct of cognitive load in interpreting
and to discuss its measurement. Cognitive load in interpreting is conceptualised as a
multi-dimensional construct that reflects the interaction between an interpreting task (in
a certain environment) and an interpreter (with certain characteristics). It is a theoretical
construct that cannot be observed directly, and its measurement relies on observable
surrogates that are indicative of cognitive load. There are four major categories of
cognitive load measures: subjective, performance, physiological and analytical
measures. They cater to different research purposes and are applicable in different
circumstances. A series of criteria that could help with the selection and combination of
the measures is also discussed. Finally, considering the lack of research on cognitive
load in consecutive interpreting, the article introduces some techniques that are
potentially useful for measuring cognitive load in consecutive interpreting. Our next
step is to validate the measurement techniques introduced in this article. Although they
have rarely been applied in empirical research so far, the techniques have provided
some exciting possibilities.
The significance of defining and operationalising cognitive load in interpreting lies
in the foundation it provides for investigating the cognitively challenging task of
interpreting experimentally. Such a theoretical framework can be helpful in identifying
triggers of cognitive overload, facilitating research on the interpreting process. Research
on cognitive load in interpreting is only starting, and hopefully more interested
researchers will be joining the effort.
64
3.8 References

AIIC. (2002). Interpreter workload study - Full report. Retrieved 7 February 2002
http://aiic.net/page/657
Altman, J. (1994). Error analysis in the teaching of simultaneous interpretation: A pilot
study. In S. Lambert & B. Moser-Mercer (Eds.), Bridging the gap: Empirical
research in simultaneous interpretation (pp. 25–38). Amsterdam: John
Benjamins.
Anderson, E. W., Potter, K. C., Matzen, L. E., Shepherd, J. F., Preston, G. A., & Silva,
C. T. (2011). A user study of visualization effectiveness using EEG and
cognitive load. Computer Graphics Forum, 30(3), 791–800. doi:10.1111/j.1467-
8659.2011.01928.x
Bannert, M. (2002). Managing cognitive load: Recent trends in cognitive load theory.
Learning and Instruction, 12, 139–146.
Barik, H. C. (1973). Simultaneous interpretation: Temporal and quantitative data.
Language and Speech, 16(3), 237–270. doi:10.1177/002383097301600307
Beatty, J. (1982). Task-evoked pupillary responses, processing load, and the structure of
processing resources. Psychological Bulletin, 91(2), 276–292.
Bendazzoli, C., Sandrelli, A., & Russo, M. (2011). Disfluencies in simultaneous
interpreting: A corpus-based analysis. In A. Kruger, K. Wallmach, & J. Munday
(Eds.), Corpus-based translation studies: Research and applications (pp. 282–
306). London and New York: Continuum.
Brünken, R., Steinbacher, S., Plass, J. L., & Leutner, D. (2002). Assessment of
cognitive load in multimedia learning using dual-task methodology.
Experimental Psychology, 49(2), 109–119. doi:10.1027//1618-3169.49.2.109
Buettner, R. (2013). Cognitive workload of humans using artificial intelligence systems:
Towards objective measurement applying eye-tracking technology. In I. J.
Timm & M. Thimm (Eds.), KI 2013: Advances in artificial intelligence (pp. 37–
48). Berlin Heidelberg: Springer. doi:10.1007/978-3-642-40942-4
Bunch, M. B. (2001). The SSTI/NAJIT translation and interpretation national
certification examination. Paper presented at the Annual Meeting of the
American Translators Association, Los Angeles.
http://www.najit.org/certification/ATAPRESR%20on%20Exam%20History%20
and%20development%20dlm.PDF

65
Cao, A., Chintamani, K. K., Pandya, A. K., & Ellis, R. D. (2009). NASA TLX:
Software for assessing subjective mental workload. Behavior Research Methods,
41(1), 113–117. doi:10.3758/BRM.41.1.113
Chao, H. (2015). Building the validity foundation for interpreter certification
performance testing (Unpublished doctoral thesis), Macquarie University,
Sydney.
Chen, J. (2009). Authenticity in accreditation tests for interpreters in China. The
Interpreter and Translator Trainer, 3(2), 257–273.
doi:10.1080/1750399X.2009.10798791
Christoffels, I. K. (2004). Cognitive studies in simultaneous interpreting (Unpublished
doctoral thesis), University of Amsterdam, Amsterdam.
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
Christoffels, I. K., & De Groot, A. M. B. (2005). Simultaneous interpreting: A cognitive
perspective. In J. F. Kroll & A. M. B. De Groot (Eds.), Handbook of
bilingualism: Psycholinguistic approaches (pp. 454–479). New York: Oxford
University Press.
Curry, R., Jex, H., Levison, W., & Stassen, H. (1979). Final report of control
engineering group. In N. Moray (Ed.), Mental workload: Its theory and
measurement (pp. 235–253). New York: Plenum Press.
De Groot, A. M. B. (1997). The cognitive study of translation and interpretation: Three
approaches. In J. H. Danks, G. M. Shreve, S. B. Fountain, & M. K. McBeath
(Eds.), Cognitive processes in translation and interpreting (pp. 25–56). London
and New Delhi: Thousand Oaks.
Gerver, D. (1976). Empirical studies of simultaneous interpretation: A review and a
model. In R. B. W. Anderson & R. W. Brislin (Eds.), Translation: Applications
and research (pp. 165–207). New York: Gardner Press.
Gile, D. (1990). Scientific research vs. personal theories in the investigation of
interpretation. In L. Gran & C. Taylor (Eds.), Aspects of applied and
experimental research on conference interpretation (pp. 28–41). Udine:
Campanotto Editore.

66
Gile, D. (2009). Basic concepts and models for interpreter and translator training
(revised edition). Amsterdam: John Benjamins.
Gunn, C. G., Wolf, S., Block, R. T., & Person, R. J. (1972). Psychophysiology of the
cardiovascular system. In N. S. Greenfield & R. A. Sternbach (Eds.), Handbook
of psychophysiology. New York: Holt, Rinehart and Winston.
Hamidi, M., & Pöchhacker, F. (2007). Simultaneous consecutive interpreting: A new
technique put to the test. Meta, 52(2), 276-289. doi:10.7202/016070ar
Hart, S. G. (2006). NASA-Task Load Index (NASA-TLX); 20 years later. Proceedings
of the Human Factors and Ergonomics Society Annual Meeting, 50(9), 904-908.
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
Publishers B.V.
Hoffman, R. R. (1997). The cognitive psychology of expertise and the domain of
interpreting. Interpreting, 2(1), 189-230.
Hu, L. (2008). Note-taking in consecutive interpreting: A psycholinguistic perspective
(Unpublished doctoral thesis), Shanghai International Studies University,
Shanghai.
Hyönä, J., Tommola, J., & Alaja, A.-M. (1995). Pupil dilation as a measure of
processing load in simultaneous interpretation and other language tasks. The
Quarterly Journal of Experimental Psychology, 48(3), 598-612.
Jesse, A., Vrignaud, N., Cohen, M. M., & Massaro, D. W. (2000). The processing of
information from multiple sources in simultaneous interpreting. Interpreting,
5(2), 95-115. doi:10.1075/intp.5.2.04jes
Johannsen, G. (1979). Workload and workload measurement. In N. Moray (Ed.),
Mental workload: Its theory and measurement (pp. 3-11). New York: Plenum
Press.
Just, M. A., & Carpenter, P. A. (1976). Eye fixations and cognitive processes. Cognitive
Psychology, 8(4), 441-480. doi:10.1016/0010-0285(76)90015-3
Just, M. A., & Carpenter, P. A. (1980). A theory of reading: From eye fixations to
comprehension. Psychological Review, 87(4), 329-354. doi:10.1037/0033-
295X.87.4.329

67
Kalina, S. (2000). Interpreting competences as a basis and a goal for teaching. The
Interpreters’ Newsletter, 10, 3-32. Retrieved from
http://hdl.handle.net/10077/2440
Kalina, S. (2002). Quality in interpreting and its prerequisites: A framework for a
comprehensive view. In G. Garzone & M. Viezzi (Eds.), Interpreting in the 21st
century: Challenges and opportunities (pp. 121-130). Amsterdam: John
Benjamins.
Klonowicz, T. (1994). Putting one's heart into simultaneous interpretation. In S.
Lambert & B. Moser-Mercer (Eds.), Bridging the gap: Empirical research in
simultaneous interpretation (pp. 213-224). Amsterdam: John Benjamins.
Köpke, B., & Nespoulous, J.-L. (2006). Working memory performance in expert and
novice interpreters. Interpreting, 8(1), 1-23. doi:10.1075/intp.8.1.02kop
Köpke, B., & Signorelli, T. M. (2012). Methodological aspects of working memory
assessment in simultaneous interpreters. International Journal of Bilingualism,
16(2), 183-197. doi:10.1177/1367006911402981
Kuperman, G. G. (1985). Pro-SWAT applied to advanced helicopter crewstation
concepts. Proceedings of the Human Factors and Ergonomics Society Annual
Meeting, 29(4), 398-402. doi:10.1177/154193128502900420
Leyman, E. L. C., Mirka, G. A., Kaber, D. B., & Sommerich, C. M. (2004).
Cervicobrachial muscle response to cognitive load in a dual-task scenario.
Ergonomics, 47(6), 625-645. doi:10.1080/00140130310001629766
Liu, M., & Chiu, Y.-H. (2009). Assessing source material difficulty for consecutive
interpreting: Quantifiable measures and holistic judgment. Interpreting, 11(2),
244–266. doi:10.1075/intp.11.2.07liu
Liu, M., Schallert, D. L., & Carroll, P. J. (2004). Working memory and expertise in
simultaneous interpreting. Interpreting, 6(1), 19–42.
Macnamara, B. (2012). Interpreter cognitive aptitudes. Journal of Interpretation, 19(1),
9–32. Retrieved from http://digitalcommons.unf.edu/joi/vol19/iss1/1
Mead, P. (2000). Control of pauses by trainee interpreters in their A and B languages.
The Interpreters' Newsletter, 10, 89–102. Retrieved from
http://hdl.handle.net/10077/2451
Mead, P. (2002). Exploring hesitation in consecutive interpreting: An empirical study.
In G. Garzone & M. Viezzi (Eds.), Interpreting in the 21st century: Challenges
and opportunities (pp. 73–82). Amsterdam: John Benjamins.

68
Mead, P. (2005). Methodological issues in the study of interpreters' fluency. The
Interpreters' Newsletter, 13, 39–63. Retrieved from
http://hdl.handle.net/10077/2469
Meister, D. (1989). Conceptual aspects of human factors. Baltimore: Johns Hopkins
University Press.
Meshkati, N. (1988). Toward development of a cohesive model of workload. In P. A.
Hancock & N. Meshkati (Eds.), Human mental workload (pp. 305–314).
Moreno, R., & Park, B. (2010). Cognitive Load Theory: Historical development and
relation to other theories. In J. L. Plass, R. Moreno, & R. Brünken (Eds.),
Cognitive Load Theory (pp. 9–28). New York: Cambridge University Press.
Moser-Mercer, B. (1997). Beyond curiosity: Can interpreting research meet the
challenge? In J. H. Danks, G. M. Shreve, S. B. Fountain, & M. K. McBeath
(Eds.), Cognitive processes in translation and interpreting (pp. 176–195).
London and New Delhi: Thousand Oaks.
Moser-Mercer, B. (2008). Skill acquisition in interpreting. The Interpreter and
Translator Trainer, 2(1), 1–28. doi:10.1080/1750399X.2008.10798764
Moser-Mercer, B., Künzli, A., & Korac, M. (1998). Prolonged turns in interpreting:
Effects on quality, physiological and psychological stress (Pilot study).
Interpreting, 3(1), 47–64.
O'Donnell, C. R. D., & Eggemeier, F. T. (1986). Workload assessment methodology. In
K. R. Boff, L. Kaufman, & J. P. Thomas (Eds.), Handbook of perception and
human performance, vol. II: Cognitive processes and performance (pp. 42-41–
42-49). New York: Wiley-Interscience.
Paas, F. G. W. C., Tuovinen, J. E., Tabbers, H., & 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). Instructional control of cognitive
load in the training of complex cognitive tasks. Educational Psychology Review,
6(4), 351–371. doi:10.1007/BF02213420
Parsons, H. M. (1978). Human factors approach to simultaneous interpretation. In D.
Gerver & H. W. Sinaiko (Eds.), Language interpretation and communication
(pp. 315–321). New York: Springer US.

69
Petsche, H., Etlinger, S. C., & Filz, O. (1993). Brain electrical mechanisms of bilingual
speech management: An initial investigation. Electroencephalography and
Clinical Neurophysiology, 86(6), 385–394. doi:10.1016/0013-4694(93)90134-H
Rennert, S. (2008). Visual input in simultaneous interpreting. Meta, 53(1), 204–217.
doi:10.7202/017983ar
Riccardi, A., Marinuzzi, G., & Zecchin, S. (1998). Interpretation and stress. The
Interpreters' Newsletter, 8, 93–106. Retrieved from
http://hdl.handle.net/10077/2493
Rinne, J. O., Tommola, J., Laine, M., Krause, B. J., Schmidt, D., Kaasinen, V., . . .
Sunnari, M. (2000). The translating brain: Cerebral activation patterns during
simultaneous interpreting. Neuroscience Letters, 294(2), 85–88.
doi:10.1016/S0304-3940(00)01540-8
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., & Kerzel, D. (2012). Cognitive load in simultaneous interpreting: Model
meets data. International Journal of Bilingualism, 16(2), 228–242.
doi:10.1177/1367006911402982
Shlesinger, M. (2000). Interpreting as a cognitive process: How can we know what
really happens? In S. Tirkkonen-Condit & R. Jääskeläinen (Eds.), Tapping and
mapping the processes of translation and interpreting: Outlooks on empirical
research (pp. 3–16). Amsterdam and Philadelphia: John Benjamins.
Sweller, J. (1988). Cognitive load during problem solving: Effects on learning.
Cognitive Science, 12, 257–285. doi:10.1207/s15516709cog1202_4
Timarová, Š., & Salaets, H. (2011). Learning styles, motivation and cognitive flexibility
in interpreter training: Self-selection and aptitude. Interpreting, 13(1), 31–52.
doi:10.1075/intp.13.1.03tim
Tommola, J., & Hyönä, Jukka. (1990). Mental load in listening, speech shadowing and
simultaneous interpreting: A pupillometric study. In J. Tommola (Ed.), Foreign
language comprehension and production (pp. 179–188). Turku: AfinLA.
Tommola, J., & Niemi, P. (1986). Mental load in simultaneous interpreting: An on-line
pilot study. In L. S. Evensen (Ed.), Nordic research in text linguistics and
discourse analysis (pp. 171–184). Trondheim: Tapir.

70
Van Orden, K. F., Limbert, W., & Makeig, S. (2001). Eye activity correlates of
workload during a visuospatial memory task. Human Factors, 43(1), 111–121.
Wickens, C. D. (2002). Multiple resources and performance prediction. Theoretical
Issues in Ergonomics Science, 3(2), 159–177. doi:10.1080/14639220210123806
Wickens, C. D., & Hollands, J. G. (1999). Engineering psychology and human
performance (3rd edition). Upper Saddle River, New Jersey: Prentice Hall.
Young, M., & Stanton, N. (2001). Mental workload: theory, measurement, and
application. In W. Karwowski (Ed.), International encyclopedia of ergonomics
and human factors (Vol. 1, pp. 507–509). London: Taylor and Francis.
Yu, K., Epps, J., & Chen, F. (2011a). Cognitive load evaluation of handwriting using
stroke-level features. Paper presented at the 16th International Conference on
Intelligent User Interfaces, Palo Alto, CA, USA. http://www.nicta.com.au/pub-
download/full/5228/
Yu, K., Epps, J., & Chen, F. (2011b). Cognitive load measurement with pen orientation
and pressure. Paper presented at the ICMI 2011 Workshop on Inferring
Cognitive and Emotional States from Multimodal Measures, Alicante, Spain.
http://www.nicta.com.au/pub-download/full/5228/

71
An introductory note to Chapter 4

The theoretical and methodological discussions in Chapter 3 have identified a series of

techniques that are potentially useful for measuring cognitive load in note-taking and CI.
But before these techniques can be applied to the main study of this PhD project, their
applicability needs to be tested, especially the ones that are seldom used in interpreting
research.
Chapter 4 reports on a pilot study which focuses on the synchronised recording and
analysis of pen and voice data. It documents how the tasks are carefully created and
controlled for variance, how the sample participants are selected through stringent
criteria, how the apparatus are set up so that abundant and varied sources empirical data
can be recorded in parallel, and how data processing and analysis are carried out. The
purpose of the pilot study is three-fold: (1) to prove that pen recording is a useful
method to tap into the process of note-taking and CI; (2) to devise some useful
indicators of cognitive load with the combined recording of pen and voice data; and (3)
to provide instrumental findings for the hypotheses to be made in the main study.

72
Chapter 4 Note-taking in consecutive interpreting: New data from
pen recording1

Abstract: Note-taking provides a unique opportunity to investigate consecutive

interpreting (CI). This study approaches note-taking from a cognitive perspective,
combining product analysis with the process research method of pen recording. It
investigates such variables as the choice of form, the choice of language, the
relationship between note-taking and interpreting performance, and the relationship
between note-taking and cognitive load in CI. In the context of CI between Chinese and
English, the study finds that interpreters prefer language to symbol, abbreviation to full
word, and English to Chinese regardless of the direction of interpreting. Interpreting
performance is not directly related to either the quantity or the quality of notes; it is a
function of both. Pen recording appears to be a powerful method to tap into the process
of note-taking and CI, and the collected data could potentially serve as useful indicators
of cognitive load.
Keywords: consecutive interpreting; note-taking; pen recording; cognitive load

4.1 Introduction

The research interest in cognitive processing in translation and interpreting is increasing,

but the focus on consecutive interpreting (CI) is very limited to date. Note-taking is a
distinctive feature of CI 2 , and provides a unique opportunity to investigate the
interpreting process. For over half a century, research on note-taking in CI has yielded
fruitful results. A series of variables have been investigated, including the choice of
form, the choice of language, and the relationship between note-taking and interpreting
performance. However, existing studies on note-taking and CI are mostly product-
oriented, revealing little information about the process.
This study attempts to address that limitation by combining product analysis with
an investigation into the interpreting process. Using pen recording and a software called
the Eye and Pen3, pen data during the note-taking process are recorded in great details.

1
This chapter is a published journal article: Chen, S. (2017). Note-taking in consecutive interpreting:
New data from pen recording. Translation and Interpreting, 9(1), 4–23. doi:10.12807/ti.109201.2017.a02
2
In this article, CI refers to long consecutive where systematic note-taking is used.
3
The website of the software is http://www.eyeandpen.net/?lng=en.

73
Pen strokes are measured in terms of distance, duration, and speed. Such a recording not
only tells us what interpreters’ note-taking choices are, but also shows us how
interpreters carry out those choices. The pen data are further investigated from a
cognitive perspective, with an aim to see if they can be used as indicators of cognitive
load in note-taking and CI.

4.2 Note-taking in CI: a brief review

The large volume of literature generated by scholars’ sustained interest in note-taking

can be roughly divided into two streams: a prescriptive stream and a descriptive stream
(see Chen (2016) for a more comprehensive review). At the earliest stage, a number of
prescriptive works have introduced some well-known note-taking systems and
principles (e.g., Kirchhoff, 1979; Matyssek, 1989; Rozan, 1956/2002). Later on,
noticing the challenges brought by the teaching and learning of note-taking in
classrooms, some scholars begin to observe how notes are actually taken by student
interpreters (e.g., Alexieva, 1994; Gile, 1991). These studies represent the beginning of
a shift in note-taking literature from being prescriptive to becoming descriptive. Some
researchers have also investigated the cognitive and linguistic aspects of note-taking,
pointing out a concurrent storage of information in memory and in notes (e.g.,
Seleskovitch, 1975) and that note-taking operates on a micro-level that stays close to the
source text (e.g., Albl-Mikasa, 2008; Kohn & Albl-Mikasa, 2002). The more recent
studies in the descriptive stream usually target specific note-taking choices, collecting
data in simulated interpreting tasks and contributing valuable empirical evidence (e.g.,
Abuín González, 2012; Andres, 2002; Dam, 2004b; Szabó, 2006). In all these studies,
three variables have received the majority of the attention: the choice of form, the
choice of language, and the relationship between note-taking and interpreting
performance.
Interpreters make choices (although not always consciously) on the form of notes:
whether to take notes in symbol or language, and if in language, whether to write the
word in full or to abbreviate it. Many prescriptive publications introducing note-taking
systems put the use of symbols and abbreviations at a prominent position. Compared to
language, symbols are easy to write and read, and can help avoid source language
influence because they represent concepts rather than specific words (Gillies, 2005, p.
99). But the prescriptive suggestion on how many symbols should be used varies from

74
system to system. At the minimalist end was Rozan, who recommended a total of 20
symbols, of which “only 10 were indispensable” (1956/2002, p. 25). At the maximalist
end was Matyssek (1989), who used a whole book volume to introduce a detailed code
of drawings and symbols. As to the use of abbreviations, it is generally suggested that
long words (more than 4 to 5 letters according to Rozan (1956/2002, p. 16)) should be
abbreviated to save time and effort spent on writing the notes.
The choice of form has also been empirically investigated in such studies as Andres
(2002), Dam (Dam, 2004a, 2004b), Lung (2003), Dai and Xu (2007), Liu (2010), and
Wang, Zhou, and Wang (2010). The results pointed to a preference for language over
symbol, whereas findings on the choice between abbreviation and full word were
inconsistent. Most studies recruited student interpreters and some interviewed them
afterwards, revealing some potential causes for the preference. Students tended to write
down everything as it was heard and were creating symbols on the spot instead of using
pre-established symbol systems. Both of these practices limited the use of symbols in
note-taking. However, it is questionable whether these findings could be generalised to
professional interpreters.
The choice of language is perhaps the most controversial variable in note-taking
literature. Traditionally, the categories used to discuss this choice are source and target
language. Source language is suggested in some prescriptive literature (e.g., Alexieva,
1994; Gile, 2009; Kirchhoff, 1979) based on the belief that interpreters can “minimize
their effort and save capacity” (Szabó, 2006, p. 131) during the listening phase under
great time pressure. However, target language is recommended in others (e.g., Herbert,
1952; Jones, 1998; Rozan, 1956/2002) because the authors believe it makes the target
speech production phase less effortful, and facilitates better processing of the source
speech.
With further empirical data available, some researchers begin to find that the
language choice is also affected by whether a language is the A or B language in an
interpreter’s language combination. In this study, A language refers to the native
language while B language refers to the active foreign language. But in order to study
the A/B language choice while accounting for the influence of the source/target
language status, both directions of interpreting need to be considered, and that has been
achieved in only a few studies (e.g., Dam, 2004b; Szabó, 2006; Wang et al., 2010).
Dam (2004b) studied the notes taken by four students with the language
combination of Danish/Spanish (three students were Danish native speakers and one

75
was a Spanish native speaker). All her participants preferred the A language regardless
of the direction of interpreting, pointing to a tendency to choose the better-mastered
native language. Szabó (2006) had eight “quasi professionals” (p. 133) interpret
between Hungarian (A language) and English (B language), and all the participants
showed a preference for English, their B language, regardless of the direction of
interpreting. According to the questionnaire results, participants preferred English
because it was “morphologically less complex” and “more economical” (p. 142) than
Hungarian, indicating that the nature of the languages themselves played an important
role in interpreters’ language choice. Wang et al. (2010) studied student interpreters
with a language combination of Chinese (A language) and English (B language). They
found a source language dominance regardless of the direction of interpreting, and
inferred that this could have resulted from the participants’ inadequate interpreting
competence (p. 15).
The relationship between note-taking and interpreting performance is a key concern
in the teaching of interpreting. Scholars have looked at the relationship between
interpreting performance and such variables as the quality (Her, 2001) and quantity
(Cardoen, 2013; Dam, 2007; Dam, Engberg, & Schjoldager, 2005) of notes, but no
consistent conclusions have been reached. It would seem that the interactions between
note-taking and interpreting performance are more complex than imagined. A pilot
study by Orlando (2014) compared the performances of interpreters in traditional
consecutive interpreting and a new hybrid mode using digital pen. Results showed that
in the new mode, which he called “consec-simul with notes” (p. 41), the accuracy was
higher, and the number of disfluencies or hesitation phenomena was lesser. The digital
pen technology was, as a result, recommended for use in consecutive interpreting
training and practice.
Through this brief review of literature on note-taking in CI, it is not difficult to find
that although some general trends could be detected, such as a dominance of language
over symbol, there are also vast inconsistencies. The collected empirical evidence is
very limited to date. Many studies that are based on empirical data either use students as
participants (whose interpreting competence varies greatly), making the data “not
enough to generalise” (Gile, 2009, p. 179), or experiment on one interpreting direction
only, making the results difficult to compare.
More importantly, the studies are largely product-oriented. That is, they only look
at the product (i.e., the notes produced) without an in-depth analysis of the note-taking

76
process. An outstanding exception was Andres (2002), who used time-coded video to
analyse the time span between the moment a source speech unit was spoken (start of
sound) and the moment it was noted down (start of pen). She found that, when
interpreting from French (B language) into German (A language), the span was between
3 and 6 seconds, although on some occasions it reached as much as 10 seconds. The
method used by Andres, however, was to determine the start of note-taking by manually
checking a video recording, and the span was measured in seconds, leaving some
questions regarding the accuracy of the data.
What could then be a promising avenue for future research? Interpreting is deemed
a cognitively demanding task by many. As Gile (2009, p. 178) points out, “note-taking
is an area in which the concept of processing capacity can be useful.” If cognitive load
can be measured during the process of note-taking, some underlying principles might be
unveiled. Considering that discussions on measuring cognitive load in interpreting,
especially CI, are very limited (see Chen (2017) for a review and a proposal for
potential measurement techniques including pen recording), investigating the cognitive
load in note-taking seems important.
This study attempts to address some of the limitations in previous research by (1)
using professional interpreters as participants; (2) investigating both directions of
interpreting; (3) combining product analysis with the process research method of pen
recording; and (4) investigating the cognitive load in note-taking. There are four
research questions (RQs), of which the first three are concerned with the three main
variables investigated in literature. The aim is to present further empirical data and to
either confirm or challenge the previous findings. The fourth RQ pertains to what
additional information pen recording can contribute to the topic. The pen data are
viewed from a cognitive perspective, and the possibility of using the data as indicators
of cognitive load in note-taking and CI is investigated.
RQ1: What do interpreters prefer when choosing the form of note-taking: language
or symbol; abbreviation or full word?
RQ2: What do interpreters prefer when choosing the language of note-taking:
source or target language; A or B language?
RQ3: What is the relationship between note-taking and interpreting performance?
RQ4: Is there a relationship between the note-taking choices and cognitive load in
CI?

77
4.3 Method

As has been mentioned above, in order to make the data more generalizable, research
needs to be carried out on professional interpreters (preferably certified and experienced)
rather than student interpreters (whose interpreting competence is not yet mature). In
order to account for both the source/target language status, and the A/B language status,
both directions of interpreting need to be involved. In addition, the note-taking process
needs to be recorded. This study was carefully designed to meet those demands.

4.3.1 Participants

In this exploratory study, five participants were recruited. They were all certified as
“Professional Interpreter” by Australia’s National Accreditation Authority for
Translators and Interpreters (NAATI). Their working language combination is
Mandarin Chinese (A language) and English (B language). Four of them had a
postgraduate interpreting degree, and one attended an intensive interpreting training
course and obtained a bachelor’s degree majoring in interpreting. The participants, aged
between 25 and 36 (average 30.2), had worked as full-time or part-time interpreters for
three to seven years (average 5.4 years). The city they most frequently worked in was
Sydney, Australia. For those who were working as part-time interpreters, their other
job(s) involved regular use of both of their working languages (e.g., interpreter trainer).
An estimated number of occasions they had provided CI services in the past 12 months
ranged from 10 to 50 (average 29).

4.3.2 Apparatus

A digital pen and a tablet were used to record pen activities during note-taking. The
tablet used was the Cintiq 13HD produced by Wacom, and it was equipped with a
Wacom Pro Pen. It was a professional digital tablet targeting graphic designers,
developed to meet very high requirements on the precise control of pen strokes. The
system has an ergonomic design, with 2048 levels of pressure sensitivity and tilt
recognition, closely simulating natural writing and painting.
The Eye and Pen software was used to control the whole experiment procedure, and
to collect and analyse pen data. The experiment was programmed into the software,
which then controlled the procedures to avoid human error. The software can report, for
each pen stroke, when the pen tip touches the tablet surface, how it travels across the
tablet (distance and duration), and when it leaves the tablet. The spatial data are

78
reported in centimetres and the temporal data are reported in milliseconds. The note-
taking and interpreting process was also video-recorded. An additional audio recorder
was used to record the retrospective verbal reports (see Section 3.4).

4.3.3 Tasks

There were two CI tasks. Stimuli consisted of one Chinese and one English speech, both
of which were carefully created through a series of procedures to control for variance.
Firstly, two English video clips on similar topics were selected from the Internet
and transcribed by the author. The transcripts were then edited by an experienced
university lecturer (a native English speaker from Australia) with respect to length,
complexity and style of language, making them as comparable as possible. The edited
texts were analysed using CPIDR, a computer programme that could automatically
determine the propositional idea density, and the results showed that they were quite
similar in the number of propositions and words, as well as idea density (Table 4.1).

Table 4.1 Text analysis results

Proposition count Word count Idea density

Text 1 324 630 0.514
Text 2 321 631 0.509

Secondly, one of the texts (text 1) was translated by the author into her A language
(Chinese), and refined stylistically and grammatically by two Chinese-speaking editors
working at a local Chinese radio station. The editors were asked to make the script oral
and suitable for recording. They understood the requirements very well due to the
nature of their work (editing scripts for radio broadcasting).
Thirdly, the edited Chinese and English scripts were recorded into audio by a native
Mandarin Chinese speaker (a radio personality from the same radio station) and a native
Australian English speaker (the English editor) in professionally soundproofed studios.
The speakers were required to record the speeches as naturally as possible, while
maintaining steady speed. They were allowed to restart any sentence at any time when
necessary.

79
 Fourthly, the recorded speeches were imported into Audacity, a sound-editing
programme, for further refinement (e.g., cutting unfinished sentences, deleting
background noises). The speeches were both about five minutes long, each divided into
three segments (Table 4.2).

Table 4.2 A summary of the tasks

Segment length
Task Topic Length
1 2 3

Chinese to English How to purchase property in Australia 4m47s 1m10s 2m07s 1m30s

English to Chinese How to register a business in Australia 4m59s 1m18s 2m02s 1m39s

4.3.4 Procedures

The experiment consisted of three sessions:

Session I: practice. First, the participants were allowed sufficient time to write
freely on the tablet using the digital pen. Then, they listened to a short practice task,
took notes, and interpreted. The purpose of this step was to get the participants
familiarised with both the equipment and the experiment procedures.
Session II: interpreting. The participants first interpreted from Chinese to English.
They were allowed a short break if required, and then performed the second task from
English to Chinese.
Session III: cued retrospection. Immediately after the tasks, the participants were
provided with their notes for cued retrospection. They were asked to provide as much
information as they could remember about the note-taking process, including but not
limited to: what each note unit was; what it stood for; whether it was symbol or
language, and if language, whether it was abbreviation or full word, Chinese or English.
This is an important step because note-taking in CI is highly individualised, and the
handwriting of interpreters could sometimes be difficult for others to decipher.

4.3.5 Data and analysis

The data collected in this study are summarised in Table 4.3. The written notes were
analysed to reveal the interpreters’ choices of form and language. The distance, duration
and speed of pen, and the ear-pen span were used as indicators of the physical, temporal,
and cognitive demands of different note-taking choices. Both the notes and the
interpreting performance were evaluated by human raters, and analysed together with

80
the note-taking choice results. The qualitative data from retrospection4 provide an emic
perspective from the interpreters, enabling finer-grained analyses of the quantitative
data, and help to explain the observed results.

Table 4.3 Data used for analysis

Source Data
All written note units;
Pen recording The distance, duration, and speed of each pen stroke;
Ear-pen span
Video of the interpreting process;
Video recording
Audio of the target speech (the interpreting performance)
Retrospection Audio of verbal report

Score of notes;
Human evaluation
Score of interpreting performance

4.3.5.1 Categorisation of note units

Based on the interpreters’ retrospection, all written notes were categorised according to
their form and language (Figure 4.1). Each note unit was first put into one of the three
form categories: symbol, language and number. All language note units were further
categorised according to form as either abbreviation or full word, and according to
language as either Chinese or English.

Note unit

Symbol Language Number

Abbreviation Full word Chinese English

Figure 4.1 Categorisation of note units

4
The retrospective data in this study is mainly used to assist the researcher to create an accurate
interpretation and documentation of the written notes.

81
 The note categories and their definitions are specified in Table 4.4 following the
rules specified in Dam (2004a, 2004b). Dam’s rules catered to Danish and Spanish, so
adaptations were made where necessary to account for the language combination of
Chinese and English. For example, Chinese characters with very simple strokes are
sometimes used by interpreters as symbols.

Table 4.4 Categories and definitions of note units (Adapted from Dam (2004b, p. 6) and
Dam (2004a, p. 253))

Category Definition Examples

A full word is a Chinese or English word

Full word written in full, including words both with and “Problem(s)” and “问题”
without morphemes of inflection.

(1) “Prob.” / “prblm” for

An abbreviation consists of parts of the letters
“problem(s)”, and “问” for “问题”;
of a long English word, or part of the
characters of a long Chinese word, or the (2) “AU” for “Australia”, and “澳” for
phonetic spelling of a word, including: (1) “澳大利亚”;
Abbreviation real abbreviations (i.e., units in which only (3) “L&G” for “ladies and
part of a word is represented); (2) acronyms; gentlemen”, and “女&先” for “女士
(3) other short forms that cannot be
们先生们” (“L” , “G”, “女” and “先”
characterised either as real abbreviations or as
acronyms, but rather as something in between. will be categorised as abbreviations;
“&” will be categorised as a symbol)

(1) Signs like pluses and colons, lines,

A symbol is a representation of (1) the
arrows, drawings, etc.;
underlying meaning of a word or expression
rather than the actual word or expression; or (2) Letter “B” for “but”, “however”,
Symbol (2) the relationship(s) between two units. “on the other hand”, “although”, etc.;
Symbols are mostly pictorial, but they can (3) Chinese character “心” for “爱
also be a pair of letters, a single letter, or (part (love)”, “喜欢(like)”, “想要
of) a Chinese character. (wanting)”, “满意(satisfied)”, etc.

The combination of full words and

Language abbreviations. Further divided into Chinese
and English5.

Independent from language and symbol,

Number numbers are seen as a special category of
notes.

4.3.5.2 Calculation of the ear-pen span

The ear-pen span is defined as the time span between the moment a speech unit is heard
(end of sound) and the moment it is written down in notes (start of pen). It was

5
Unlike in Dam (2004a, 2004b), the author found no notes written in a third language or an unidentifiable
language.

82
calculated using the following steps. First, identifying correspondence between the
source speech and the notes. The content and meaning of each note unit (identified with
the help from retrospective reports) were checked to determine if there was a one-to-one
correspondence between the note unit and a source speech unit. The ear-pen span could
not be calculated for notes that did not correspond to specific source speech units (e.g.,
symbols indicating hidden links).
Second, determining the end of sound and the start of pen. For each note unit that
corresponded to a source speech unit, two points in time were determined: (1) the end of
sound of the source speech unit; and (2) the start of pen stroke. This was different from
what Andres (2002) did in her study, where the time lag was calculated from the start of
sound to the start of pen. The consideration was that a span calculated from the start of
sound would be heavily influenced by the length of the sound unit. To avoid that
influence, this study calculated the span from the end of sound to the start of pen. The
start of pen in time was automatically reported by the software in milliseconds. The end
of sound was determined by checking the sound waves of the source speech audio using
Audacity, also reported in milliseconds. The software kept an experiment log which
recorded the time that the source speech started to play, so for each note unit, the end of
sound and the start of pen could be pinpointed on the same timeline.
Third, calculating the span. The ear-pen span was calculated as “start of pen minus
end of sound”. It was usually positive, indicating that there was a lag between hearing a
source speech unit and noting it down. But on some rare occasions, the span was
negative, indicating that the interpreter started to write down the note before hearing the
end of a source speech unit, or even predicted an entire incoming unit.

83
4.3.5.3 Human evaluation

Both the notes and the interpreting performance were rated by two raters: the author and
a colleague. Both raters had previous experience of rating interpreter certification tests.
Rating the notes. Each note unit that has a one-to-one correspondence with the
source speech was rated. It was given a score of either 1 or 0. When a note successfully
represented a source speech unit and was correctly interpreted in the target speech, it
was scored 1. If it falsely represented the source speech, did not appear in the target
speech or was falsely interpreted in the target speech, it was scored 0. For example, if a
note unit was written as “invest” (standing for “investment”), and interpreted as
“investor” (because the interpreter could no longer identify which meaning it stood for),
it would be scored 0.
For each note unit raters were given the content, meaning, corresponding source
text unit, source text sentence, and target text sentence (both orthographic transcription).
The scores of all note units were added up, and divided by the total number of notes
being rated, thus forming the score of notes (i.e., percentage of notes correctly
interpreted). The scores given by two raters were averaged.
Rating the interpreting performance. The purpose of performance rating in this
study was quite different from those in interpreter education or testing. There was no
need to judge whether the performance reached certain standards, because all
participants were nationally accredited, experienced interpreters. The goal was to
differentiate the performances as finely as possible, so that the relationship between
note-taking and interpreting performance could be revealed. Considering that all the
participants were expected to give high-quality performance, it would be very difficult
to use holistic scores to differentiate the performances. A stringent rating system
therefore needed to be developed.
The criterion chosen for performance rating in this study was accuracy, a core
component of interpreting quality (e.g., Gile, 1999; Pöchhacker, 2002). Many
researchers have applied it as a yardstick to evaluate interpreting performance (e.g.,
Dam & Engberg, 2006; Gerver, 1969/2002; M. Liu & Chiu, 2009). It is also
“particularly relevant and central” to studies on note-taking because notes function as
memory triggers to ensure an accurate rendition (Dam & Engberg, 2006, p. 216).
The method used for performance rating in this study was a proposition-based one.
As has been mentioned, the two original English texts were analysed using CPIDR, and

84
the proposition count of the two were 339 and 321 respectively. Based on the
proposition analysis results, the texts were divided into scoring units. The rule was that
each unit contained an average of three propositions, and natural sentence breaks were
kept. The two raters first divided the units separately, and then discussed the
inconsistencies and reached agreement. The Chinese text (an edited translation of one of
the English texts) was divided following the units marked on the original English text,
and the raters discussed the units on the Chinese text and reached agreement. The final
number of scoring units were 101 in the Chinese to English task, and 112 in the English
to Chinese task.
The interpreting performances were transcribed orthographically by the author and
the target texts were provided to the raters for rating. The accuracy was determined by
checking how closely each scoring unit was matched by the target text. A score of 1 was
given when the meaning of a unit was correctly interpreted; otherwise a score of 0 was
given. Following the principles in M. Liu and Chiu (2009), added information was not
penalised and erroneous renderings of the same proposition were penalized only the
first time they appeared. A performance score was calculated as the percentage of
scoring units correctly interpreted. The two raters did a trial rating session individually
on some randomly chosen target texts (covering both tasks), discussed the
inconsistencies, and reached agreement. The raters then performed all ratings
independently. The final score was an average of the scores given by the two raters.

4.4 Results and discussion of note-taking choices and their relationship with
interpreting performance

This part reports findings that are directly related to previous literature, and answers the
first three RQs: interpreters’ preferred choice of form in note-taking; interpreters’
preferred choice of language in note-taking; the relationship between note-taking and
interpreting performance.

85
4.4.1 Choice of form and language

4.4.1.1 Choice between language and symbol

Descriptive statistics concerning the choice between language and symbol of each
participant in the two tasks are summarised in Table 4.5. The quantity of notes taken by
each participant varied, with participant 5 (P5) taking down the least (120 in task 1 and
112 in task 2), and P2 taking the most (233 in task 1 and 261 in task 2), indicating that
note-taking is a highly individualised activity. But the quantity of notes taken in the two
tasks was quite similar, both averaged across individuals (177 vs. 179) and within each
individual, indicating that the information density of the two tasks are well controlled.

Table 4.5 Distribution over form: language vs. symbol

Task 1: Chinese to English

Participant Total Language Symbol Number
1 197 96 (49%) 88 (45%) 13 (7%)
2 233 118 (51%) 99 (42%) 16 (7%)
3 164 79 (48%) 71 (43%) 14 (9%)
4 172 108 (63%) 52 (30%) 12 (7%)
5 120 79 (66%) 30 (25%) 11 (9%)
Avg. of Task 1 177 96 (54%) 68 (38%) 13 (8%)
Task 2: English to Chinese
Participant Total Language Symbol Number
1 188 113 (60%) 67 (36%) 8 (4%)
2 261 135 (52%) 118 (45%) 8 (3%)
3 164 88 (54%) 65 (40%) 11 (7%)
4 171 119 (70%) 45 (26%) 7 (4%)
5 112 70 (66%) 30 (27%) 11 (7%)
Avg. of Task 2 179 106 (59%) 65 (36%) 8 (5%)
Avg. across
178 100 (57%) 67 (37%) 11 (6%)
participants & tasks
Note: percentages may not add up to 100% due to rounding.

As can be seen from the table, there was a clear preference for language over
symbol (57% vs. 37% when averaged across participants and tasks), and the trend was
consistently reflected in all individual cases and in both directions of interpreting.

86
4.4.1.2 Choice between abbreviation and full word

There was a preference for abbreviation (34%) to full word (22%) averaged across
participants and tasks (Table 4.6). This trend was consistently reflected in both
directions, but not in all cases (the exceptions are P3 in task 2 and P5 in both tasks).

Table 4.6 Distribution over form: abbreviation vs. full word

Task 1: Chinese to English

Participant Abbreviation Full word
1 61 (31%) 35 (18%)
2 89 (38%) 29 (12%)
3 45 (27%) 34 (21%)
4 69 (40%) 39 (23%)
5 37 (31%) 42 (35%)
Avg. of Task 1 60 (34%) 36 (20%)
Task 2: English to Chinese
Participant Abbreviation Full word
1 69 (37%) 44 (23%)
2 95 (36%) 40 (15%)
3 39 (24%) 49 (30%)
4 70 (41%) 49 (29%)
5 37 (33%) 37 (33%)
Avg. of Task 2 62 (35%) 44 (24%)
Avg. across
61 (34%) 40 (22%)
participants & tasks
Note: percentages do not add up to 100% because the rest are symbols and numbers.

87
4.4.1.3 Choice of language

The distribution of notes over language (Table 4.7) shows that the participants as a
group preferred B language (English, accounting for 36%) to A language (Chinese,
accounting for 20%). This preference was consistent in both tasks (i.e., both interpreting
directions), and in most participants (with the only exception of P4 in task 1). The trend
was stronger in task 2 (18% Chinese vs. 41% English) than in task 1 (23% Chinese vs.
31% English), showing that when the source language and B language coincided, the
preference for B language was strengthened.

Table 4.7 Distribution over language

Task 1: Chinese to English

Participant Chinese English
1 44 (22%) 52 (26%)
2 50 (21%) 68 (29%)
3 32 (20%) 47 (29%)
4 73 (42%) 35 (20%)
5 5 (4%) 74 (62%)
Avg. of Task 1 41 (23%) 55 (31%)
Task 2: English to Chinese
Participant Chinese English
1 48 (26%) 65 (35%)
2 41 (16%) 94 (36%)
3 9 (5%) 79 (48%)
4 42 (25%) 77 (45%)
5 18 (16%) 56 (50%)
Avg. of Task 2 32 (18%) 74 (41%)
Avg. across
36 (20%) 65 (36%)
participants & tasks
Note: percentages do not add up to 100% because the rest are symbols and numbers.

4.4.2 Relationship between note-taking and interpreting performance

As we can see in Table 4.8, neither the score of notes nor the quantity of notes alone
could explain the variances in performance:

88
Table 4.8 Relationship between note-taking and interpreting performance

Task 1: Chinese to English

Participant Score of performance Score of notes Quantity of notes
1 79.21 87.27 197
2 89.61 92.29 233
3 81.19 87.80 164
4 67.83 85.57 172
5 74.76 89.82 120
Task 2: English to Chinese
Participant Score of performance Score of notes Quantity of notes
1 87.95 95.00 188
2 92.86 94.48 261
3 68.30 83.93 164
4 69.20 88.41 171
5 69.65 92.22 112

Rather, performance seemed to be a function of both the quality and quantity of notes.
For example, P1 had high note counts, but low note scores, and his/her performance
ranked in the middle. P5 had high note scores, but influenced by his/her low note counts,
his/her performance also ranked in the middle. P2 had both high note counts and high
note scores, and his/her performance ranked the highest.

4.4.3 Discussion

Some tentative answers could be suggested for the first three RQs. It needs to be noted
that the answers are based on empirical results found on a small sample of professional
interpreters working between the language combination of Chinese and English.
RQ1: What do interpreters prefer when choosing the form of note-taking: language
or symbol; abbreviation or full word?
Interpreters in our study showed a clear preference for language over symbol. This
finding corroborates previous studies, using either student interpreters (Dai & Xu, 2007;
Dam, 2004b; J. Liu, 2010; Lung, 2003; Wang et al., 2010) or professional interpreters
(Andres, 2002; Dam, 2004a) as participants. The interpreters preferred abbreviation to
full word, a finding corroborating some studies (Dai & Xu, 2007; Wang et al., 2010),
but contradicting the findings of others (Dam, 2004a; J. Liu, 2010; Lung, 2003). The
contradiction could be caused by such factors as the nature of the language pair, the
type of participants used, or the texture or genre of the source speech (Setton &

89
Dawrant, 2016, p. 211), but there is not enough empirical evidence to pinpoint the cause
at the moment.
RQ2: What do interpreters prefer when choosing the language of note-taking:
source or target language; A or B language?
The interpreters showed a preference for English (their B language) over Chinese
(their A language), and this preference was strengthened when the B language and the
source language coincided. That is to say, the interpreters opted for a language that is
weaker in their language combination, a choice intuitively implausible. Sifting through
the retrospective reports, it was found that in many cases, the interpreters chose English
for note-taking because it was easier and faster to write than Chinese characters. What
also needs to be noted is that the interpreters in this study are based in Australia, an
English-speaking country, and they are likely to have a very strong B language.
The result relating to the choice of language in this study contradicts what Wang et
al. (2010) found in student interpreters with the same language combination, where a
strong preference for source language was detected regardless of the direction of
interpreting. It also contradicts what Dam (2004b) found in students with the
Danish/Spanish language combination, where a strong preference for the A language
was found, indicating a tendency to choose the better-mastered native language. It is in
line with what Szabó (2006) found in professional interpreters with the
Hungarian/English language combination. Szabóobserved a preference for English, the
B language, regardless of the direction of interpreting, and pointed to the morphological
complexity of Hungarian and the economy of writing in English as an explanation.
Szabó also mentioned that the participants had a very strong B language, as is the case
in this study.
Based on the above discussion, some conclusions could be suggested on the choice
of language in note-taking. The language choice is a function of the combined influence
of a series of factors, including: (1) the nature of the languages themselves (e.g.,
morphological complexity and economy of writing); (2) the A/B language status; (3) the
source/target language status; and (4) interpreter characteristics (e.g., working
experience and language competence).
When two languages do not differ too much in morphological complexity or
economy of writing (the case in Dam (2004b)), the A/B language status plays a major
role in determining the language choice, and interpreters are more likely to use their A
language for note-taking. When one language is morphologically simpler or easier and

90
faster to write (the case in Szabó (2006) and this study), this language would be the
preferred choice regardless of the A/B language status, especially when the interpreter
has a strong B language. When the interpreter lacks experience (the case in Wang et al.
(2010)), the language choice is subject mainly to the source/target language status in a
task.
The empirical data collected so far are insufficient to identify how the factors
interact, and what their respective and combined influences are on the choice of
language. These are interesting directions for future research.
RQ3: What is the relationship between note-taking and interpreting performance?
With a small sample size of five, it is difficult to draw any concrete conclusions
using the data in this explorative study. However, it would seem that the interpreting
performance is subject to variances in both the quality and quantity of notes. The
following are tentative explanations. The quality of notes is based on two levels of
equivalence (between source speech/notes and between notes/target speech). For all
notes to faithfully represent the source speech and be successfully rendered in the target
speech, an interpreter needs to allocate enough cognitive capacity to activities such as
listening/analysing and memorising. This would sometimes lead to a decrease in the
amount of notes that can be written down, reducing the amount of information that can
be stored in notes. A good interpretation is related to the concurrent storage of
information both in notes and in memory. That is why sometimes we could observe a
set of notes with high score but low quantity to be associated with a middle-ranking
performance. Previous studies have also detected potential relationships between
performance and the quality (Her, 2001) and quantity (Dam, 2007) of notes. But the
interactions between the variables and their individual and combined influences on
performance remain unclear with the available data. Further empirical evidence needs to
be gathered before the mechanism could be revealed.

91
4.5 Results and discussion of the pen data: potential indicators of cognitive load
in CI

This part reports on data collected via pen recording. Different note-taking choices are
compared on the distance, duration, and speed of pen, as well as the ear-pen span. The
data are examined from a cognitive perspective, with an attempt to answer the last RQ:
RQ4: Is there a relationship between the note-taking choices and cognitive load in
CI?

4.5.1 Pen data on the choice of form

4.5.1.1 Between language, symbols, and numbers

Consistent differences are found between language and symbol notes in terms of the
distance, duration, and speed of pen, and the ear-pen span (Table 4.9). The average
distance and duration of language notes (7.17 cm and 1256.13 ms respectively) were
much longer than those of symbol notes (2.99 cm and 367.48 ms respectively), and the
writing speed of symbol (9.14 cm/s) was faster than that of language (6.04 cm/s). That
is to say, compared to language, symbols are easier and faster to write. The ear-pen span
of symbol (3039.33 ms) was longer than that of language (2504.99 ms), indicating it
took longer for interpreters to transfer a source speech unit into symbol than into
language notes.
Interestingly, the distance, duration, and speed of pen of numbers all lie between
those of language and symbol, but the ear-pen span of numbers (1428.31 ms) was much
shorter than both. This means that, after hearing a number, the participants would take
very swift responses and write it down, about one second faster than language and 1.5
seconds faster than symbols.

92
Table 4.9 Pen data on the choice of form

Task 1: Chinese to English

Language Symbol Number Abbreviation Full word
Distance (cm) 7.17 3.21 4.97 6.48 8.14
Duration (ms) 1237.56 379.04 796.99 1094.46 1433.94
Speed (cm/s) 6.20 9.13 6.64 6.22 6.21
Ear-pen span (ms) 2620.10 2980.78 1682.21 2412.34 2925.95
Task 2: English to Chinese
Language Symbol Number Abbreviation Full word
Distance (cm) 7.17 2.77 5.24 6.03 8.66
Duration (ms) 1274.70 355.93 1021.87 1088.51 1527.51
Speed (cm/s) 5.88 9.15 6.10 5.80 5.91
Ear-pen span (ms) 2389.87 3097.89 1174.41 2436.28 2356.79
Averaged across participants and tasks
Language Symbol Number Abbreviation Full word
Distance (cm) 7.17 2.99 5.10 6.25 8.40
Duration (ms) 1256.13 367.48 909.43 1091.48 1480.72
Speed (cm/s) 6.04 9.14 6.37 6.01 6.06
Ear-pen span (ms) 2504.99 3039.33 1428.31 2424.31 2641.37

4.5.1.2 Between abbreviations and full words

Consistent differences are found between abbreviation and full word notes in terms of
the distance and duration of pen (Table 4.9). The average distance and duration of pen
of abbreviations (6.25 cm and 1091.48 cm respectively) were shorter than those of full
words (8.40 cm and 1480.72 cm respectively), indicating that abbreviations were easier
to write, but the speed of pen was similar (6.01 cm/s for abbreviations and 6.06 cm/s for
full words).
In task 1, the average ear-pen span of abbreviations (2424.31 ms) was shorter than
that of full words (2641.37 ms), and this difference was consistent in all participants.
However, no consistent trend could be detected in task 2.

4.5.2 Pen data on the choice of language

The distance, duration, and speed of pen showed no consistent difference between the
language choices. But the ear-pen span (Table 4.10) of notes in A language was longer
than B language in almost all cases (except for P1 in task 2). That is to say, after hearing
a source speech unit (no matter in what language), it takes longer before the participants
write down a Chinese note than an English one.

93
Table 4.10 Ear-pen span data on the choice of language

Task 1: Chinese to English

Participant Chinese English
1 2137.35 2105.78
2 2489.00 2390.92
3 2774.55 2179.00
4 3104.04 3055.06
5 3808.80 2983.53
Avg. of Task 1 2862.75 2542.86
Task 2: English to Chinese
Participant Chinese English
1 1855.02 2717.71
2 2245.11 2142.57
3 2648.63 1584.61
4 2969.07 2476.84
5 3227.28 3035.06
Avg. of Task 2 2589.02 2391.36
Avg. across participants & tasks 2725.88 2467.11

4.5.3 Interpreting the findings from a cognitive load perspective

No matter what choices interpreters make during note-taking, the basic question, as Gile
(2009, p. 178) points out, is “how to reduce processing capacity and time requirements
of note-taking while maintaining the efficiency of notes as memory reinforcers”. On the
cognitive side, since interpreting is a highly demanding task, an important goal of
interpreters’ skills and strategies is to save cognitive effort. On the physical and
temporal side, the physical effort and time cost associated with note-taking are of great
concern to consecutive interpreters (Alexieva, 1994). Therefore we have good reasons
to infer that, for professional interpreters with sufficient experience, their overall
choices should reflect a balanced weighting of the physical, temporal and cognitive
demands of note-taking.
In this study, the distribution data showed that interpreters preferred language (57%)
to symbol (37%), abbreviation (34%) to full word (22%), and English (36%) to Chinese
(20%) during note-taking. We would like to make the bold hypothesis that the overall
physical, temporal and cognitive demands associated with different note-taking choices
for Chinese interpreters working between English and Chinese is: language lower than

94
symbol, abbreviation lower than full word, and English lower than Chinese, regardless
of the direction of interpreting.
The pen data of distance and duration could be straightforward indicators of the
physical effort and temporal cost associated with the note-taking choices. Notes that
induce lower demands should be those with shorter pen distance and duration, meaning
the pen tip travels a shorter distance and for a shorter period of time. According to our
results, the distance and duration of language and full words are longer than those of
symbols and abbreviations respectively, suggesting that the use of symbols and
abbreviations could reduce physical and temporal demands. This finding corroborates
the note-taking principles proposed by many (e.g., Alexieva, 1994; Gillies, 2005;
Schweda-Nicholson, 1993). No clear difference is found in the physical and temporal
demands between Chinese and English notes, suggesting that the choice of language
does not significantly affect the physical or temporal demand of note-taking.
The ear-pen span data are potentially indicative of the cognitive load in note-taking.
Since interpreting is an externally paced task, high cognitive load tends to increase the
time lag, causing participants to “lag farther and farther behind the input” (Treisman,
1965, p. 378). In our study, the ear-pen span results were: symbol longer than language,
full word longer than abbreviation, and Chinese longer than English. Assuming the ear-
pen span is an indicator of cognitive load (longer span means higher load), then the
cognitive load associated with different note-taking choices are: language lower than
symbol, abbreviation lower than full word, and English lower than Chinese.
If we put the two pieces of the puzzle together (Table 4.11), we can see how the
physical, temporal and cognitive demands act together to affect interpreters’ note-taking
choices. It would seem that physical and temporal demands do not affect note-taking as
much as cognitive load. In particular, despite their lower physical and temporal
demands, symbols are used less than language by interpreters in note-taking.

95
Table 4.11 How physical, temporal and cognitive demands affect interpreters’ note-
taking choices

Form Language
Language vs. symbol Abbreviation vs. full word Chinese vs. English
Physical and
Symbol < Language Abbreviation < Full word Chinese ≈ English
temporal demands
Cognitive
Language < Symbol Abbreviation< Full word English < Chinese
load
Note-taking
Language Abbreviation English
preference

4.6 Conclusions

This study investigates note-taking in CI in terms of the choice of form and language,
and the relationship between note-taking and interpreting performance. It reports new
data from pen recording, interprets the data from a cognitive perspective, and presents
preliminary findings on the relationship between note-taking and cognitive load.
It was found that, firstly, interpreters preferred language to symbol, abbreviation to
full word, and English to Chinese, regardless of the direction of interpreting. Secondly,
the interpreting performance seemed to be subject to variances in both the quality and
quantity of notes. Thirdly, the physical and temporal demands of different note-taking
choices, as indicated by the pen data of distance and duration, appeared to be: language
higher than symbol, full word higher than abbreviation, and Chinese similar to English.
Fourthly, the cognitive load induced by different note-taking choices, as indicated by
the ear-pen span, appeared to be: symbol higher than language, full word higher than
abbreviation, and Chinese higher than English.
On the whole, pen recording is found to be a powerful method to tap into the
process of note-taking and CI. The data collected can provide us with an accurate and
encompassing picture of the interpreting process with moment-to-moment changes in
pen position reported in coordinates. The data also appear to be useful indicators of
cognitive load. Although the digital pen and tablet system used in this study is
particularly useful for research purposes, it is not recommended for application in
training or practice. There are other types of digital pens which are less powerful in data
collection but much easier to use in classrooms and field interpreting (see Orlando,
2010, 2014).

96
 It has to be admitted that the empirical data collected in this study are very limited.
The sample size is small, and only one language combination (Chinese and English) is
involved, confining the generalizability of the findings. But at the same time, the
limitations have pointed to some interesting directions for future research. For example,
can the findings be replicated with a larger sample size? Can the same results be
reached using a different language combination? The author will continue to seek
answers to these questions, and hopefully they will attract the interests from other
researchers as well.

4.7 References

Abuín González, M. (2012). The language of consecutive interpreters’ notes:

Differences across levels of expertise. Interpreting, 14(1), 55–72.
doi:10.1075/intp.14.1.03abu
Albl-Mikasa, M. (2008). (Non-) sense in note-taking for consecutive interpreting.
Interpreting, 10(2), 197–231. doi:10.1075/intp.10.2.03alb
Alexieva, B. (1994). On teaching note-taking in consecutive interpreting. In C. Dollerup
& A. Lindegaard (Eds.), Teaching translation and interpreting 2: Insights, aims,
visions (pp. 199–206). Amsterdam: John Benjamins.
Andres, D. (2002). Konsekutivdolmetschen und notation [Consecutive interpreting and
note-taking]. Frankfurt: Peter Lang.
Cardoen, H. (2013). The Effect of Note-taking on Target-text Fluency. In G. González
Núñez, Y. Khaled, & T. Voinova (Eds.), Emerging Research in Translation
Studies: Selected Papers of the CETRA Research Summer School 2012 (pp. 1-
22). Leuven: CETRA.
Chen, S. (2016). Note-taking in consecutive interpreting: A review with special focus
on Chinese-English literature. The Journal of Specialised Translation, 26, 151–
171.
Chen, S. (2017). The construct of cognitive load in interpreting and its measurement.
Perspectives. doi:10.1080/0907676X.2016.1278026
Dai, W., & Xu, H. (2007). 汉英交替传译过程中译员笔记特征实证研究——以职业
受训译员和非职业译员为例 [An empirical study of the features of interpreters'
notes in Chinese-English consecutive interpreting: The examples of

97
 professionally trained and unprofessional interpreters]. Foreign Language
Teaching and Research, 39(2), 136–144.
Dam, H. V. (2004a). Interpreters' notes: On the choice of form and language. In G.
Hansen, K. Malmkjær, & D. Gile (Eds.), Claims, changes and challenges in
Translation Studies (pp. 251–261). Amsterdam: John Benjamins.
Dam, H. V. (2004b). Interpreters' notes: On the choice of language. Interpreting, 6(1),
3–17. doi:10.1075/intp.6.1.03dam
Dam, H. V. (2007). What makes interpreters’ notes efficient?: Features of (non-
)efficiency in interpreter's notes for consecutive. In Y. Gambier, M. Shlesinger,
& R. Stolze (Eds.), Doubts and directions in Translation Studies: Selected
contributions from the EST Congress, Lisbon 2004 (pp. 183–197). Amsterdam:
John Benjamins.
Dam, H. V., & Engberg, J. (2006). Assessing accuracy in consecutive interpreting: a
comparison of semantic network analyses and intuitive assessments. In C. Heine,
K. Schubert, & H. Gerzymisch-Arbogast (Eds.), Text and translation: Theory
and methodology of translation (pp. 215–234). Tübingen: Narr Francke
Attempto Verlag.
Dam, H. V., Engberg, J., & Schjoldager, A. (2005). Modelling semantic networks on
source and target texts in consecutive interpreting: A contribution to the study of
interpreters’ notes. In H. V. Dam, J. Engberg, & H. Gerzymisch-Arbogast (Eds.),
Knowledge systems and translation (Vol. 7, pp. 227–254). Berlin: Mouton de
Gruyter.
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. 53–66). London and New
York: Routledge.
Gile, D. (1991). Prise de notes et attention en début d’apprentissage de l’interprétation
consécutive–une expérience– démonstration de sensibilisation [Note-taking and
attention at the beginning of consecutive interpreting learning –an experience–
demonstration of awareness]. Meta, 36(2–3), 431–439.
Gile, D. (1999). Variability in the perception of fidelity in simultaneous interpretation.
Hermes, 22, 51–79. doi:10.7146/hjlcb.v12i22.25493
Gile, D. (2009). Basic concepts and models for interpreter and translator training
(revised edition). Amsterdam: John Benjamins.

98
Gillies, A. (2005). Note-taking for consecutive interpreting: A short course. Manchester
and Kinderhook: St. Jerome.
Her, H. (2001). Notetaking in basic interpretation class: An initial investigation. Studies
of Translation and Interpretation, 6, 53-77.
Herbert, J. (1952). Manuel de l'interprète: comment on devient interprète de
conférences [Interpreter's manual: How to become a conference interpreter].
Geneva: Georg.
Jones, R. (1998). Conference interpreting explained. Manchester: St. Jerome.
Kirchhoff, H. (1979). Die notationssprache als hilfsmittel des konferenzdolmetschers im
konsekutivvorgang [The language of note-taking as a tool for the conference
interpreter in consecutive interpreting]. In W. Mair & E. Sallager (Eds.),
Sprachtheorie und sprachpraxis [Language theory and language practice] (pp.
121-133). Tübingen: Gunter Narr.
Kohn, K., & Albl-Mikasa, M. (2002). Note-taking in consecutive interpreting: On the
reconstruction of an individualised language. Linguistica Antverpiensia, 1, 257-
272.
Liu, J. (2010). 英语专业本科学生汉英交传笔记特征——一项基于学生交传笔记的
实证研究 [Note-taking characteristics of English majored undergraduates in
Chinese-English consecutive interpreting: an empirical study based on students'
consecutive interpreting notes]. Foreign Language World, 2, 47–53.
Liu, M., & Chiu, Y.-H. (2009). Assessing source material difficulty for consecutive
interpreting: Quantifiable measures and holistic judgment. Interpreting, 11(2),
244–266. doi:10.1075/intp.11.2.07liu
Lung, R. (2003). Taking "notes" seriously in the interpretation classroom. In Á.
Collados Aís, M. M. Fernández Sánchez, & D. Gile (Eds.), La evaluación de la
calidad en interpretación: investigación [The evaluation of quality in
interpretation: Research] (pp. 199–205). Granada: Comares.
Matyssek, H. (1989). Handbuch der notizentechnik für dolmetscher [Handbook of note-
taking for interpreters]. Heidelberg: Julius Groos.
Orlando, M. (2010). Digital pen technology and consecutive interpreting: Another
dimension in note-taking training and assessment. The Interpreters’ Newsletter,
15, 71–86.

99
Orlando, M. (2014). A study on the amenability of digital pen technology in a hybrid
mode of interpreting: Consec-simul with notes. Translation and Interpreting,
6(2), 39–54.
Pöchhacker, F. (2002). Researching interpreting quality: Models and methods. In G.
Garzone & M. Viezzi (Eds.), Interpreting in the 21st century: Challenges and
opportunities (pp. 95–106). Amsterdam: John Benjamins.
Rozan, J.-F. o. (1956/2002). Note-taking in consecutive interpreting (A. Gillies, Trans.).
Cracow: Tertium Society for the Promotion of Language Studies.
Schweda-Nicholson, N. (1993). An introduction to basic note-taking skills for
consecutive interpretation. In E. Losa (Ed.), Keystones of communication:
Proceedings of the 34th Annual Conference of the American Translators
Association (pp. 197–204). Medford: Learned Information.
Seleskovitch, D. (1975). Langage, langues et mémoire: étude de la prise de notes en
interprétation consécutive [Speech, language, and memory: A study of note-
taking in consecutive interpreting]. Paris: Minard Lettres Modernes.
Setton, R., & Dawrant, A. (2016). Conference interpreting: A trainer's guide.
Amsterdam: John Benjamins.
Szabó, C. (2006). Language choice in note-taking for consecutive interpreting.
Interpreting, 8(2), 129–147. doi:10.1075/intp.8.2.02sza
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, W., Zhou, D., & Wang, L. (2010). 口译笔记特征与口译产出质量实证研究
[An empirical study of note-taking characteristics and output quality in
interpreting]. Foreign Language World, 4, 9–18.

100
An introductory note to Chapter 5

Findings in Chapter 4 allow specific hypotheses to be realised about the form and
language of note-taking and its relationship with interpreting performance. With these
hypotheses available, the PhD research moves on to the next level to carry out a main
study involving a larger sample of participants. The main study implemented the full-
fledged design of the PhD project, triangulating the methods of pen recording, eye
tracking and voice recording to explore the cognitive processing and cognitive load in
note-taking and CI.
Chapter 5 reports the pen and voice recording data of the main study, focusing on
Phase I of CI in which interpreters listen to the source speech and write notes. Issues
investigated include the content of notes, the timing of note-writing, and the choices of
form and language in note-taking as well as their associated cognitive load. Most of the
findings in the pilot study (Chapter 4) are successfully replicated in the main study.
Several trends in the seemingly individualised note-taking choices are also unveiled.
Discussions on how the findings in this study add to the literature on note-taking and CI
and pedagogical suggestions are provided on the basis of the empirical data.

101
Chapter 5 The process of note-taking and consecutive interpreting:
Evidence from digital pen recording1

Abstract: This article reports the findings of an empirical study on the process of note-
taking and consecutive interpreting (CI). Combining digital pen recording with video
recording, the note-taking and interpreting process is recorded in detail and
automatically synchronised with the source speech. The study seeks answers to four
questions about CI and note-taking: (1) What do interpreters note down? (2) When do
interpreters take notes? (3) How do interpreters take notes? (4) Why do interpreters
make certain choices in note-taking? Data that have been collected include the note-
taking and interpreting process (e.g., all note units; details of the pen movement which
reveal information on the physical, temporal and cognitive demands of different note-
taking choices) and the interpreting performance. By triangulating the different types of
data, some general trends in the process of note-taking and CI were detected, and some
fundamental principles guiding interpreters’ note-taking choices were unveiled. This
forms the basis for a number of pedagogical recommendations.
Keywords: note-taking; consecutive interpreting; process research; cognitive load;
digital pen recording

5.1 Introduction

The field of translation and interpreting process research has seen significant
development over the past years and in particular, the focus on the cognitive aspects of
translation and interpreting is increasing. This is evident in the growing number of
volumes that have been published in the last two decades (e.g., Ehrensberger-Dow,
Dimitrova, Hubscher-Davidson, & Norberg, 2015; Martín, 2016; Shreve & Angelone,
2010; Tirkkonen-Condit & Jääskeläinen, 2000). Researchers have benefited from an
interdisciplinary effort, which resulted in the application of many methods that are new
to the field, such as verbal reporting, key logging, video and screen recording and eye
tracking.
However, in the field of interpreting studies, consecutive interpreting (CI) has
received very limited attention so far. To look into the process of CI, note-taking serves

1
This chapter is under review in the journal of Interpreting as: Chen, S. (under review). The process of
note-taking and consecutive interpreting: Evidence from digital pen recording.

102
as a good starting point. It is a unique characteristic of CI2 and can reveal important
information about the cognitive processes. Consecutive note-taking has been a recurring
topic in the field, creating fruitful results over the past half a century. But most of the
studies have been product-oriented, without collecting and analysing process data. A
potentially important reason for this is a lack of existing methods to tap into the process
of CI within the field of interpreting research. Nevertheless, there are a handful of
studies that have taken a process-oriented approach to note-taking and CI. Two
outstanding examples would be Andres (2002) who used video recording and Orlando
(2010) who used the Livescribe Smartpen. However, neither method could provide
comprehensive data on the entire process. For example, video recording involves
determining the start of note-taking by manually checking the video and its timestamp
and the Smartpen does not report the moment-to-moment change in pen position in
coordinates.
In writing research, there is a new pen recording method which involves the use of
a software called the Eye and Pen3. With the software, a digital tablet and a digital pen,
any writing process can be recorded in fine details (with time data reported in
milliseconds and position data reported in centimetres). This study applies this method
to investigate the process of note-taking and CI. The software is programmed to control
the play of the source speech (hence listening) and align it with the note-taking process
(for more details please see Section 3.2). With this synchronous recording of the
multiple tasks happening in the interpreting process, this study explores the what, when
and how questions about CI and note-taking, and tries to find an explanation to why
interpreters make certain note-taking choices.

5.2 Research background

5.2.1 A review of studies on note-taking in CI

There are two main streams of note-taking literature: a prescriptive stream and a
descriptive stream (for a detailed review please see Chen (2016)). In the prescriptive
stream a number of books and articles put forward various note-taking systems and
principles. Prescriptive suggestions are given on what, when and how to take notes,
usually starting from the authors’ experiences in the profession and/or in teaching.

2
In this article, CI refers to long consecutive in which systematic note-taking is used.
3
More introduction to the software can be found on http://www.eyeandpen.net/?lng=en.

103
 The first and foremost principle in any note-taking system is to take notes based on
speech analysis. This is described by Gillies (2005) as identifying the “Subject Verb
Object” unit. Things that are recommended to be noted down are (1) those to cue the
interpreter and ensure an efficient rendering of the speech, including the main ideas,
links, verb tenses and modal verbs; and (2) those to relieve memory, including numbers,
dates, proper names and lists (e.g., Gillies, 2005; Jones, 1998; Rozan, 2002).
Slightly different suggestions have been given on when to take notes. Some
recommend starting to take notes as quickly as possible (Jones, 1998, pp. 61-63); others
suggest waiting until the subject of a speech is fully understood (Lung, 1999, p. 311);
still others suggest that notes can be taken either sooner or later according to the
circumstances (Gillies, 2005, pp. 156-158). It is generally agreed that crucial details
(especially those difficult to store in memory) such as names, numbers and dates should
be noted down as soon as possible and that interpreters should not be bound by the
original order of things in the source speech. They are free to change the order to
coordinate note-taking and memory (e.g., Gillies, 2005, p. 159; Jones, 1998, p. 63).
Discussions on how to take notes are generally focused on the use of symbols and
abbreviations, as well as the language of note-taking. The use of symbols is an essential
part of any established note-taking system. Symbols are easy to write and read, and
because they represent concepts rather than words, they can help to avoid source
language influence (Gillies, 2005, p. 99). Although the amount of exemplary symbols
given by different authors varies (for example Rozan (2002) on the minimalist end and
Matyssek (1989) on the maximalist end), there is general agreement that fully-mastered,
unambiguous symbols are very useful. The use of abbreviations is suggested for noting
long words. This could be done by writing the first and last letters, preferably with the
latter written as superscript (Matyssek, 1989, p. 115; Rozan, 2002, pp. 16-18; Schweda-
Nicholson, 1993, p. 200), and by using phonetic spelling and misspelling (Gillies, 2005,
p. 162). As to the language of note-taking, authors tend to not give any specific
suggestions, although sometimes a slight preference is given to either the source
language (e.g., Alexieva, 1994; Ilg, 1988) or the target language (e.g., Rozan, 2002;
Seleskovitch, 1975; Seleskovitch & Lederer, 1989/1995).
It would seem that the note-taking systems are well-developed, and once students
are made aware of them and practice accordingly, note-taking should not be a problem.
However, when it comes to the teaching and learning of note-taking, both the teachers
and the students find it challenging. The problem has been documented in a number of

104
studies (e.g., Alexieva, 1994; Gile, 1991), in which researchers found that note-taking
could divert the attention of student interpreters and could even lead to a degradation in
interpreting performance. Studies that describe how notes are actually taken by student
interpreters represent the beginning of a shift from prescriptive to descriptive
approaches in note-taking literature. This shift is strengthened by researchers who
approach the topic from cognitive and linguistic perspectives.
Motivated by an attempt to theorise note-taking and CI, some researchers started
investigations on the cognitive and linguistic aspects of note-taking. Studies by
pioneering scholars such as Kirchhoff (1979) and Seleskovitch (1975) and the more
recent ones by Albl-Mikasa (2006, 2008) have pointed to a concurrent storage of
information in notation texts and in memory, as well as a competition for resources
between note-taking and other activities in the interpreting process.
Unlike the early descriptive studies which have a general interest in what real notes
look like, and set out to discover some overall trends, the latest studies on note-taking
usually have more specific targets. They investigate different note-taking choices,
collect data in simulated interpreting tasks, and have contributed some valuable
empirical evidence. In this group of studies, variables that have attracted the most
attention include the choice of form, the choice of language, and the relationship
between note-taking and interpreting performance.
The choice of form refers to the choice between language and symbol and the
choice between abbreviation and full word. Overall, studies found a dominance of
language over symbol but reached inconsistent conclusions as to whether interpreters
preferred abbreviation or full word (Chen, 2016).
The choice of language refers to the choice between source and target language and
the choice between L1 and L24. It has been found that the notes of student interpreters
are largely source-language dominated whereas the notes of professional interpreters are
more varied. While some studies found that professionals preferred L1 (e.g., Abuín
González, 2012; Dam, 2004a), others found that they preferred L2 such as English in
Szabó(2006).
Research on the relationship between note-taking and interpreting performance
appears to be even more inconclusive. Some studies claim to have found evidence

4
In this article, L1 refers to the native language and L2 refers to the non-native language. They are called
A and B language in some studies.

105
pointing to a potential link between certain note-taking features and the quality of
interpreting (Cardoen, 2013; Dam, 2007; Dam, Engberg, & Schjoldager, 2005; Her,
2001); others have failed to detect any such relationship (Dai & Xu, 2007).
Most of these studies are product-oriented, meaning they only look at the notes
produced without an in-depth analysis of the interpreting process. An outstanding
exception to this was Andres (2002), who used time-coded video to analyse the time lag
between listening and note-taking in CI. She found that, when interpreting from French
(L2) into German (L1), the lag was between 3 and 6 seconds, although on some
occasions it reached as much as 10 seconds. Since the time lag in CI refers to the span
between the moment a source speech unit is heard and the moment it is written down, it
could be called the ear-pen span (EPS). It is similar to the ear-voice span (EVS) in
simultaneous interpreting, which has been studied by many (e.g., Barik, 1973;
Christoffels & De Groot, 2004; Gerver, 1969/2002; Goldman-Eisler, 1972). EVS and
EPS can provide rich information about the cognitive processing in both simultaneous
(Timarová, Dragsted, & Hansen, 2011; Treisman, 1965) and consecutive (Chen, 2017)
interpreting. In this study, EPS will be carefully analysed and used as an indicator of
cognitive load in CI.
On the whole, the amount of empirical data that has been collected on the topic is
still limited. Studies that do collect empirical data are usually limited in several aspects:
the participants, the method and technology used, the stimulus task, and the statistical
analysis. Many of the studies used students as participants, rendering the data
insufficient to allow for generalisation (Gile, 2009, p. 179). Studies that have involved
professional interpreters usually have small sample sizes (less than 15 participants). The
method and technology used cannot capture the note-taking and CI process with enough
details, leading to a lack of multiple strands of empirical evidence to elucidate the same
process. The stimulus task often involves one interpreting direction only, making it
difficult to compare the results and to consider the variation brought by directionality.
More importantly, the specifics of the stimulus task (such as length and speed) are
simply not reported in many studies, so it is impossible to replicate the experiment. The
recorded empirical data are usually descriptive and only demonstrate some general
trends (e.g., a preference for L1 or L2) but on many occasions, no significance testing
has been performed. All these limitations have confined our ability to understand CI and
note-taking in more depth.

106
5.2.2 Research questions

Considering the limitations in the previous studies, this study is designed to address
some of the challenges. It uses professional interpreters as participants and combines the
process research method of pen recording with product analysis on both the notes and
the interpreting performance. It involves two directions of interpreting and performs
significance testing on all the results. With these attempts, the study will try to answer
four questions about note-taking and CI.
First, what do interpreters note down? This question looks at the linguistic features
of interpreters’ notes in relation to the source text. Attempts will be made to analyse
which source text elements are noted down and whether the findings conform to what
has been suggested in the literature.
Second, when do interpreters take notes? This question examines how interpreters
coordinate the parallel tasks in the first phase of CI, namely listening/analysis and note-
taking. Special attention will be paid to how far interpreters lag behind the source
speech in CI and to what extent they engage in multi-tasking.
Third, how do interpreters take notes? This question is concerned with the choice of
form (whether interpreters prefer language or symbol, abbreviation or full word) and the
choice of language (whether interpreters prefer the source or target language, L1 or L2)
in note-taking.
Fourth, why do interpreters make certain note-taking choices? The physical,
temporal and cognitive demands associated with the different note-taking choices will
be compared to seek possible explanations for interpreters’ note-taking preferences. The
interpreting performance will also be examined to see if there is a relation between
note-taking choices and the quality of interpreting.

5.3 Method

5.3.1 Participants

The study involves 26 professional interpreters. They were paid for their participation.
They all had NAATI Professional Interpreter accreditation, with a working language
combination of Mandarin Chinese (L1) and English (L2). Most of them had a
postgraduate degree in interpreting (65%); some had an interpreting diploma (15%);
some attended an intensive interpreting training course (15%); and one was self-trained
(4%) . With an average age of 36.4, the participants had worked as full-time or part-

107
time interpreters for an average of 7.4 years. The country they most frequently worked
in was Australia (with only two exceptions who worked more frequently in China). For
those who were working as part-time interpreters, their other job(s) had a bilingual
feature (e.g., interpreter trainer). The number of occasions they had provided CI
services in the past 12 months was averaged at 167.
The participants used a digital pen and a tablet for note-taking. A post-hoc
questionnaire was used to ask whether the interpreters felt comfortable with the digital
pen and tablet. If the rating was too low (lower than 4 in a 1-7 scale rating 5 ), the
participant was excluded from the data analysis. The final pen recording data came from
22 participants.

5.3.2 Apparatus

The digital pen used in the study was the Wacom Pro Pen and the tablet was the Cintiq
13HD. The system is ergonomically designed to mimic natural writing and painting,
targeting graphic designers who have very high requirements on the precise control of
the pen on the tablet surface. It was linked to a laptop computer piloted by Eye and Pen
software. The experiment procedure was programmed into the software, which
controlled the experiment and interacted with the participants. For example, the
software would play a new segment of the source speech when participants indicated
that the interpretation was completed (by clicking a button displayed on the screen with
the pen). It could also create as many new pages as needed when participants listened to
the source speech and took notes. The software was also responsible for reporting the
moment-to-moment position and state of the pen tip on the tablet and collecting the data.
The interpreting process was video-recorded. An additional audio recorder was used to
record the retrospective verbal reports (see Section 3.4).

5.3.3 Tasks

A series of procedures were carried out to create two comparable source speeches. First,
two English scripts (1 and 2) on similar topics were created and edited by an Australian
university lecturer with respect to length, complexity and style of language. The
resulting scripts were put into a programme called CPIDR 5 for analysis, and results

5
The question was: “On a scale from 1 (strongly disagree) to 7 (strongly agree), do you agree that the
digital tablet and pen are sufficiently similar to real pen and paper, and therefore did NOT affect your
note-taking behaviour?”

108
showed that they were comparable in word count, proposition count and idea density6
(Table 5.1). Second, script 2 was translated into Chinese, refined stylistically and
grammatically by two Chinese editors working at a local Chinese radio station. Third,
an English speech was recorded using script 1 by a native English speaker in Australia
(the same person who edited the original English scripts). A Chinese speech was
recorded by a native Chinese (Mandarin) speaker (a radio personality from the radio
station mentioned above) using the Chinese version of script 2. Both recordings were
made in professional sound-proof studios. Fourth, the recordings were edited using
Audacity. Each speech was divided into three segments and controlled for variables
such as pauses, duration and speed (Table 5.1).

Table 5.1 A summary of the two tasks

Word/character Proposition Idea Segment duration

Task Topic Duration
count count density 1 2 3
How to register
E-C a business in 631 321 0.509 4'59" 1'18" 2'02" 1'39"
Australia
How to purchase 630 (English) *
C-E a property in 324* 0.514* 4'47" 1'10" 2'07" 1'30"
Australia 944 (Chinese)

* The calculations were based on the original English script to enable comparison between the two tasks.

5.3.4 Procedures

First, participants were allowed sufficient time to practice using a short task. During this
practice session they get familiarised with the digital pen and tablet and the experiment
procedures. Second, the participants performed the CI tasks, the order of which was
randomised. Third, after both tasks were completed, the participants were provided with
their written notes for cued retrospection. During the retrospection, the participants told
the experimenter as much as they could remember about the note-taking process,
including but not limited to: what each note unit was; what it stood for; whether it was
symbol or language, and if language, whether it was abbreviation or full word, Chinese
or English. This is an important step because note-taking in CI is highly individualised
and interpreters’ notes could be difficult for others to identify.

6
More details about the programme and the analysis can be found at http://ai1.ai.uga.edu/caspr/ and
Brown, Snodgrass, Kemper, Herman, and Covington (2008).

109
5.3.5 Data and analysis

Data in this study came from several sources: pen recording, video recording, human
evaluation, and verbal report. Pen recording produces all the written note units (see
3.5.1) and details of the pen movement (see 3.5.2 and 3.5.4). The video recording of the
interpreting process was used to produce transcripts of the interpreting performance and
was provided to a group of raters for evaluation (see 3.5.4). The verbal report produced
via cued retrospection provided additional subjective data about the interpreting process
and served as a basis for categorising the note units (see 3.5.3).

5.3.5.1 Tagging the source speech

All note units with a one-to-one correspondence with the source speech (identified with
the help of retrospective reports) were mapped onto the source speeches, which were
analysed using part-of-speech (POS) tagging. The POS tags are those of the Penn
Treebank (Santorini, 1990; Xia, 2000). The most important ones in this study are:
adjective, adverb, conjunction, determiner, noun, number, preposition, pronoun, proper
noun, verb, localiser and measure word (for the Chinese speech only), and modal (for
the English speech only). The data can reveal which elements in the source speech have
a higher frequency of being noted down and whether the empirical findings conform to
the prescriptive literature. For example, to see if interpreters give priority to numbers,
each source speech element tagged as “number” is checked to see the percentage of
interpreters who have noted it down. This percentage is then averaged across all
elements tagged as numbers in the source speech, producing the frequency of numbers
to be noted down by all interpreters in the task. If the frequency is very high, it is proof
that when deciding what to note, interpreters do give priority to numbers.

5.3.5.2 Calculating the EPS

The source speech and the note-taking process were mapped onto the same timeline.
The EPS was calculated as the time span between the moment a speech unit is heard
and the moment it is written down in notes. As has been mentioned earlier, Andres
(2002) used time-coded video to analyse the time span in CI. She calculated the span
from the start of sound to the start of pen, and determined the time points by manually
checking the videos. But this raises the concern that the data would be influenced by the
length of the source speech unit and that the accuracy and objectivity of the data could
be affected by human error. To improve the situation, the measurement points in this

110
study were from the end of sound to the start of pen using data automatically reported
by software.
Since the EPS can only be calculated for those note units that have a one-to-one
correspondence with the source speech, the number of notes included in EPS analysis is
smaller than the total number of notes. Moreover, time lag data is sensitive to extreme
values, as has been pointed out by researchers who studied EVS in simultaneous
interpreting (Oléron & Nanpon, 1965/2002, p. 47; Timarová et al., 2011, p. 142). In
note-taking, extreme values are often observed in cases such as note additions at the end
of a source speech segment. Therefore, for the analysis of EPS, all observations that
were three standard deviations below or above the mean for each condition (a certain
participant in a certain task) were excluded from analysis. In the end, about 90% of the
notes were included in the EPS analysis.

5.3.5.3 Categorising the notes

Each written note was first categorised according to form as symbol, language, or
number7. Then, if a note unit was labelled as language, it was further categorised as
either abbreviation or full word (according to form) and as either Chinese or English
(according to language). The categories and their definitions followed the rules
specified in Dam (2004a, 2004b), and adaptations were made where necessary to
account for the language pair of Chinese and English. For example, Chinese characters
with a few simple strokes are sometimes used by interpreters as symbols.

5.3.5.4 Detailed pen recording data and the interpreting performance

For each note unit, the experiment software recorded the distance of pen (how far the
pen moved across the surface, reported in centimetres) and the duration of pen (for how
long the pen moved, reported in milliseconds). The distance and duration of pen were
used as indicators of the physical and temporal demands of different note-taking choices.
Notes that induce lower demands should be those with shorter pen distance and duration,
meaning the pen tip moves a shorter distance and for a shorter period of time. The EPS
was used as an indicator of the cognitive load. Since interpreting is an externally paced
task, high cognitive load tends to increase the time lag, causing participants to “lag

7
Numbers are put into an independent category because they are often treated with special care by
interpreters, which is evident both in literature (see Section 2) and from the results in this study (see
Section 4.2).

111
farther and farther behind the input” (Treisman, 1965, p. 378). Notes that induce lower
cognitive load should be those with shorter EPS.
To explore the relationship between note-taking and interpreting performance, a
group of three raters evaluated the performances. The purpose of evaluation is to
differentiate the performances as finely as possible, so that the relationship between
note-taking and interpreting performance could be revealed. What is needed here is a
stringent rating system to differentiate the performances of professional interpreters.
Accuracy is a core component of interpreting quality (e.g., Gile, 1999; Pöchhacker,
2002) and is applied in many studies to quantify interpreting performance (e.g., Dam &
Engberg, 2006; Gerver, 1969/2002; Liu & Chiu, 2009).
For studies on note-taking, accuracy is “particularly relevant and central” because
the purpose of note-taking is to ensure an accurate interpretation (Dam & Engberg,
2006, p. 216). Therefore, the current study used a proposition-based rating method,
using accuracy as the only criterion. With the proposition analysis results provided by
CPIDR 5, the Chinese speech was divided into 101 scoring units, and the English
speech was divided into 112 scoring units, with each scoring unit containing an average
of three propositions. Raters were given the source and target texts in parallel to each
other with the scoring units marked onto them. Each unit was given a score of either 1
or 0 depending on how closely the target text unit matches the source. The percentage
of units correctly interpreted was used as the final score to indicate the interpreting
performance. More details about how the scoring units are determined can be found in
(Chen, 2017).
Each interpreting performance was rated by all three raters. A high degree of
reliability was found between the raters. The average measures intraclass correlation
coefficient (ICC) in the E-C task was .95 with a 95% confidence interval from .69 to
.98; F(21, 42) = 52.2, p < .001. The ICC in the C-E task was .94 with a 95% confidence
interval from .87 to .98; F(21, 42) = 20.7, p < .001. The mean of the three ratings was
used as the score of the interpreting performance.

112
5.4 Results

5.4.1 What do interpreters note down

A paired-samples t-test was conducted to compare the total number of note units taken
down in the two tasks. No significant difference was found between the E-C task (M =
182.1, SD = 27.7) and the C-E task (M = 179.3, SD = 27.6); t(21) = 0.90, p = .380. This
shows that on average participants took similar amounts of notes in both tasks. A further
step was taken to see what proportion of the source speech has been put into notes. All
note units that have a one-to-one correspondence with the source speech (about 90% of
the notes) were selected and mapped onto the source speech. There was no difference
between the quantities of notes selected in the two tasks. On average, the interpreters
put about one third of the source speech elements in notes, but there was a difference
between the two tasks. The proportion of source speech noted down in the E-C task was
27.4% (SD = 4.16%), significantly lower than the 33.8% that was noted down in the C-
E task (SD = 5.0%); t(21) = -9.12, p < .001, d = 1.948.
The frequencies of different categories of source speech units to be noted down by
interpreters are shown in Table 5.2. Results show that the frequency of numbers
(including dates) being noted down is the highest in both tasks (over 95%), followed by
proper nouns, adjectives, and nouns (between 50% and 70%). Adverbs, verbs, and
conjunctions in both tasks, along with localisers and pronouns in the C-E task, were
noted down in between 15% and 40% of the frequencies. Noted down at the lowest
frequencies (below 10%) were: determiners and prepositions (in both tasks), modal
verbs and pronouns (in the E-C task), and measure words (in the C-E task).
Additionally, each task contains two lists. The frequency of each list being noted down
was 100% (although not all participants noted down the complete list).

8
Cohen's d is the effect size used in this article to compare between two means. It is calculated as the
difference between the means divided by the pooled standard deviation. Effect sizes are classified as
small (d = 0.2), medium (d = 0.5), and large (d = 0.8).

113
Table 5.2 Categories of the source speech units and their frequencies of being noted
down

E-C task C-E task

Number 98.4% Number 97.9%
Proper noun 68.4% Proper noun 69.2%
Adjective 58.4% Adjective 48.1%
Noun 50.9% Noun 43.6%
Adverb 38.1% Localiser 37.1%
Verb 21.6% Verb 25.5%
Conjunction 14.7% Conjunction 23.3%
Preposition 9.0% Adverb 16.8%
Modal 2.4% Pronoun 15.9%
Pronoun 1.9% Determiner 8.2%
Determiner 0.8% Preposition 6.1%
Measure word 5.2%

5.4.2 When do interpreters take notes

Interpreters were engaged in simultaneous listening (and analysis) and note-writing for
about the same percentages of time in the E-C task (M = 56.9%, SD = 6.08%) and the
C-E task (M = 56.4%, SD = 5.27%); there was no significant difference between the
tasks, t(21) = 0.37, p = .715. That is to say, regardless of the direction of interpreting,
interpreters spent about 60% of the time on parallel processing of multiple tasks in the
first phase of CI.
The time lag between listening and note-writing, namely the EPS, was 2447
milliseconds averaged across the two directions of interpreting (2262 milliseconds in
the E-C task and 2632 milliseconds in the C-E task). The EPS of numbers, which were
treated as one special category of notes (see Section 3.5.3), was of particular interest.
The EPS of number notes in the E-C and C-E task was 1005 milliseconds (SD = 964)
and 1794 milliseconds (SD = 583) respectively, both significantly shorter than the
average EPS. This means that interpreters reduced the EPS when they come across
numbers and started taking notes as soon as they heard one.

5.4.3 How do interpreters take notes

On the choice of form between language and symbol, interpreters preferred language to
symbol in both tasks (Table 5.3). In the E-C task, interpreters took 75.2% notes in

114
language, significantly higher than the 19.8% in symbol. In the C-E task, interpreters
took 68.6% notes in language, also significantly higher than the 22.5% in symbol. On
the choice of form between abbreviation and full word, no significant difference was
found between the distributions of the two note forms in either task (Table 5.3).

Table 5.3 The choice of form

Language Symbol
M SD M SD t df d
75.2% 7.34% 19.8% 7.59% 17.42*** 21 3.71
E-C
Abbreviation Full word
M SD M SD t df d
39.6% 10.2% 35.5% 12.0% 0.91 21
Language Symbol
M SD M SD t df d
68.6% 8.11% 22.5% 8.54% 13.1*** 21 2.79
C-E
Abbreviation Full word
M SD M SD t df d
37.6% 7.27% 31.0% 10.4% 1.94 21
*** p < .001

On the choice of language, interpreters preferred English to Chinese in both tasks

(Table 5.4). In the E-C task, interpreters took 16.4% notes in Chinese, significantly less
than the 58.8% in English. In the C-E task, interpreters took 26.7% notes in Chinese,
also significantly less than the 41.9% in English. There were also differences between
the two tasks. The percentage of Chinese notes in the E-C task (16.4%) was
significantly lower than that in the C-E task (26.7%), t(21) = -2.60, p = .017, d = 0.55;
the percentage of English notes in the E-C task (58.8%) was significantly higher than
that in the C-E task (41.9%), t(21) = 4.16, p < .001, d = 0.89. That is to say, interpreters’
choice of language (preferring English over Chinese) was affected by the direction of
interpreting, and to be more specific, by the source/target status of the languages. When
English and the source language coincided, the preference for English was strengthened,
but when English and the source language contradicted (when Chinese was the source
language), this preference was weakened.

115
Table 5.4 The choice of language

Chinese English
Task M SD M SD t df d
E-C 16.4% 13.4% 58.8% 15.2% -7.15*** 21 1.52
C-E 26.7% 16.4% 41.9% 16.5% -2.24* 21 0.48
* p < .05, *** p < .001

5.4.4 Why do interpreters make different note-taking choices

5.4.4.1 The physical, temporal and cognitive demands of different note-taking choices

Significant differences were found between language and symbol notes in terms of the
distance and duration of pen and the EPS in both directions of interpreting (Table 5.5).
The distance and duration of language notes were significantly longer than those of
symbol notes, and EPS of language notes was significantly shorter than that of symbol
notes. This shows that the physical and temporal demands of language notes were
significantly higher than those of symbols, but the cognitive load associated with
language notes was significantly lower than symbols. The fact that interpreters preferred
language over symbol and that the cognitive load associated with language notes was
lower than symbol notes (even though the physical and temporal demands of the former
were higher) seemed to indicate that the choice between language and symbol was
governed by the cognitive demand (rather than the physical or temporal demands).

Table 5.5. A comparison of language and symbol notes in terms of distance, duration
and EPS

Language Symbol
Task Pen data M SD M SD t df d
Distance (cm) 7.87 2.13 3.18 1.12 13.3*** 21 2.84
E-C Duration (ms) 1214 209 378 63 20.0*** 21 4.26
EPS (ms) 2282 610 2694 991 -3.41** 21 0.73
Distance (cm) 8.14 2.28 3.10 1.14 12.3*** 21 2.61
C-E Duration (ms) 1236 218 365 91 17.4*** 21 3.70
EPS (ms) 2653 695 3047 741 -4.33*** 21 0.92
** p < .01, *** p < .001

Significant differences were found between abbreviation and full word notes in
terms of the distance and duration of pen in both directions of interpreting, but not in the
EPS (Table 5.6). The distance and duration of abbreviation notes were significantly

116
shorter than those of full word notes, but the EPS was similar between the two. That is
to say, the physical and temporal demands of abbreviations were significantly lower
than those of full words, but the cognitive load was similar. The fact that interpreters did
not have a preference between abbreviation and full word and that the levels of
cognitive load associated with the two forms of notes were similar (even though the
physical and temporal demands of the former were lower) seemed to indicate that the
choice between abbreviation and full word was also governed by the cognitive demand
(rather than the physical or temporal demands).

Table 5.6 A comparison of abbreviation and full word notes in terms of distance,
duration and EPS

Abbreviation Full word

Task Pen data M SD M SD t df d
Distance (cm) 6.74 1.92 9.17 2.54 -7.27*** 21 1.55
E-C Duration (ms) 1068 212 1384 232 -6.61*** 21 1.41
EPS (ms) 2338 633 2203 629 1.62 21
Distance (cm) 7.12 1.93 9.40 2.86 -6.57*** 21 1.40
C-E Duration (ms) 1080 178 1423 289 -6.78*** 21 1.45
EPS (ms) 2660 737 2627 700 0.43 21
*** p < .001

No significant difference was found between Chinese and English notes in terms of
the distance and duration of pen in either direction of interpreting, but the EPS of
Chinese notes was significantly longer than that of English notes in the C-E task (but
not in the E-C task) (Table 5.7). This means that the physical and temporal demands of
different language choices were similar. However, the cognitive load associated with
Chinese notes was significantly higher than that with English notes in the C-E task. It
would seem that in L1-L2 interpreting, the choice of language was determined by the
cognitive demand (rather than the physical or temporal demands) associated with
different language choices. But in L2-L1 interpreting, the choice of language was
determined by the source/target language status and was dominated by the source
language.

117
Table 5.7 A comparison of Chinese and English notes in terms of distance, duration and
EPS

Chinese English
Task Pen data M SD M SD t df d
Distance (cm) 7.61 1.91 8.04 2.22 -1.24 21
E-C Duration (ms) 1199 444 1189 187 0.13 21
EPS (ms) 2368 766 2213 576 1.68 21
Distance (cm) 8.10 2.81 8.22 2.17 -0.35 21
C-E Duration (ms) 1204 245 1242 218 -0.75 21
EPS (ms) 2859 857 2611 669 2.71* 21 0.58
* p < .05

5.4.4.2 Note-taking and interpreting performance

The relationship between interpreting performance and note-taking was investigated

using the Pearson’s correlation (Table 5.8). The performance was positively correlated
with the total number of note units taken in the C-E task but not in the E-C task. There
was a positive correlation between the interpreting performance and the percentage time
spent on note-taking (i.e., engaged in simultaneous listening/analysis and note-writing)
in both tasks. The performance was negatively correlated with the EPS in the E-C task
but not in the C-E task.
The performance was also correlated with the distribution of notes, i.e., the
percentage of notes in different categories. In the E-C task, there was a negative
correlation between the performance and the percentage of language notes and a
positive correlation between the performance and the percentage of symbol notes, but
no significant results was found in the C-E task. There was a negative correlation
between the interpreting performance and the percentage of English notes in both the E-
C task and the C-E task. No significant correlation was found between the performance
and the percentage of Chinese, abbreviation, and full word notes.

Table 5.8 The Pearson’s correlation between interpreting performance and note-taking

Quantity Time spent on Percentage of notes

Task EPS
of notes note-taking Language Symbol English
E-C .52* -.44* -.55** .56** -.48*
C-E .48* .50* -.43*
N = 22, * p < .05, ** p < .01

118
 As we can see, the two directions of interpreting shared some common trends. A
higher percentage of time spent on multi-tasking (listening/analysing and note-taking at
the same time) was related to a better interpreting performance, but a higher percentage
of English notes was related to a worse interpreting performance. However, there were
also some differences between the two directions. In L2-L1 interpreting, the interpreting
performance was better when the EPS was shorter, when interpreters used more symbol
notes and when they used less language notes. But in L1-L2 interpreting, the quality of
performance was better when interpreters took more notes.

5.5 Discussions

5.5.1 The content of note-taking

The empirical data in this study supports most (but not all) of what has been suggested
in literature on what should be noted down (see Section 2). According to the results,
interpreters give top priority to lists (100% noted down), numbers and dates (almost
100% noted down), and proper names (about 70% noted down). In the “Subject Verb
Object” unit, the frequency of nouns (usually subjects or objects) being noted down is
about 50%; the frequency of verbs being noted down is about 20%. The frequency of
conjunctions (usually links) being noted down is also about 20%.
Standing in contrast to what has been recommended in literature, verb tenses and
modal verbs are rarely noted down (less than 5%). This is likely to due to the features of
the specific language pair of Chinese and English. Marking verb tenses is not always
necessary (especially in the E-C direction) because the form of a Chinese verb never
changes; rather, tenses in Chinese are inferred from context or marked by temporal
words (Liu, 2008, p. 69). As to modal verbs, the Chinese modals are functionally weak
compared with English, and Chinese does not differentiate all the meanings associated
by the English modals (Li, 2010, p. 359). These differences between Chinese and
English could probably explain why the data in this study partly contradicts the
prescriptive studies, many of which do not involve Chinese.
When it comes to interpreter education, students should be trained to pay special
attention to lists, numbers and proper names, and try to note down these elements as
completely as possible. Students should also be trained to analyse the “Subject Verb
Object” structures in the source speech, locate the main ideas and note them down.
Teachers should also be careful with the unique features brought by specific language

119
pairs and make pedagogical adaptations accordingly, rather than strictly follow the
established note-taking systems.

5.5.2 The timing of note-taking

On average, interpreters lag behind the source speech for about 2.45 seconds during the
listening and note-taking phase of CI. This is shorter than what Andres (2002) found in
her study, which was between 3 and 6 seconds. The difference could be caused by the
method of calculating the span. Andres calculated the span from the start of sound and
manually determined the start of note-taking by checking video recordings. This study
calculated the EPS from the end of sound, and the start of note-taking was automatically
reported by software. Another possible cause for the difference is the language pair. In
studies on simultaneous interpreting, the reported EVS in interpreting between different
languages pairs varies (e.g., Barik, 1973; Lee, 2002; Oléron & Nanpon, 1965/2002;
Timarováet al., 2011; Treisman, 1965).
One particularly interesting finding is that interpreters note down numbers
exceptionally quickly. The EPS data shows that numbers are noted down about 1
second faster than language notes and about 1.5 seconds faster than symbols. It would
seem that in order to avoid cognitive overload (numbers are a well-known problem-
trigger in interpreting (Alessandrini, 1990; Cheung, 2008)), interpreters opted for a
strategy to shorten the time lag and lower the cognitive load when they come across
numbers.
Since the data in this study is gathered from professional interpreters, it could serve
as a reference as to how much interpreters should lag behind the source speech in note-
taking. The time lag found in students’ interpretations could be checked against the data
to get an idea of whether students are following the source speech too closely or lagging
too far behind. In particular, students should be instructed to keep the time lag short and
take down notes as quickly as possible when it comes to numbers.

5.5.3 The choice of form and the choice of language

Interpreters showed a clear preference for language over symbol, corroborating the
previous studies (Andres, 2002; Dai & Xu, 2007; Dam, 2004a, 2004b; Lung, 2003). No
preference was found between abbreviation and full word. Previous studies did not
perform significance testing on their data, but the descriptive data showed a preference
for abbreviations in some studies (e.g., Dai & Xu, 2007) but full words in others (e.g.,

120
Dam, 2004a; Lung, 2003). The inconsistent findings could be related to such factors as
the language pair and the type of participants used (professional vs. student interpreters),
but there is not enough empirical evidence to draw conclusions.
Interpreters preferred English (L2) in both directions of interpreting, and this
preference was stronger when English was the source language. The result contradicts
what Dam (2004b) found in students with the Danish/Spanish language pair, in which
all participants preferred their L1, the better mastered native language. The result is in
line with what Szabó (2006) found in professional interpreters with the
Hungarian/English language pair. Szabó observed a preference for English (L2),
regardless of the direction of interpreting, and pointed to the morphological complexity
and economy of writing as an explanation.
A post hoc questionnaire of this study shows that interpreters prefer to use English
in note-taking mainly because English can be written down using phonetic spelling and
even misspelling, whereas the written form of Chinese is not always available
immediately. The participants find it easier to take notes in English. This has to do with
the fact that English is an alphabetic language but Chinese is not. However, it should
also be noted that the participants in this study are based in Australia, an English-
speaking country, and they are likely to have a very strong L2, same as the participants
in Szabó’s study.
Findings in this study and various previous studies show that the choice of language
in note-taking is a function of the combined influence from a series of factors, including:
(1) the language combination itself (e.g., alphabetic/non-alphabetic nature;
morphological complexity); (2) the L1/L2 language status; (3) the source/target
language status; and (4) interpreter characteristics (e.g., working experience and
language competence).

5.5.4 Note-taking, cognitive load and interpreting performance

The time needed for writing, the effort of the hand and the mental effort costs are
believed to be the main concerns in making note-taking choices (e.g., Alexieva, 1994,
pp. 203-204; Allioni, 1989, p. 195; Gile, 2009, p. 178). But the data in this study
suggests that the temporal and physical demands are less of a concern to interpreters
than the cognitive demand. This is evident in the different note-taking choices made by
interpreters (see Table 5.9). Although the physical and temporal demands of language
notes are higher than those of symbols, the cognitive demand of language notes is lower

121
and interpreters preferred language to symbol. The cognitive demands of abbreviation
and full word notes are similar and interpreters do not have a preference between the
two, even though abbreviations are easier and faster to write. Chinese and English notes
are similar in their physical and temporal demands, but the cognitive load associated
with English is lower and interpreters prefer English in the C-E task9 (although it is the
non-source and non-native language).
However, the relation between note-taking and interpreting performance seems to
show that the choices made by interpreters to lower cognitive load in the listening and
note-taking phase of CI are at the expense of interpreting quality (see Table 5.9). A
better interpreting performance is correlated with a lower percentage of language notes
and a higher percentage of symbol notes (only the E-C task reached significance); it is
also correlated with a lower percentage of English notes (but not with Chinese notes).

Table 5.9 The physical, temporal and cognitive demands and interpreters' preferred
note-taking choices

Choice of form Choice of language

Language vs. symbol Abbreviation vs. full word Chinese vs. English
Physical and
Language > Symbol Abbreviation < Full word Chinese ≈ English
temporal demands
Chinese > English
Cognitive load Language < Symbol Abbreviation ≈ Full word
(C-E only)
Preferred note-
Language No preference English
taking choice
Less language, more Less English, better
Relation with
symbol, better No relation performance; no
performance
performance (E-C only) relation with Chinese

Considering the positive contribution of symbols to interpreting performance, more

emphasis should be put on symbols in CI training. The aims should be two-fold: to
decrease the cognitive costs of symbols and to increase their usage. Instead of merely
pointing to the highly individualised nature of symbols in note-taking and leaving it to
the students to develop a system of their own (usually painstaking and involving a lot of
trial and error), teachers could lend a set of symbols rich enough for the students to start
with. The students could then spend more time on practising using the symbols instead
of inventing them.

9
In the E-C task there is no difference in cognitive load and the language choice seems to be determined
by the source language status of English.

122
 Some important differences exist between the two directions of interpreting. In the
L2-L1 direction, the performance is negatively correlated with the EPS, whereas in the
L1-L2 direction, the performance is positively correlated with the quantity of the notes.
This suggests that when interpreting from the non-native language, the time lag matters
more. Because the cognitive resources needed for listening and analysis in the non-
native language is higher, interpreters need to maintain a shorter EPS to ensure that they
do not overload their cognitive system. But when interpreters work from the native
language, the cognitive demands of listening and analysis is lower. Therefore, they can
afford to allocate more resources to note-taking and take down more notes, which in
turn stores more information and facilitates a complete production.
The findings on directionality could inform teachers to take different pedagogical
approaches to the two directions of interpreting and remind students to develop different
strategies to deal with the specific challenges.

5.6 Conclusion

This study contributes a process-oriented approach towards note-taking and CI. It

makes a methodological contribution by showing how digital pen recording can be
applied in interpreting research. Pen recording is found to be a powerful tool to tap into
the cognitive processes in note-taking and CI. The data not only provides us with an
accurate and encompassing picture of the interpreting process but also serves as an
indicator of cognitive load.
The study conducts a comprehensive investigation on the note-taking choices,
cognitive load and interpreting performance in CI. Some fundamental principles
underlying the interpreters’ note-taking choices (which seem highly individualised at a
first glance) are unveiled and pedagogical suggestions are made based on empirical data.
It has to be admitted that this study is limited in several ways. Only one language
pair (Chinese and English) is involved, so the findings may not generalise to another
language combination. Most of the participants live and work in a country where
English (their L2) is spoken, and the results may not generalise to another type of
interpreters, such as Chinese-English interpreters based in China. Only one group of
participants is involved in the study. This design does not allow the investigation of the
role interpreting experience plays in the process. It would be interesting to see how
different groups of participants with varying degrees of experience differ in the note-

123
taking and CI process. These limitations have pointed to some interesting directions for
future research. It is hoped that more researchers will be joining the effort to create a
more comprehensive picture of note-taking and CI.

5.7 References

Abuín González, M. (2012). The language of consecutive interpreters’ notes:

Differences across levels of expertise. Interpreting, 14(1), 55–72.
doi:10.1075/intp.14.1.03abu
Albl-Mikasa, M. (2006). Reduction and expansion in notation texts. In C. Heine, K.
Schubert, & H. Gerzymisch-Arbogast (Eds.), Text and translation: Theory and
methodology of translation (pp. 195–214). Tübingen: Narr.
Albl-Mikasa, M. (2008). (Non-) sense in note-taking for consecutive interpreting.
Interpreting, 10(2), 197–231. doi:10.1075/intp.10.2.03alb
Alessandrini, M. S. (1990). Translating numbers in consecutive interpretation: An
experimental study. The Interpreters' Newsletter, 3, 77–80.
Alexieva, B. (1994). On teaching note-taking in consecutive interpreting. In C. Dollerup
& A. Lindegaard (Eds.), Teaching translation and interpreting 2: Insights, aims,
visions (pp. 199–206). Amsterdam: John Benjamins.
Allioni, S. (1989). Towards a grammar of consecutive interpretation. In L. Gran & J. M.
Dodds (Eds.), The theoretical and practical aspects of teaching conference
interpretation (pp. 191–197). Udine: Campanotto.
Andres, D. (2002). Konsekutivdolmetschen und notation [Consecutive interpreting and
note-taking]. Frankfurt: Peter Lang.
Barik, H. C. (1973). Simultaneous interpretation: Temporal and quantitative data.
Language and Speech, 16(3), 237–270. doi:10.1177/002383097301600307
Brown, C., Snodgrass, T., Kemper, S. J., Herman, R., & Covington, M. A. (2008).
Automatic measurement of propositional idea density from part-of-speech
tagging. Behavior Research Methods, 40(2), 540–545.
doi:10.3758/BRM.40.2.540
Cardoen, H. (2013). The Effect of Note-taking on Target-text Fluency. In G. González
Núñez, Y. Khaled, & T. Voinova (Eds.), Emerging Research in Translation
Studies: Selected Papers of the CETRA Research Summer School 2012 (pp. 1-
22). Leuven: CETRA.

124
Chen, S. (2016). Note-taking in consecutive interpreting: A review with special focus
on Chinese-English literature. The Journal of Specialised Translation, 26, 151–
171.
Chen, S. (2017). Note-taking in consecutive interpreting: New data from pen recording.
Translation and Interpreting, 9(1), 4–23. doi:10.12807/ti.109201.2017.a02
Cheung, A. K. F. (2008). Simultaneous interpreting of numbers: An experimental study.
Forum, 6(2), 23–38. doi:10.1075/forum.6.2.02kfc
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
Dai, W., & Xu, H. (2007). 汉英交替传译过程中译员笔记特征实证研究——以职业
受训译员和非职业译员为例 [An empirical study of the features of interpreters'
notes in Chinese-English consecutive interpreting: The examples of
professionally trained and unprofessional interpreters]. Foreign Language
Teaching and Research, 39(2), 136–144.
Dam, H. V. (2004a). Interpreters' notes: On the choice of form and language. In G.
Hansen, K. Malmkjær, & D. Gile (Eds.), Claims, changes and challenges in
Translation Studies (pp. 251–261). Amsterdam: John Benjamins.
Dam, H. V. (2004b). Interpreters' notes: On the choice of language. Interpreting, 6(1),
3–17. doi:10.1075/intp.6.1.03dam
Dam, H. V. (2007). What makes interpreters’ notes efficient?: Features of (non-
)efficiency in interpreter's notes for consecutive. In Y. Gambier, M. Shlesinger,
& R. Stolze (Eds.), Doubts and directions in Translation Studies: Selected
contributions from the EST Congress, Lisbon 2004 (pp. 183–197). Amsterdam:
John Benjamins.
Dam, H. V., & Engberg, J. (2006). Assessing accuracy in consecutive interpreting: a
comparison of semantic network analyses and intuitive assessments. In C. Heine,
K. Schubert, & H. Gerzymisch-Arbogast (Eds.), Text and translation: Theory
and methodology of translation (pp. 215–234). Tübingen: Narr Francke
Attempto Verlag.
Dam, H. V., Engberg, J., & Schjoldager, A. (2005). Modelling semantic networks on
source and target texts in consecutive interpreting: A contribution to the study of

125
 interpreters’ notes. In H. V. Dam, J. Engberg, & H. Gerzymisch-Arbogast (Eds.),
Knowledge systems and translation (Vol. 7, pp. 227–254). Berlin: Mouton de
Gruyter.
Ehrensberger-Dow, M., Dimitrova, B. E., Hubscher-Davidson, S., & Norberg, U. (Eds.).
(2015). Describing cognitive processes in translation: Acts and events.
Amsterdam: John Benjamins.
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. 53–66). London and New
York: Routledge.
Gile, D. (1991). Prise de notes et attention en début d’apprentissage de l’interprétation
consécutive–une expérience– démonstration de sensibilisation [Note-taking and
attention at the beginning of consecutive interpreting learning –an experience–
demonstration of awareness]. Meta, 36(2–3), 431–439.
Gile, D. (1999). Variability in the perception of fidelity in simultaneous interpretation.
Hermes, 22, 51–79. doi:10.7146/hjlcb.v12i22.25493
Gile, D. (2009). Basic concepts and models for interpreter and translator training
(revised edition). Amsterdam: John Benjamins.
Gillies, A. (2005). Note-taking for consecutive interpreting: A short course. Manchester
and Kinderhook: St. Jerome.
Goldman-Eisler, F. (1972). Segmentation of input in simultaneous translation. Journal
of Psycholinguistic Research, 1(2), 127–140.
Her, H. (2001). Notetaking in basic interpretation class: An initial investigation. Studies
of Translation and Interpretation, 6, 53-77.
Ilg, G. (1988). La prise de notes en interprétation consécutive. Une orientation générale
[Note-taking in consecutive interpretation: A general approach]. Parallèles, 9,
9–13.
Jones, R. (1998). Conference interpreting explained. Manchester: St. Jerome.
Kirchhoff, H. (1979). Die notationssprache als hilfsmittel des konferenzdolmetschers im
konsekutivvorgang [The language of note-taking as a tool for the conference
interpreter in consecutive interpreting]. In W. Mair & E. Sallager (Eds.),
Sprachtheorie und sprachpraxis [Language theory and language practice] (pp.
121-133). Tübingen: Gunter Narr.

126
Lee, T.-H. (2002). Ear voice span in English into Korean simultaneous interpretation.
Meta, 47(4), 596-606. doi:10.7202/008039ar
Li, R. (2010). A contrastive study of the semantic functions and pragmatic uses of
modal verbs in English and Chinese. In D. Shu & K. Turner (Eds.), Contrasting
meaning in languages of the East and West (pp. 335-360). Bern: Peter Lang.
Liu, M. (2008). 逐步口譯與筆記 [Consecutive interpretation and note-taking]. Taipei:
Bookman.
Liu, M., & Chiu, Y.-H. (2009). Assessing source material difficulty for consecutive
interpreting: Quantifiable measures and holistic judgment. Interpreting, 11(2),
244–266. doi:10.1075/intp.11.2.07liu
Lung, R. (1999). Note-taking skills and comprehension in consecutive interpretation.
Babel, 45(4), 311–317. doi:10.1075/babel.45.4.04lun
Lung, R. (2003). Taking "notes" seriously in the interpretation classroom. In Á.
Collados Aís, M. M. Fernández Sánchez, & D. Gile (Eds.), La evaluación de la
calidad en interpretación: investigación [The evaluation of quality in
interpretation: Research] (pp. 199–205). Granada: Comares.
Martín, R. M. (Ed.) (2016). Reembedding translation process research. Amsterdam:
John Benjamins.
Matyssek, H. (1989). Handbuch der notizentechnik für dolmetscher [Handbook of note-
taking for interpreters]. Heidelberg: Julius Groos.
Oléron, P., & Nanpon, H. (1965/2002). Research into simultaneous translation. In F.
Pöchhacker & M. Shlesinger (Eds.), The interpreting studies reader (pp. 43–50).
London and New York: Routledge.
Orlando, M. (2010). Digital pen technology and consecutive interpreting: Another
dimension in note-taking training and assessment. The Interpreters’ Newsletter,
15, 71–86.
Pöchhacker, F. (2002). Researching interpreting quality: Models and methods. In G.
Garzone & M. Viezzi (Eds.), Interpreting in the 21st century: Challenges and
opportunities (pp. 95–106). Amsterdam: John Benjamins.
Rozan, J.-F. o. (2002). Note-taking in consecutive interpreting (A. Gillies, Trans.).
Cracow: Tertium Society for the Promotion of Language Studies.
Santorini, B. (1990). Part-of-speech tagging guidelines for the Penn Treebank Project
(3rd revision). Retrieved from University of Pennsylvania:
http://repository.upenn.edu/cis_reports/570
127
Schweda-Nicholson, N. (1993). An introduction to basic note-taking skills for
consecutive interpretation. In E. Losa (Ed.), Keystones of communication:
Proceedings of the 34th Annual Conference of the American Translators
Association (pp. 197–204). Medford: Learned Information.
Seleskovitch, D. (1975). Langage, langues et mémoire: étude de la prise de notes en
interprétation consécutive [Speech, language, and memory: A study of note-
taking in consecutive interpreting]. Paris: Minard Lettres Modernes.
Seleskovitch, D., & Lederer, M. (1989/1995). A systematic approach to teaching
interpretation (J. Harmer, Trans.). Silver Spring, MD: Registry of Interpreters
for the Deaf.
Shreve, G. M., & Angelone, E. (Eds.). (2010). Translation and cognition. Amsterdam
and Philadelphia: John Benjamins.
Szabó, C. (2006). Language choice in note-taking for consecutive interpreting.
Interpreting, 8(2), 129–147. doi:10.1075/intp.8.2.02sza
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 and Philadelphia: John Benjamins.
Tirkkonen-Condit, S., & Jääskeläinen, R. (Eds.). (2000). Tapping and mapping the
processes of translation and interpreting: Outlooks on empirical research.
Amsterdam and Philadelphia: 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
Xia, F. (2000). The part-of-speech tagging guidelines for the Penn Chinese Treebank
(3.0) (38). Retrieved from University of Pennsylvania:
http://repository.upenn.edu/ircs_reports/38

128
An introductory note to Chapter 6

Following the analysis of pen and voice data in Chapter 5, Chapter 6 reports the eye
tracking data collected in the main study of the PhD project. The focus is on Phase II of
CI in which interpreters read back their notes and produce a target speech. The chapter
contributes a discussion on the eye movement measures used in the study, accompanied
by a definition to each measure and examples of application. It demonstrates how eye
tracking can be applied in interpreting research, especially research involving note-
taking and CI.
Through a combined analysis of the qualitative and quantitative data, Chapter 6
presents how visual processing is carried out during note-reading. It investigates the
cognitive load in the interpreting process using a series of eye movement measures. In
addition, Chapter 6 puts two pieces of the puzzle together by synthesising the findings
on both Phase I and Phase II of CI. It reveals that there is a trade-off between the
cognitive costs of different note-taking choices in the two phases, and further discloses
the interplay between note-taking, cognitive load and interpreting performance.

129
Chapter 6 An eye-tracking approach to note-reading in consecutive
interpreting: Reading patterns and cognitive load1

Abstract ： This study reports the eye tracking data collected from professional
interpreters while they perform consecutive interpreting (CI) with notes. It is among the
first to visualise the way in which note-reading proceeds. Data collected in this study
provide important indicators of cognitive processing and cognitive load in the
interpreting process. It then transpires that the note-taking choices made during Phase I
of CI (in which interpreters listen to the source speech and write notes) affect the level
of cognitive load in Phase II (in which interpreters read back their notes and produce a
target speech). The results indicate that there is a trade-off between the cognitive costs
in Phase I and Phase II, with interpreters strategically lowering the load in Phase I
(during which interpreters are dealing with the input information for the first time and
are paced by the speaker) and leaving the burden of increased cognitive load to Phase II
(during which interpreters are more familiar with the information content and set their
own pace).
Keywords: eye tracking; note-reading; note-taking; consecutive interpreting; cognitive
load

6.1 Introduction

Translation and interpreting process research has made important progress in the past
two decades. However, in terms of interpreting, simultaneous interpreting (SI) has
received the bulk of the attention, with little attention devoted to studying the process of
consecutive interpreting (CI). It may be the case that CI has been largely replaced by SI
in some markets (such as western Europe) and contexts (such as multilateral and
multilingual conferences), but it remains an important mode of interpreting in other
markets (such as Asia) and the preferred mode in various other contexts such as
diplomatic and business negotiations (Andres, 2015; Dam, 2010). CI is also an
important part in the curriculum of interpreter training (Setton & Dawrant, 2016, pp. 82-
83). Therefore, it is somewhat surprising that CI has been left behind in the vigorous
development of process-oriented research on translation and interpreting.

1
This chapter has been prepared for publication as: Chen, S. (manuscript prepared for publication). An
eye-tracking approach to note-reading in consecutive interpreting: Reading patterns and cognitive load.

130
 CI is an interesting activity from both a cognitive and a linguistic point of view.
Similar to SI, it requires a high level of bilingual language processing and challenges
the interpreter’s cognitive system by requiring multi-tasking under strict time
constraints. But CI 2 also introduces a new challenge: note-taking. In addition to
listening to the source speech and producing a target speech, CI requires the interpreter
to perform the tasks of note-writing and note-reading and to deliver the output relying
on memory with the assistance of notes. In Phase I of CI, interpreters listen to and
analyse the source speech, keep parts of the speech in their working memory, and write
down notes. In Phase II, interpreters read back their notes, retrieve information from
their working memory, and produce a target speech.
Phase I of CI has received some attention, with most of the efforts devoted to note-
taking and in particular its product, the notes (see Chen (2016) for a review). Phase II of
CI, however, has been under-researched relative to its importance to professional
practice and its potential for providing insights into cognitive processes. It is evident
that researchers either stop at observing the products of note-writing or shift their
attention from written notes directly to interpreting performance, ignoring the actual
processes involved in the note-reading and production phase. This gap in the literature
may very well be due to the fact that process-oriented research on CI has been impeded
by inadequate research methods.
Digital pen recording is a valuable research method for Phase I, but it has not been
applied in interpreting research until quite recently (Chen, 2017b; Orlando, 2010, 2014).
Before that, process research on CI relied on less accurate and more cumbersome
methods such as video recording (which involved manually checking the timing of
note-writing) and even then, very little can be found in the literature (see however
Andres (2002)). Another research method with substantial potential for investigating the
cognitive processes in Phase II of CI is eye tracking (Chen, 2017a), a method that has
long been applied in fields such as reading, scene perception, and cognitive sciences,
but only became a popular research method in translation studies during roughly the
past decade (Hvelplund, 2017).
This study uses eye tracking to study the process of CI. It aims to make a
methodological contribution to interpreting research, especially process-oriented
research. The focus of the study is Phase II of CI, in which interpreters read back their

2
In this article, CI refers to long consecutive in which systematic note-taking is used.

131
notes and produce a translated speech. Through an analysis of the eye movement data,
the study is among the first to visualise the way note-reading proceeds in CI and
provides us some insight into interpreters’ “black box” during this complex cognitive
and linguistic process.

6.2 Note-taking and note-reading in CI

Note-taking in CI is a particularly interesting topic in interpreting research. Both

prescriptive and descriptive accounts of note-taking have been generated over the years.
On the prescriptive side, various note-taking systems have been established in a
substantial volume of literature, including those that address note-taking as a subject in
its own right (e.g., Gillies, 2005; Kirchhoff, 1979; Matyssek, 1989; Rozan, 1956) and
those that include note-taking as part of their discussions on interpreter training and
education (e.g., Bowen & Bowen, 1980; Ilg & Lambert, 1996; Jones, 1998;
Seleskovitch & Lederer, 1995). A series of skill-oriented issues have been discussed,
such as how to arrange the layout of the notes, how to use symbols and abbreviations,
and which language to use (source or target language) for taking notes.
On the descriptive side, note-taking in CI has been the focus of a number of
empirical studies. Important variables investigated include the choice of form (e.g.,
Dam, 2004a), the choice of language (e.g., Abuín González, 2012; Dam, 2004b; Szabó,
2006), and the relationship between note-taking and interpreting performance (e.g.,
Cardoen, 2013; Dai & Xu, 2007; Dam, 2007). In terms of the choice of form,
interpreters can take notes either in language or in symbols; if they take notes in the
form of language, they could use either abbreviations or full words. In terms of the
choice of language, interpreters can choose between the source and the target language,
or between the native and non-native language.
Findings in those empirical studies have pointed to some general trends.
Interpreters take much more notes in language than in symbol, but no consistent
preference for abbreviation or full word has been found. Their choice of language in
note-taking is dependent on a series of factors, such as the nature of the languages
themselves (e.g., morphological complexity and economy of writing), the native/non-
native language status, the source/target language status, task characteristics (e.g.,
texture or genre of the source speech) and interpreter characteristics (e.g., working
experience and language competence).

132
 Compared to the amount of research efforts devoted to note-taking in Phase I of CI,
the attention granted to note-reading in Phase II has been very limited. Note-reading is
in fact a very important process in CI. After interpreters have listened to the source
speech and taken down notes, they read back their notes and produce a translated
speech. The reading process takes place simultaneously with the speaking process and
provides information for the interpretation to cue and supplement the working memory.
Poor quality notes have a detrimental effect on production because too much effort will
be allocated to note-reading. Good quality notes, on the other hand, can facilitate target
speech production, “telling the interpreter when to pause, when to add emphasis and
when not to” (Gillies, 2005, p. 7).
Moreover, the note-taking choices made by interpreters during Phase I can have a
direct bearing on Phase II of CI. A recent study (Chen, 2017c) has shown that the varied
note-taking choices made in Phase I are associated with different levels of cognitive
load, and that interpreters’ note-taking choices are mainly determined by cognitive
rather than physical and temporal demands (Table 6.1). However, very little is known
about how these varied note-taking choices made in Phase I affect the cognitive
processing and cognitive load in Phase II. This is one of the questions that the current
article will attempt to answer.

Table 6.1 Note-taking choices and cognitive load in Phase I of CI

Choice of form Choice of language

Language vs. symbol Abbreviation vs. full word Chinese vs. English
Physical and Abbreviation <
Language > Symbol Chinese ≈ English
temporal demands Full word
Cognitive load in Language < Abbreviation ≈ Chinese > English
Phase I Symbol Full word (C-E only)
Preferred note-
Language No preference English
taking choice
Adapted from (Chen, 2017c)

What makes note-reading even more intriguing is that it is a very unique reading
process. Interpreters do not “read” their notes in the usual sense of the word, so note-
reading both resembles and differs from reading in general. Similar to general reading,
interpreters need to read the notes to grab information. Therefore, in terms of
information intake, note-reading should not deviate too much from normal reading. But
what sets note-reading apart from general reading is that interpreters are simultaneously

133
performing the cognitive process of producing a translation of the source speech from
memory supported by what they are reading. The prescriptive literature suggests that
interpreters should be reading ahead and preparing the next section of the target speech
while producing a translated speech for a former section (Gillies, 2005, p. 73). But so
far there is little empirical evidence of this particular reading behaviour.
Eye tracking is a particularly useful tool for studying note-reading in CI. By
recording the eye movements of interpreters while they are reading their notes and
producing a target speech, Phase II of CI can be approached in unprecedented ways. In
addition to providing information on attention allocation, eye tracking data contribute
important indicators of cognitive processing, allowing us to see how the note-taking
choices made by interpreters during Phase I affect the level of cognitive load in Phase II.
Eye tracking also allows us to visualise the way in which note-reading progresses and
find traces of how note-reading interacts with memory and target speech production.

6.3 Eye tracking and cognitive processing in language tasks

In the late 19th century, Javal first observed that eye movements in reading consist of a
series of rapid movements (saccades) and intermittent short stops (fixations) (Huey,
1908). Since those very early days, many findings about eye movements in language
processing have been reported (see Liversedge, Gilchrist, and Everling (2011) for a
review). One important discovery was that information intake could only occur during
fixations but not during saccades. Just and Carpenter (1980) thus formulated the
influential eye–mind hypothesis. It states that there should be no appreciable lag
between what is fixated and what is being processed. When people look at a word or an
object, they also simultaneously process it for the duration of the fixation. In this way,
eye movements open a window of opportunity for studying the cognitive processes of
perception and comprehension.
Eye tracking is a technique that captures where, how and when people look using
devices called eye trackers. Most of the eye trackers in use today are reliant on video-
based pupil and corneal reflection (Duchowski, 2007, p. 51). The eye trackers
photograph the pupil and the reflection of an infrared light source from the fovea and
combine the two to determine the point of gaze (i.e., where people look). The eye image
(with the pupil and corneal reflection) forms one coordinate system; the stimulus (such
as a screen) forms the other coordinate system. A calibration procedure is carried out to

134
relate one system to the other by asking the participant to fixate on specific points on
the stimulus and then mapping the gaze point to that.
Once the raw eye tracking data (time-stamped points of gaze) are obtained, they can
be processed to produce various events (such as fixations and saccades). There are two
main types of algorithms used for event detection. Dispersion-based algorithms identify
fixations by finding data samples that are close enough to one another for a
predetermined minimal duration of time. Velocity-based algorithms find saccades as
periods longer than a minimal period of time during which the eye velocity is below a
predefined threshold. Most commercial eye trackers come with built-in software for
extracting events.
After processing raw data samples into events such as fixations and saccades, a
further step can be taken to define areas of interest (AOIs). AOIs are certain regions in
the stimulus within which the eye movement data are of central concern to a study.
Usually, research questions and hypotheses should decide what AOIs to create for the
stimulus. For example, in reading research, an AOI can be a single word or a string of
sentences; in scene perception research, an AOI can be a human face or a moving object.
When using eye tracking to study note-reading, the unit of analysis is a note unit (or a
group of notes that share one common feature). In this case, an AOI needs to be drawn
for every note unit.
When the unit of analysis is as small as a word or a note unit, the appropriate
measure to use is controversial (Inhoff & Radach, 1998). In this context, any single eye
movement measure is a pale reflection of the reality of cognitive processing. A
recommended remedy is to examine multiple measures so that inferences drawn about
the cognitive processing can be as valid as possible (Rayner, 1998; Rayner, Sereno,
Morris, Schmauder, & Clifton, 1989; Schmauder, 1992).
In this study, a series of eye movement measures is investigated for the purpose of
capturing and approximating the cognitive reality of note-reading in CI (also see
Section 4.5.2). Each of these measures is defined below and a brief review is given of
how they have been applied in relevant research fields.
Dwell refers to a visit in an AOI from entry to exit. A dwell can be measured by the
dwell time, which can be based on either the raw data (the sum of all data points
including fixations and saccades falling within an AOI) or fixations (the sum of all
fixations in an AOI). When the unit of analysis is a word (similar to the note unit in this
study), the effect of adding saccade duration is minimal because intra-word saccade

135
durations are quite brief (Rayner, 1998). Therefore, the dwell time in this study is
defined as the sum of fixations only, excluding all other events. Dwell time is a
commonly applied eye movement measure in various fields, but it has been suggested
by a number of researchers that the measure needs to be further refined in order to draw
reasonable inferences about cognitive processing (e.g., Holmqvist et al., 2011; Inhoff &
Radach, 1998). To provide a fine-tuned account of the dwell time, this study
distinguishes first-pass dwell time, second-pass dwell time and total dwell time.
First-pass dwell time (usually termed gaze duration in reading research) refers only
to the first dwell in an AOI from entry to exit. Rayner (1998) reviewed reading research
and concluded that first-pass dwell time (gaze duration) was a good index both of word
frequency and of comprehension processes integrating several words. The measure is
proposed as a candidate for measuring early processing and object recognition
(Liversedge, Paterson, & Pickering, 1998).
A measure which can be confused with the first-pass dwell time is the first fixation
duration. This is the duration of the first fixation in an AOI. It is indicative of fast
processes such as recognition and identification (Holmqvist et al., 2011, p. 385) and
sensitive to cognitive difficulty experienced immediately on processing an AOI
(Liversedge et al., 1998).
Second-pass dwell time includes all subsequent dwells (excluding the first dwell)
on the same AOI. This measure is a useful indicator for global text processing (Hyönä,
Lorch, & Rinck, 2003) and reflects more delayed effects in sentence processing
(Murray, 2000).
Total dwell time subsumes the fixation time in an AOI during an entire trial. It is
sensitive to slow and long-term cognitive processes (Holmqvist et al., 2011). In reading
research, total dwell time was found to be sensitive to linguistic processes that operated
after a word had been identified (Daneman, Reingold, & Davidson, 1995). In translation
studies, total dwell time is often used as an indicator of cognitive load. Many studies
found that the total dwell time on the target text was longer than that on the source text,
indicating more processing efforts spent on target text production and monitoring (Carl
& Kay, 2011; Dragsted, 2010; Jensen, Sjørup, & Balling, 2009). Studies also found that
student translators behaved differently and spent more time looking at the source text,
indicating a more effortful source text analysis (Jakobsen & Jensen, 2008; Sharmin,
Špakov, Räihä, & Jakobsen, 2008). Sjørup (2008) reported that the total dwell time was
longer when translators came across metaphors, suggesting that dealing with metaphors

136
required increased cognitive processing. In audio-visual translation research, it was
reported that the total dwell time on subtitles was longer for deaf and hard-of-hearing
viewers compared to hearing viewers, indicating a larger effort needed to process
subtitled content and more difficulty in extracting information (Szarkowska, Krejtz,
Pilipczuk, Dutka, & Kruger, 2016).
Another widely used eye movement measure is the average fixation duration. This
is the sum of the duration of all fixations divided by the number of fixations in an AOI
during an entire trial. As has been demonstrated in various fields (e.g., reading, scene
perception and usability research), a longer average fixation duration is usually
associated with a deeper and more effortful cognitive processing (Holmqvist et al.,
2011, p. 381). This measure typically varies across different tasks and stimuli. The
average fixation duration is about 225 milliseconds in silent reading (for
comprehension) and 275 milliseconds in oral reading (Rayner, 1998); it ranges from
205 milliseconds (Jakobsen & Jensen, 2008) to 245 milliseconds (Dragsted, 2010) in
reading in preparation for translation, and from 235 milliseconds (Jakobsen & Jensen,
2008) to 252 milliseconds in sight translation/interpretation (Dragsted & Hansen, 2009).
Apart from the time-based measures, a series of count-based eye movement
measures are also related to cognitive processing in language tasks. The number of
fixations refers to the fixation count inside an AOI during a trial. In translation research,
the number of fixations has been found to index cognitive load. Doherty, O'Brien, and
Carl (2010) provided participants with machine translated sentences with good and bad
acceptability. They found that there were more fixations and longer dwell times on
“bad” sentences than on “good” sentences, indicating that the former was more difficult
to process. Dragsted (2012) provided words with high versus low target text variability
(number of alternative renditions in the target text) to participants and found that the
number of fixations on high variation words was higher than that on low variation
words.
Revisit is defined as a transition to an AOI already visited. In picture viewing, the
number of revisits is an indicator of a semantically informative AOI (Loftus &
Mackworth, 1978) or a need to confirm (Mello-Thoms et al., 2005). In reading research,
the number of revisits is related to incomplete lexical processes (Pollatsek & Rayner,
1990; Pynte, 1996).
Regressions are right-to-left movements along the line of reading or movements
back to previously read lines (Rayner, 1998). It has been found that people make more

137
regressions when the reading task was cognitively demanding (Frazier & Rayner, 1982).
The regression rate reported in reading research (reading in native language) is about
10–15% (Rayner, 1998); in translation studies, the regression rate is 20–25% in reading
a foreign language and 30–35% in sight translation/interpretation (Shreve, Lacruz, &
Angelone, 2010).
Skip rate in this study is defined as the percentage of AOIs that are not fixated. In
reading research, a word is skipped when it does not receive a direct fixation during first
pass (Rayner, Slattery, & Drieghe, 2011). In translation studies, skip rate concerns
whether a word has been fixated at all (Schaeffer, Dragsted, Hvelplund, Balling, & Carl,
2016). In this study, skip rate is defined as the percentage of AOIs that do not receive
any fixation during the entire trial, because note-reading is a non-linear type of reading
(see Section 6.1) and AOIs that are initially skipped have a high probability of being
fixated later. In reading research, the skip rate of content words is about 15%, whereas
the skip rate of function words is about 65% (Rayner, 1998), leading to an average
skipping rate of about one third (Rayner et al., 2011).
Eye movement measures are usually used in combination to gauge cognitive efforts
in language processing. For example, if an effect is found on first fixation duration
and/or first-pass dwell time, the difficulty was usually experienced immediately on
processing that AOI; if an effect is only observed for total and/or second-pass dwell
time, this could be indicating a relatively late effect on processing (Liversedge et al.,
1998). Examining a series of eye movement measures could provide researchers with
abundant information in explaining the cognitive processes during language tasks.

6.4 Purpose of the study and research questions

This study attempts to contribute further empirical data to the process research on CI
and note-taking using eye tracking. In particular, it reports new data on Phase II where
interpreters read their notes and produce a target speech, and seeks answers to the
following research questions.
First, how do interpreters read back their notes? This study will be among the first
to visualise note-reading in CI. Moreover, when people read, their eye movements differ
as a function of the task, such as reading for comprehension, reading for translation and
note-reading in CI. This study tries to find out how note-reading resembles and deviates

138
from other forms of reading. To address this research question, the eye tracking
measures of average fixation duration, regression rate, and skip rate are consulted.
Second, what is the relationship between the note-taking choices and cognitive load
in Phase II of CI? Interpreters make choices on the form and language to be used for
note-taking in Phase I of CI, and these choices are likely to impact the cognitive load in
Phase II. In Phase I, differences in cognitive load have been observed to be associated
with different note-taking choices (Chen, 2017b, 2017c). This study will reveal whether
the same patterns of differences will be found in Phase II. A series of eye tracking
measures are used to answer this research question, including the first fixation duration,
first-pass dwell time, second-pass dwell time, total dwell time, average fixation
duration, number of fixations, number of revisits, and skip rate.

6.5 Method

6.5.1 Participants

An advertisement of the study was sent out via the researchers’ professional network to
recruit qualified participants (in terms of accreditation, working language combination
and experience). A total of 26 professional interpreters participated in the study and
they were paid for their participation. They all had the Professional Interpreter
accreditation from Australia’s National Accreditation Authority for Translators and
Interpreters (NAATI), with a working language combination of Mandarin Chinese
(native language) and English (non-native language). Most of them had a postgraduate
degree in interpreting (65%); some had an interpreting diploma (15%); some attended
an intensive interpreting training course (15%); and one was self-trained (4%)3. With an
average age of 36.4, the participants had worked as full-time or part-time interpreters
for an average of 7.4 years. The country they most frequently worked in was Australia
(with only two exceptions who worked more frequently in China). For those who were
working as part-time interpreters, their other job(s) had a bilingual feature (e.g.,
interpreter trainer).
The participants used a digital pen and a tablet for note-taking 4 and their eye
movements were recorded using a head-mounted eye tracker (there was a practice
session to familiarise them with the equipment, see Section 5.4). A post-experiment

3
Percentages do not add up to 100% due to rounding.
4
The pen recording data has been reported in another article currently under review.

139
questionnaire was used to ask whether the interpreters felt comfortable with the digital
pen and tablet. If the participants indicated a high level of discomfort with the
equipment (measured as a rating below 50% on the comfort scale), the participant was
excluded from the final data analysis due to the high likelihood that the use of the
equipment would be too much of a barrier to that participant to allow for ecologically
valid data. Eye tracking also led to some further data loss (e.g., the eye tracker did not
work well with bi-focal glasses), which was not uncommon in eye tracking studies. In
the end, 18 participants had both their pen and eye data successfully collected, meaning
that roughly 31% of the participants were excluded. The data loss in eye tracking
research varies significantly, ranging from 20–60% of participants/trials (Holmqvist et
al., 2011). Considering that the data loss in this study is a result of the combination of
two methods (pen recording and eye tracking), it is within an acceptable range.

6.5.2 Apparatus

The eye tracker used in the study was the SMI ETG 2W. It is a lightweight (47 g), head-
mounted eye tracker that uses dark pupil tracking. It has a tracking accuracy of .5°over
all distances and a sampling rate of 60 Hz. The eye tracker has a built-in high-definition
camera for scene recording. Participants sat in front of a 13-inch LCD tablet with a
resolution of 1366768 pixels. The tablet was the Wacom Cintiq 13HD, equipped with a
Wacom Pro Pen, which the participants used for note-taking. The Eye and Pen
software5 was used for collecting and analysing the pen data. The experiment took place
in a sound-proof studio with constant artificial illumination. The SMI software iView
ETG and BeGaze were used with default settings for eye data recording and analysis
respectively.

6.5.3 Tasks

There were two CI tasks, one from English to Chinese (E-C), and one from Chinese to
English (C-E). A series of steps were taken to ensure that the two tasks were as
comparable as possible.

5
More introduction to the software can be found on http://www.eyeandpen.net/?lng=en. The software can
be used for the synchronous recording of eye and pen data, but unfortunately it does not support the type
of eye tracker used in this study. So all the eye data was synchronised with the pen data post-hoc. This
added more work load to the researcher, but did not affect data quality.

140
 Step one, source text selection and editing. Two English video clips on similar
topics (see Table 6.2) were selected from the Internet and transcribed by the author.
This created two English scripts (1 and 2) which were later edited to make them
comparable by an Australian university lecturer. The focus was on the length,
complexity and style of language. The edited scripts were put into CPIDR 5 6 for
analysis and results showed that they were comparable in word count, proposition count
and idea density (Table 6.2).
Step two, translation and editing. Script 2 was translated into Chinese and refined
to make it suitable to be read aloud and recorded as a speech. This was done by two
Chinese editors working at a local Chinese radio station.
Step three, audio recording. A native English speaker in Australia 7 (the same
person who edited the original English scripts) read aloud the English speech and the
audio was recorded. A native Chinese (Mandarin) speaker (a radio personality from the
radio station mentioned above) read aloud the Chinese speech and the audio was
recorded. Both recordings were made in professional sound-proof studios.
Step four, audio editing. The recordings were edited using Audacity. Each speech
was divided into three segments and controlled for variables such as pauses, duration
and speed (Table 6.2).

Table 6.2 Task specifications

Word/character Proposition Idea Segment duration

Task Topic Duration
count count density 1 2 3
How to register
E-C a business in 631 321 0.509 4'59" 1'18" 2'02" 1'39"
Australia
How to purchase 630 (English) *
C-E a property in 324* 0.514* 4'47" 1'10" 2'07" 1'30"
Australia 944 (Chinese)

* The calculations were based on the original English script to enable comparison between the two tasks.

6.5.4 Experimental procedure

First, participants were allowed sufficient time to practice using a short task. During this
practice session they became familiarised with the digital pen and tablet, the eye tracker,
and the experiment procedures. Second, the participants performed the two interpreting
6
More details about the programme and the analysis can be found at http://ai1.ai.uga.edu/caspr/ and
Brown, Snodgrass, Kemper, Herman, and Covington (2008).
7
The person speaks standard English with a neutral accent.

141
tasks with randomised order. Third, after both tasks were completed, the participants
were provided with their written notes for cued retrospection. Fourth, the participants
completed a post-experiment questionnaire. The purpose of the questionnaire was to
collect information such as the participants’ familiarity with the task topics, how they
felt about using the digital pen and the eye tracker, and other feedback about the
experiment. A calibration 8 procedure was carried out for each participant at the
beginning of each trial. It was followed by a validation procedure and recalibration was
performed when needed. The calibration was further checked at the end of each trial.
The experimental procedure is outlined in Table 6.3.

Table 6.3 Experimental procedure

Session Steps
Eye tracker set up, calibration, and start a trial;
Start the Eye and Pen software;
I Practice Let the participant practice note-taking with the digital pen and tablet
using the practice material, until they feel comfortable with the
equipment and familiar with the procedures.
Calibration, validation (recalibration if needed), and start a trial;
Start the Eye and Pen software;
Task 1 Perform Task 1 (could be either E-C or C-E, randomised);
II
Recheck the calibration;
Rest.
Task 2 Same procedures as Task 1.

III Retrospection Retrospection for Task 1 and Task 2.

IV Questionnaire Complete a post-experiment questionnaire.

6.5.5 Data and analysis

6.5.5.1 Semantic Gaze Mapping and AOI drawing

Semantic Gaze Mapping is a function provided by SMI’s analysis software BeGaze.

This function is used to map the gaze data points from scene videos to corresponding
reference images. There are two ways to perform the mapping: an event-based mapping
where there is one gaze mapping for each event at a certain frame and the other frame
mappings for that event are generated automatically; a frame-based mapping where

8
The SMI ETG offers two types of calibration: a one-point and a three-point calibration. The one-point
calibration was used in this study because it generated good quality data in test runs and it reduced the
time and money cost of the experiment, during which professional interpreters were paid for their
participation.

142
there is one gaze mapping for each frame in the scene video. The latter is very labour-
intensive and is usually used when multiple analysers are available to do the mapping.
In this study, there was only one researcher to do the mapping so an event-based
mapping method was selected.
Each page of notes taken by each interpreter was saved automatically by the Eye
and Pen software as a picture file. These pictures were imported into BeGaze to be used
as reference images. The gaze point data on the scene video were mapped onto the
reference images. Figure 6.1 shows two screenshots during Semantic Gaze Mapping.
On the right is the scene video of one participant reading one page of notes, showing the
current gaze position. On the left is a corresponding reference image (a picture file of
the page of notes being read). When the analyser holds the mouse button while clicking
the correct gaze position on the reference image, a magnified image of the area under
the mouse cursor is shown for improved positioning.

Figure 6.1 An example of Semantic Gaze Mapping

After all relevant eye tracking data were mapped onto the reference images, AOIs
were drawn on the images instead of the scene video, which increased the accuracy and
efficiency of analysis. An AOI was drawn for each note unit (Figure 6.2).

143
Figure 6.2 An example of drawing the AOIs

This study compares how interpreters read notes in different forms (language vs.
symbol notes; abbreviation vs. full word notes) and in different languages (Chinese vs.
English notes). Therefore, each AOI was labelled post hoc according to its form and
language (Figure 6.3). The labelling of the notes was informed by the cued retrospection,
during which participants were asked to provide as much information as they could
remember about the note-taking process as well as the form and language of each note
unit. This was important because interpreters’ notes were highly individualised and
could be difficult for others to decipher. One note unit could have multiple labels. For
example, an English abbreviation could be labelled simultaneously as “language”,
“abbreviation” and “English”. AOIs with the same label are treated as one single
(distributed) AOI. For example, all notes labelled with “Chinese” belong to an AOI
called Chinese and all notes labelled with “English” belong to an AOI called English.

144
Figure 6.3 Labelling the AOIs

6.5.5.2 Eye tracking measures used in this study

A range of eye tracking measures (Table 6.4) has been used in this study to answer the
two research questions. These measures are summarised with a brief definition
provided.

Table 6.4 Eye tracking measures used in this study

Eye tracking measure Definition Target research question

Number of regressive fixations to previous How do interpreter read
Regression rate
AOIs divided by the total number of fixations. back their notes?
First fixation duration Duration of the first fixation on an AOI.
Sum of the fixation durations during the first
First-pass dwell time
dwell on an AOI from entry to exit.
Second-pass dwell What is the relationship
Total dwell time – first-pass dwell time.
time between the note-taking
Sum of the fixation durations on an AOI in an choices and cognitive load in
Total dwell time Phase II of CI?
entire trial.
Fixation count inside an AOI during an entire
Number of fixations
trial.
Number of revisits Count of transitions to an AOI already visited.
Average fixation
Total dwell time / number of fixations
duration
Both research questions
Number of AOIs skipped divided by the total
Skip rate
number of AOIs.

6.5.5.3 Statistical analysis

The unit of analysis in this study was a note unit and all data were standardised in the
same manner. For example, if the number of Chinese notes written by a participant is n,
then the total dwell time (T) of Chinese notes of that participant is calculated as:

145
 𝑛
1
× ∑ 𝑇𝑖
𝑛
𝑖=1

Paired-samples t-tests were conducted to compare the note-taking choices in

different forms (language vs. symbol; abbreviation vs. full word) and languages
(Chinese vs. English). All statistical analyses were performed by running the IBM SPSS
Statistics 22. Two-tailed p values less than 0.05 were considered to be statistically
significant. Cohen's d (the difference between the means divided by the pooled standard
deviation) was used to indicate the effect sizes, which were classified as small (d = 0.2),
medium (d = 0.5), and large (d = 0.8).
Normality was checked for all data using the Shapiro-Wilk (S-W) test. Of the 96
sets of data involved in t-tests (8 eye movement measures × 2 interpreting directions × 6
note-taking choices), only three were not normally distributed, all found in the C-E task:
the total dwell time of Chinese notes (SW = .87, df = 18, p = .015, skewness = 1.09,
kurtosis = 0.24), second-pass dwell time of Chinese notes (SW = .87, df = 18, p = .017,
skewness = 1.41, kurtosis = 1.82), and the first-pass dwell time of full word notes (SW
= .79, df = 18, p = .001, skewness = 2.14, kurtosis = 6.60). For comparisons involving
these three sets of data, additional non-parametric tests (the related-samples Wilcoxon
signed-rank test) were run and the results (non-significant) did not change. Considering
that parametric statistics are robust to violations of the normality assumption
(Tabachnick & Fidell, 2007), and for the sake of consistency in data report and
interpretation, only the t test results were presented in this study.

6.6 Results

6.6.1 How interpreters read back their notes

A visual check of the scanpath video revealed that note-reading progressed in a non-
linear manner. A sample AOI sequence chart is presented in Figure 6.4. This shows how
one participant read one page of notes. The x axis indicates the passing of time and the
y axis shows the AOIs (notes) being fixated (the little bars are fixations). If note-reading
was linear, the participant would read the notes one by one in a sequential manner as
time passed. However, the data showed that notes were read in groups, with each group
of notes being meaning-related chunks. The participant read the first group of notes
(AOIs 17 to 20) between time 0 to 8750 milliseconds, moved to the second group of
notes (AOIs 22 to 26) and read them between time 8750 and 21,250 milliseconds, and

146
then read the third group of notes (AOIs 28 to 33) for the remaining time. The
participant also moved frequently between the notes within each group, indicating that
during speech production around one note unit, the interpreter did read ahead to
examine other related note units. But this read-ahead usually happened within the note
groups (meaning-based chunks) and rarely between the groups.

Figure 6.4 A sample AOI sequence chart

Descriptive statistics of the eye movement measures during note-reading are

reported in Table 6.5. The average fixation duration was around 277 milliseconds, the
regression rate was about 23% and the skip rate was about 12% averaged across the two
tasks. All note units were included in the analysis regardless of their form and language.

Table 6.5 Descriptive statistics of note-reading

E-C C-E Average

Measure M SD M SD M SD
Average fixation duration (millisecond) 271.48 70.99 281.77 76.34 276.63 72.93
Regression rate (percentage) 23.18 2.91 21.86 3.00 22.52 2.73
Skip rate (percentage) 11.81 6.03 11.40 5.52 11.61 5.46

6.6.2 The relationship between the note-taking choices and cognitive load in Phase II
of CI

6.6.2.1 The choice between language and symbol

Significant differences were found between language and symbol notes in all the
measures (Table 6.6). The total, first-pass and second-pass dwell times, the first fixation
duration, and the average fixation duration of language notes were significantly longer
than those of symbol notes. The number of fixations and revisits of language notes were

147
significantly higher than those of the symbol notes, whereas the skip rate of the former
was significantly lower than that of the latter. Further, Cohen’s effect size values
(Cohen’s d, M = 2.53) suggested large magnitude of these differences, much bigger than
the differences found between abbreviation and full word (6.2.2) and between Chinese
and English (6.2.3).

Table 6.6 The choice between language and symbol

Language Symbol
Task Measure M SD M SD t df d
First fixation duration 260.03 61.13 141.44 53.81 13.00*** 17 3.07
First-pass dwell time 369.82 86.68 165.39 71.63 15.80*** 17 3.72
Second-pass dwell time 1261.76 576.24 288.38 163.76 8.67*** 17 2.04
Total dwell time 1631.57 626.20 453.77 212.86 9.83*** 17 2.32
E-C
Avg fixation duration 266.76 73.89 151.12 61.09 9.99*** 17 2.35
Number of fixations 5.98 1.97 1.73 0.74 10.58*** 17 2.49
Number of revisits 2.72 1.06 0.79 0.46 10.89*** 17 2.57
Skip rate 0.05 0.05 0.41 0.13 -12.93*** 17 3.05
First fixation duration 276.65 78.96 153.56 68.08 10.80*** 17 2.55
First-pass dwell time 418.49 132.10 190.46 81.63 8.28*** 17 1.95
Second-pass dwell time 1265.83 530.43 280.11 165.94 8.88*** 17 2.09
Total dwell time 1684.32 591.79 470.57 226.51 9.76*** 17 2.30
C-E
Avg fixation duration 282.51 82.48 158.59 63.27 10.01*** 17 2.36
Number of fixations 6.06 1.75 1.87 0.89 11.73*** 17 2.77
Number of revisits 2.60 0.86 0.87 0.54 12.45*** 17 2.93
Skip rate 0.04 0.03 0.37 0.18 -8.11*** 17 1.91

Notes:
1. *** p < .001;
2. All time measures were calculated in milliseconds;
3. Avg = average.

6.6.2.2 The choice between abbreviation and full word

Data on the choice between abbreviation and full word notes are summarised in Table
6.7. In both directions of interpreting, significant differences were found between
abbreviations and full words in terms of the total and second-pass dwell times, the
number of fixations and the number of revisits (the number of revisits in the C-E task
was only approaching significance, p = .051). Full words had longer dwell times, more
fixations and more revisits than abbreviations. The skip rate of abbreviations was
significantly higher than that of full words in the E-C task but not in the C-E task.
148
Cohen’s d values (M = 0.81) suggested that most of the effect sizes were large. In
neither direction were the first-pass dwell time, first fixation duration and average
fixation duration significantly different.

Table 6.7 The choice between abbreviation and full word

Abbreviation Full word

Task Measure M SD M SD t df d
First fixation duration 264.37 65.26 256.57 63.49 1.04 17
First-pass dwell time 370.70 99.24 378.77 85.07 -0.55 17
Second-pass dwell time 1106.25 536.17 1401.71 632.42 -3.19** 17 0.75
Total dwell time 1476.96 578.05 1780.48 678.39 -3.51** 17 0.83
E-C
Avg fixation duration 268.87 76.30 263.96 70.94 1.17 17
Number of fixations 5.37 1.98 6.62 2.02 -4.54*** 17 1.07
Number of revisits 2.52 1.11 2.94 1.02 -3.55** 17 0.84
Skip rate 0.06 0.06 0.04 0.04 3.66** 17 0.86
First fixation duration 278.46 84.33 271.96 73.56 0.57 17
First-pass dwell time 409.75 120.55 434.10 191.17 -0.62 17
Second-pass dwell time 1149.45 393.43 1374.86 635.71 -2.88** 17 0.68
Total dwell time 1559.19 436.88 1808.96 731.32 -2.76* 17 0.65
C-E
Avg fixation duration 284.40 87.56 277.59 75.26 0.70 17
Number of fixations 5.60 1.38 6.58 2.21 -3.37** 17 0.80
4
Number of revisits 2.48 0.69 2.73 1.07 -2.10 17
Skip rate 0.05 0.04 0.04 0.05 1.26 17
Notes:
1. * p < .05, ** p < .01, *** p < .001;
2. All time measures were calculated in milliseconds.
3. Std = standardised, avg = average.
4. The number of revisits in the C-E task was closely significant (p = .051, d = 0.49).

Considering that there might be a consistent difference between the length of

abbreviation and full word notes, and that this difference might have caused the
differences observed in some of the measures, this study took a further step to control
for length. The notes were grouped according to the number of characters they contain
(e.g., one-character group, two-character group, etc.). Then, notes in each group were
compared. When abbreviations and full words with the same length were compared
with each other, the differences were no longer significant. That is to say, full word
notes were not looked at more than abbreviation notes when the variable of length was
controlled.

149
6.6.2.3 The choice between Chinese and English

Data on the choice between Chinese and English notes are summarised in Table 6.8.
When the direction of interpreting was English to Chinese, the total and second-pass
dwell time of Chinese notes were significantly shorter than those of English notes; the
average fixation duration of Chinese notes were shorter than those of English notes
(approaching significance, p = .058); the number of fixations and revisits on the Chinese
notes were significantly lower than those on the English notes; and the Chinese notes
had a significantly higher skip rate. Notes in the two languages did not differ in the first-
pass dwell time or the first fixation duration. However, when the direction of
interpreting was Chinese to English, no significant difference was found on any of the
measures between the two languages.

Table 6.8 The choice between Chinese and English

Chinese English
Task Measure M SD M SD t df d
First fixation duration 250.96 81.27 261.02 63.69 -0.81 15
First-pass dwell time 354.56 152.25 381.96 90.41 -0.92 15
Second-pass dwell time 983.18 543.39 1284.23 492.54 -7.00*** 15 1.75
Total dwell time 1337.74 631.96 1666.19 563.95 -6.20*** 15 1.55
E-C
Avg fixation duration5 255.61 84.02 271.28 77.08 -2.06 17 0.51
Number of fixations 4.86 2.22 6.10 1.70 -4.70*** 15 1.17
Number of revisits 2.24 0.95 2.70 0.75 -3.54** 15 0.89
Skip rate 0.10 0.13 0.04 0.05 2.33* 15 0.58
First fixation duration 277.38 93.57 271.41 82.71 0.46 17
First-pass dwell time 414.65 151.36 405.75 141.55 0.39 17
Second-pass dwell time 1256.62 588.73 1246.50 519.32 0.20 17
Total dwell time 1671.27 606.34 1652.24 597.69 0.32 17
C-E
Avg fixation duration 294.04 93.70 276.63 87.04 1.79 17
Number of fixations 5.79 1.59 6.07 1.80 -1.11 17
Number of revisits 2.52 0.82 2.55 0.84 -0.23 17
Skip rate 0.05 0.04 0.04 0.04 0.58 17
Notes:
1. * p < .05, ** p < .01, *** p < .001;
2. All time measures were calculated in milliseconds;
3. The degree of freedom (df) in the E-C task was 15 because two interpreters took notes in English
only;
4. Std = standardised, avg = average.
5. The average fixation duration in the E-C task was closely significant (p = .058, d = 0.51).

150
6.7 Discussion

6.7.1 Note-reading, reading for comprehension and reading for translation

Qualitative data such as scanpaths and AOI sequence charts confirm that note-reading
proceeds in a non-linear manner. Instead of reading the notes one by one, interpreters
move from one group of notes to another. Each group of notes are possibly meaning
related chunks, corresponding to one part of the target speech. Reading-ahead happens
frequently within each note group, but interpreters do not usually read ahead to examine
another group of notes until they complete reading the current note group.
Quantitative data shows that note-reading is a unique reading behaviour, differing
from other forms of reading in many ways. The average fixation duration in note-
reading is found to be 277 milliseconds. In reading research, the reported average
fixation duration is 225 milliseconds in silent reading and 275 milliseconds in oral
reading (Rayner, 1998). This seems to indicate that the reading of notes corresponds
with oral reading. In translation research, the reported average fixation durations tend to
vary between studies: from 205 milliseconds (Jakobsen & Jensen, 2008) to 245
milliseconds (Dragsted, 2010) in reading silently (in preparation for translation); from
235 milliseconds (Jakobsen & Jensen, 2008) to 252 milliseconds in sight
translation/interpretation (Dragsted & Hansen, 2009).
The regression rate in note-reading found in this study is 23%, almost twice as
much as the regression rate reported in reading research, which is 10–15% (Rayner,
1998). It is similar to the regression rate found in reading in a foreign language (20–
25%) but lower than that in sight translation/interpretation (30–35%) (Shreve et al.,
2010).
The skip rate in note-reading is found to be 12% across all note forms. Compared to
the average skip rate of about one third in reading (Rayner et al., 2011), the skip rate in
note-reading is much lower. This makes sense due to the high meaning load carried by
note units.
Considering that a longer average fixation duration, higher regression rate, and
lower skip rate usually indicate higher cognitive load (see Section 3), the data seem to
suggest that the level of cognitive load during note-reading is higher than that during
reading silently (both reading for comprehension and reading in preparation for
translation) and more similar to that during reading in sight translation/interpretation.
This is plausible considering that when interpreters read their notes, they are

151
simultaneously involved with the additional tasks of memory retrieval, translation and
speech production.

6.7.2 Note-taking choices and cognitive load during Phase II of CI

Highly significant results (p < .001) with large effect sizes (Cohen’s d, M = 2.53) were
found between language and symbol notes, suggesting that language notes are much
more cognitively demanding to read than symbol notes. Notes in these two forms differ
not only in early and fast cognitive processing such as recognising and identifying the
note unit and its meaning (as indicated by first fixation duration and first-pass dwell
time), but also in slower operations such as linking the note unit to the context and
producing a target speech around it (as indicated by second-pass dwell time and total
dwell time). Furthermore, eye tracking measures such as average fixation duration,
number of revisits, and skip rate seem to suggest that reading language and reading
symbol could almost be treated as different task types. Their differences are even bigger
than those between different types of reading tasks (e.g., reading for comprehension and
reading for translation).
In terms of the choice between abbreviation and full word notes, significant
differences are observed in second-pass and total dwell time, but not in first fixation
duration or first-pass dwell time. The two note forms differ in number of fixations and
revisits, but are similar in average fixation durations. That is to say, full word notes are
not more difficult to recognise or comprehend than abbreviations, but they receive more
fixations and revisits, which lead to longer second-pass and total dwell times. The
observed differences could be a function of length: it is possible that full words are
simply being looked at more than abbreviations because they are longer. A further step
needs to be taken to examine the difference between these two forms while controlling
for length.
The notes were grouped according to the number of characters they contain. When
notes in each group were compared (i.e., abbreviations and full words with the same
length), the differences were no longer significant. That is to say, full words are looked
at more mainly because they are longer than abbreviations. However, considering that
time is a valuable resource in interpreting and that full words do cost more time during
note-reading, abbreviations might be a more favourable choice because they attract less
visual attention and take less time to process.

152
 The observed difference between Chinese and English notes is only significant in
the English to Chinese direction. Notes in the two languages have similar first fixation
durations and first-pass dwell times, indicating that the levels of cognitive load
associated with initial processes such as identifying and recognising notes are similar.
However, Chinese notes have shorter average fixation durations, second-pass and total
dwell times and receive fewer fixations and revisits. This indicates that Chinese notes
are easier to process when it comes to later cognitive processes such as integrating the
note unit with the context and producing a target speech around it.
There are two possible contributing factors. First, Chinese is the native language of
the interpreters, hence reading Chinese notes (native language processing) is likely to be
less demanding than reading English notes (non-native language processing). Second,
note-reading and speech production in the same language (within-language processing)
is likely to be less cognitively demanding than reading one language and speaking in
another language (between-language processing). Therefore, when interpreting from
English to Chinese, note-reading in Chinese is both native-language processing and
within-language processing, leading to a significantly lower level of cognitive load
associated with Chinese notes. However, when the direction of interpreting is reversed,
the effect is diminished by the extra costs brought by between-language processing
(from Chinese notes to English speech).
As has been mentioned in Section 2 (Table 6.1), the varied note-taking choices
made by interpreters are associated with different levels of cognitive load in Phase I of
CI. Now it is time to put the two pieces of the puzzle together and see how note-taking
choices and cognitive load interplay during the entire process of CI (Table 6.9). Taken
together, the evidence suggests that cognitive demand (rather than physical or temporal
demands) is the main determining factor in interpreters’ note-taking choices.
Interpreters would take proactive actions to reduce cognitive load during Phase I of CI,
leaving (either consciously or unconsciously) the extra burden to Phase II. However, the
note-taking choices made by interpreters with the intention to ease the cognitive system
in Phase I are not always optimal when all factors are taken into account. For example,
interpreters would benefit more from using more symbols considering that they can
save effort both in note-writing and note-reading. They could also use more
abbreviations, especially considering that abbreviations save time and that they are not
more difficult to process than full words.

153
 Table 6.9 Physical, temporal and cognitive demands of different note-taking choices in
both Phase I and II of CI

Choice of form Choice of language

Language vs. Abbreviation vs. full
Chinese vs. English
symbol word
Physical and Abbreviation <
Language > Symbol Chinese ≈ English
temporal demands Full word
Cognitive load in Language < Abbreviation ≈ Chinese > English
Phase I Symbol Full word (only significant in the C-E task)
Preferred note-taking
Language No preference English
choice
Cognitive load in Abbreviation < Chinese < English
Language > Symbol
Phase II Full word (only significant in the E-C task)

6.8 Conclusions

This study is among the first to reveal the processes that occur during Phase II of CI. It
contributes both qualitative and quantitative data to help explain the unique task of note-
reading and to present the relationship between note-taking choices and cognitive load
in the entire interpreting process. It reveals that there is a trade-off between the
cognitive costs associated with the different note-taking choices in the two phases of CI.
Generally speaking, note-taking choices that help to reduce cognitive load in Phase I are
more difficult to process during Phase II. This study is also pioneering in the use of eye
tracking to investigate CI. Eye tracking has proven to be a useful and powerful method
for visualising and quantifying the process of note-reading and CI.
This exploratory study is admittedly limited in several ways. First, the only head-
mounted eye tracker available to the study was the SMI ETG, with a relatively low
sampling rate of 60 Hz. If other types of head-mounted eye trackers with higher
sampling rates are available, more high quality data can be collected and more eye
movement measures (such as saccades and velocity-based measures) could be used for
analysis. Second, the inferences in the study are made from data successfully collected
from a small group of 17 professional interpreters. Given the common issue of data loss
in eye tracking research (see Section 5.1), the study could have benefited from a larger
sample size. Third, the comparisons between note-reading and other types of reading
are being made even though the stimulus used are possibly very different. The study has
been careful not to over-interpret the data because the difficulty of the stimulus
materials might contribute to the differences observed in data. Fourth, it is not sure

154
whether the findings of this study can be generalised to other language pairs or to other
types of participants (e.g., student interpreters). Hopefully these limitations can provide
some ideas for future research to be conducted on related topics.
Cognitive and process-oriented investigations on the task of interpreting is an
attractive field as it could reveal much about the way in which the human mind
processes two languages simultaneously. There is much that remains unknown, opening
the door for many future studies. The findings of this study will not only benefit the
field of interpreting itself, but also contribute to our understanding of language
processing and the human mind and articulate further links between disciplines.

6.9 References

Abuín González, M. (2012). The language of consecutive interpreters’ notes:

Differences across levels of expertise. Interpreting, 14(1), 55–72.
doi:10.1075/intp.14.1.03abu
Andres, D. (2002). Konsekutivdolmetschen und notation [Consecutive interpreting and
note-taking]. Frankfurt: Peter Lang.
Andres, D. (2015). Consecutive interpreting. In F. Pöchhacker (Ed.), Routledge
encyclopedia of interpreting studies (pp. 84–87). Abingdon: Routledge.
Bowen, D., & Bowen, M. (1980). Steps to consecutive interpretation. Washington, DC:
Pen and Booth.
Brown, C., Snodgrass, T., Kemper, S. J., Herman, R., & Covington, M. A. (2008).
Automatic measurement of propositional idea density from part-of-speech
tagging. Behavior Research Methods, 40(2), 540–545.
doi:10.3758/BRM.40.2.540
Cardoen, H. (2013). The effect of note-taking on target-text fluency. In G. González
Núñez, Y. Khaled, & T. Voinova (Eds.), Emerging research in translation
studies: Selected papers of the CETRA Research Summer School 2012 (pp. 1–
22). Leuven: CETRA.
Carl, M., & Kay, M. (2011). Gazing and typing activities during translation: A
comparative study of translation units of professional and student translators.
Meta, 56(4), 952–975. doi:10.7202/1011262ar

155
Chen, S. (2016). Note-taking in consecutive interpreting: A review with special focus
on Chinese-English literature. The Journal of Specialised Translation, 26, 151–
171.
Chen, S. (2017a). The construct of cognitive load in interpreting and its measurement.
Perspectives. doi:10.1080/0907676X.2016.1278026
Chen, S. (2017b). Note-taking in consecutive interpreting: New data from pen recording.
Translation and Interpreting, 9(1), 4–23. doi:10.12807/ti.109201.2017.a02
Chen, S. (2017c). The process of note-taking and consecutive interpreting: Evidence
from digital pen recording. Manuscript submitted for publication.
Dai, W., & Xu, H. (2007). 汉英交替传译过程中译员笔记特征实证研究——以职业
受训译员和非职业译员为例 [An empirical study of the features of interpreters'
notes in Chinese-English consecutive interpreting: The examples of
professionally trained and unprofessional interpreters]. Foreign Language
Teaching and Research, 39(2), 136–144.
Dam, H. V. (2004a). Interpreters' notes: On the choice of form and language. In G.
Hansen, K. Malmkjær, & D. Gile (Eds.), Claims, changes and challenges in
Translation Studies (pp. 251–261). Amsterdam: John Benjamins.
Dam, H. V. (2004b). Interpreters' notes: On the choice of language. Interpreting, 6(1),
3–17. doi:10.1075/intp.6.1.03dam
Dam, H. V. (2007). What makes interpreters’ notes efficient?: Features of (non-
)efficiency in interpreter's notes for consecutive. In Y. Gambier, M. Shlesinger,
& R. Stolze (Eds.), Doubts and directions in Translation Studies: Selected
contributions from the EST Congress, Lisbon 2004 (pp. 183–197). Amsterdam:
John Benjamins.
Dam, H. V. (2010). Consecutive interpreting. In Y. Gambier & L. van Doorslaer (Eds.),
Handbook of translation studies: Volume 1 (pp. 75–79). Amsterdam: John
Benjamins.
Daneman, M., Reingold, E. M., & Davidson, M. (1995). Time course of phonological
activation during reading: Evidence from eye fixations. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 21(4), 884–898.
Doherty, S., O'Brien, S., & Carl, M. (2010). Eye tracking as an MT evaluation
technique. Machine translation, 24(1), 1–13.

156
Dragsted, B. (2010). Coordination of reading and writing processes in translation: An
eye on uncharted territory. In G. M. Shreve & E. Angelone (Eds.), Translation
and cognition (pp. 41–62). Amsterdam: Jonh Benjamins.
Dragsted, B. (2012). Indicators of difficulty in translation: Correlating product and
process data. Across Languages and Cultures, 13(1), 81–98.
doi:10.1556/Acr.13.2012.1.5
Dragsted, B., & Hansen, I. G. (2009). Exploring translation and interpreting hybrids.
The case of sight translation. Meta, 54(3), 588–604.
Duchowski, A. (2007). Eye tracking methodology: Theory and practice (2 ed.). London:
Springer-Verlag.
Gillies, A. (2005). Note-taking for consecutive interpreting: A short course. Manchester
and Kinderhook: St. Jerome.
Holmqvist, K., Nyström, M., Andersson, R., Dewhurst, R., Jarodzka, H., & van de
Weijer, J. (2011). Eye tracking: A comprehensive guide to methods and
measures. Oxford: Oxford University Press.
Huey, E. B. (1908). The psychology and pedagogy of reading. New York: Macmillan.
Hvelplund, K. T. (2017). Eye tracking in translation process research. In J. W.
Schwieter & A. Ferreira (Eds.), The handbook of translation and cognition (pp.
248-264). Hoboken, NJ: John Wiley & Sons.
Hyönä, J., Lorch, R. F., Jr, & Rinck, M. (2003). Eye movement measures to study
global text processing. In J. Hyönä, R. Radach, & H. Deubel (Eds.), The mind's
eye: Cognitive and applied aspects of eye movement research (pp. 313-334).
Amsterdam: Elsevier.
Ilg, G., & Lambert, S. (1996). Teaching consecutive interpreting. Interpreting, 1(1), 69-
99.
Inhoff, A. W., & Radach, R. (1998). Definition and computation of oculomotor
measures in the study of cognitive processes. In G. Underwood (Ed.), Eye
guidance in reading and scene perception (pp. 29-54). Oxford: Elsevier.
Jakobsen, A. L., & Jensen, K. T. H. (2008). Eye movement behaviour across four
different types of reading task. In S. Göpferich, A. L. Jakobsen, & I. M. Mees
(Eds.), Looking at eyes: Eye-tracking studies of reading and translation
processing (pp. 103-122). Frederiksberg: Samfundslitteratur.
Jensen, K. T. H., Sjørup, A. C., & Balling, L. W. (2009). Effects of L1 syntax on L2
translation. In I. M. Mees, F. Alves, & S. Göpferich (Eds.), Methodology,

157
 technology and innovation in translation process research: A tribute to Arnt
Lykke Jakobsen (pp. 319-336). Frederiksberg: Samfundslitteratur.
Jones, R. (1998). Conference interpreting explained. Manchester: St. Jerome.
Just, M. A., & Carpenter, P. A. (1980). A theory of reading: From eye fixations to
comprehension. Psychological Review, 87(4), 329-354. doi:10.1037/0033-
295X.87.4.329
Kirchhoff, H. (1979). Die notationssprache als hilfsmittel des konferenzdolmetschers im
konsekutivvorgang [The language of note-taking as a tool for the conference
interpreter in consecutive interpreting]. In W. Mair & E. Sallager (Eds.),
Sprachtheorie und sprachpraxis [Language theory and language practice] (pp.
121-133). Tübingen: Gunter Narr.
Liversedge, S. P., Gilchrist, I. D., & Everling, S. (Eds.). (2011). The Oxford handbook
of eye movements. Oxford: Oxford University Press.
Liversedge, S. P., Paterson, K. B., & Pickering, M. J. (1998). Eye movements and
measures of reading time. In G. Underwood (Ed.), Eye guidance in reading and
scene perception (pp. 55–75). Oxford: Elsevier.
Loftus, G. R., & Mackworth, N. H. (1978). Cognitive determinants of fixation location
during picture viewing. Journal of Experimental Psychology, 4(4), 565–572.
doi:10.1037/0096-1523.4.4.565
Matyssek, H. (1989). Handbuch der notizentechnik für dolmetscher [Handbook of note-
taking for interpreters]. Heidelberg: Julius Groos.
Mello-Thoms, C., Hardesty, L., Sumkin, J., Ganott, M., Hakim, C., Britton, C., . . .
Maitz, G. (2005). Effects of lesion conspicuity on visual search in mammogram
reading. Academic Radiology, 12(7), 830–840. doi:10.1016/j.acra.2005.03.068
Murray, W. S. (2000). Sentence processing: Issues and measures. In A. Kennedy, R.
Radach, D. Heller, & J. Pynte (Eds.), Reading as a perceptual process (pp. 649–
664). Oxford: Elsevier.
Orlando, M. (2010). Digital pen technology and consecutive interpreting: Another
dimension in note-taking training and assessment. The Interpreters’ Newsletter,
15, 71–86.
Orlando, M. (2014). A study on the amenability of digital pen technology in a hybrid
mode of interpreting: Consec-simul with notes. Translation and Interpreting,
6(2), 39–54.

158
Pollatsek, A., & Rayner, K. (1990). Eye movements and lexical access in reading. In D.
A. Balota, G. B. F. d'Arcais, & K. Rayner (Eds.), Comprehension processes in
reading (pp. 143–164). Hillsdale, NJ: Erlbaum.
Pynte, J. (1996). Lexical control of within-word eye movements. Journal of
Experimental Psychology: Human Perception and Performance, 22(4), 958–969.
doi:10.1037/0096-1523.22.4.958
Rayner, K. (1998). Eye movements in reading and information processing: 20 years of
research. Psychological Bulletin, 124(3), 372–422.
Rayner, K., Sereno, S. C., Morris, R. K., Schmauder, A. R., & Clifton, C., Jr. (1989).
Eye movements and on-line language comprehension processes. Language and
Cognitive Processes, 4(3-4), SI21–SI49. doi:10.1080/01690968908406362
Rayner, K., Slattery, T. J., & Drieghe, D. (2011). Eye movements and word skipping
during reading: Effects of word length and predictability. Journal of
Experimental Psychology: Human Perception and Performance, 37(2), 514–528.
Rozan, J.-F. o. (1956). La prise de notes en interprétation consécutive [Note-taking in
consecutive interpreting]. Genève: Georg.
Schaeffer, M., Dragsted, B., Hvelplund, K. T., Balling, L. W., & Carl, M. (2016). Word
translation entropy: Evidence of early target language activation during reading
for translation. In M. Carl, S. Bangalore, & M. Schaeffer (Eds.), New directions
in empirical translation process research: Exploring the CRITT TPR-DB (pp.
183–210). Cham: Springer.
Schmauder, A. R. (1992). Eye movements and reading processes. In K. Rayner (Ed.),
Eye movements and visual cognition: Scene perception and reading (pp. 369–
378). New York: Springer-Verlag.
Seleskovitch, D., & Lederer, M. (1995). A systematic approach to teaching
interpretation (J. Harmer, Trans.). Silver Spring, MD: Registry of Interpreters
for the Deaf.
Setton, R., & Dawrant, A. (2016). Conference interpreting: A trainer's guide.
Amsterdam: John Benjamins.
Sharmin, S., Špakov, O., Räihä, K.-J., & Jakobsen, A. L. (2008). Where on the screen
do translation students look while translating, and for how long? In S. Göpferich,
A. L. Jakobsen, & I. M. Mees (Eds.), Looking at eyes: Eye-tracking studies of
reading and translation processing (pp. 31–51). Frederiksberg:
Samfundslitteratur.

159
Shreve, G. M., Lacruz, I., & Angelone, E. (2010). Cognitive effort, syntactic disruption,
and visual interference in a sight translation task. In G. M. Shreve & E.
Angelone (Eds.), Translation and cognition (pp. 63–84). Amsterdam and
Philadelphia: John Benjamins.
Sjørup, A. C. (2008). Metaphor comprehension in translation: Methodological issues in
a pilot study. In S. Göpferich, A. L. Jakobsen, & I. M. Mees (Eds.), Looking at
eyes: Eye-tracking studies of reading and translation processing (pp. 53–77).
Frederiksberg: Samfundslitteratur.
Szabó, C. (2006). Language choice in note-taking for consecutive interpreting.
Interpreting, 8(2), 129–147. doi:10.1075/intp.8.2.02sza
Szarkowska, A., Krejtz, I., Pilipczuk, O., Dutka, Ł., & Kruger, J.-L. (2016). The effects
of text editing and subtitle presentation rate on the comprehension and reading
patterns of interlingual and intralingual subtitles among deaf, hard of hearing
and hearing viewers. Across Languages and Cultures, 17(2), 183–204.
doi:10.1556/084.2016.17.2.3
Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th edition).
Boston: Pearson.

160
Chapter 7 Summary, discussion and conclusion

From a cognitive perspective, CI is a remarkable language processing task because it

involves multiple tasks in both phases: listening, analysing, note-taking and memorising
in Phase I, and remembering, note-reading, and producing a translated speech in Phase
II. Note-taking is a unique activity that channels the two phases of CI, leaving traces of
the cognitive processes involved in the interpreting process. This PhD study has
approached note-taking and CI from a cognitive perspective, with particular focus on
cognitive load in the process. This concluding chapter summarises the main findings of
the PhD research and discusses their implications, pinpoints the strengths and
limitations of the study, and proposes some promising areas for future research.

7.1 Main findings of the research

Three main questions concerning note-taking and CI have been addressed in this study:
(1) What are the preferred note-taking choices of professional interpreters in terms
of form and language?
(2) What is the relationship between note-taking choices and cognitive load in CI?
(3) What is the relationship between note-taking and interpreting performance?
In this section, the main findings of the research are elaborated and organised
around the above questions.

7.1.1 Note-taking preferences

When interpreters take notes, they make choices on the form of notes: whether to use
language or symbol, and if language, whether to use abbreviation or full word.
Literature suggests that there is a consistent preference for language over symbol
established in both professional interpreters (Andres, 2002; Dam, 2004a) and student
interpreters (Andres, 2002; Dam, 2004b; Liu, 2010; Lung, 2003; Wang, Zhou, & Wang,
2010). Findings on the choice between abbreviations and full words, however, appear to
be much more inconsistent. While some studies find their participants prefer
abbreviations (Dai & Xu, 2007; Wang et al., 2010), others find the opposite (Dam,
2004a; Liu, 2010; Lung, 2003).
In this study (see Chapter 4 and Chapter 5), a clear dominance of language notes
over symbol notes was detected in the notes taken by professional interpreters in both
directions of interpreting, corroborating the previous studies. The empirical data to date

161
conjointly support a general language-based note-taking system, rather than a symbol-
based system. This study did not find a significant difference between the distribution of
notes over abbreviation and full word. Although some individual participants in the
study showed a personal preference for one or the other, the interpreters as a group did
not share a common trend. The inconsistencies concerning the choice between
abbreviation and full word in note-taking can be subject to several factors, such as the
language pair and the type of participants, but the empirical evidence to date is too thin
to draw conclusive results.
Another choice that interpreters need to make during note-taking is the language of
notes. Traditionally, the categories used to discuss this choice have been source and
target languages. In the prescriptive stream of literature, while some publications have
given a preference to the source language (e.g., Alexieva, 1994; Kirchhoff, 1979),
others have recommended the target language (e.g., Herbert, 1952; Jones, 1998; Rozan,
1956/2002). More recently, researchers have found that the choice of language is also
affected by whether a language is the native or non-native language. Empirical studies
have been carried out to answer the questions centred on the choice of language. In
general, student interpreters have been found to take their notes predominantly in the
source language, but professional interpreters appear to have very varied language
preferences (Abuín González, 2012; Andres, 2002; Dai & Xu, 2007; Lim, 2006; Liu,
2010; Lung, 2003; Wang et al., 2010).
It is important to note that both directions of interpreting need to be examined to
simultaneously account for the source/target language factor and the native/non-native
language factor. This has been achieved in very few studies and inconsistent findings
have been reported. Dam (2004b) studied four student interpreters, of which three were
Danish native speakers and one was a Spanish native speaker. The notes were
dominated by the native language and she explained that it was because the students
chose their better mastered language. Szabó (2006) studied eight professional
interpreters with a language pair of Hungarian (native language) and English (non-
native language). She found a contradicting preference for the non-native language,
English, and pinpointed the morphological complexity and economy of writing as
impacting factors. Wang et al. (2010) studied 12 student interpreters with a language
combination of Chinese (native language) and English (non-native language). They
found a source language dominance regardless of the direction of interpreting, and
inferred that this resulted from the participants’ inadequate interpreting competence.

162
 This study analysed the notes taken by 22 professional interpreters while they
performed CI between Chinese and English (see Chapter 5). The interpreters used
significantly more English (the non-native language) than Chinese (the native language)
in both directions of interpreting. A post-experiment questionnaire was used to capture
the interpreters’ personal account of their note-taking preferences (see Appendix D).
Results of the questionnaire show that the interpreters prefer English because it can be
written down faster, using phonetic spelling and even misspelling, whereas the written
form of Chinese is not linked to its pronunciation in such a straightforward way. The
study also found that the preference for English was stronger when English was the
source language and that this preference was weakened when Chinese became the
source language. Taken together, data in this study suggest that during note-taking,
interpreters are subject to the influence from both the characteristics of the specific
language pair and the source/target language status.
What is worth noting is that the participants in this study were recruited based on
stringent criteria (see Chapter 1) and although they were native Chinese speakers, they
were based in an English-speaking country (Australia). Therefore, the participants were
experienced professionals with a strong non-native language, similar to the participants
in Szabó’s study. Their language competence and experience should also have played a
role in how they chose the note-taking language.
The case of language choice in note-taking is far from closed because the empirical
data is still limited and many of the previous findings have not been tested statistically.
Even so, combining what has been found in literature and in this study, some tentative
conclusions can be drawn. The choice of language in note-taking seems to be a function
of the combined influence from a series of factors, including (1) the nature of the
languages themselves (e.g., morphological complexity and economy of writing); (2) the
native/non-native language status; (3) the source/target language status; and (4)
interpreter characteristics (e.g., working experience and language competence). More
specifically, when one of the languages is morphologically simpler and/or easier to
write, this language would be the preferred choice regardless of the native/non-native
language status, especially when the interpreter has a strong non-native language. When
two languages do not differ too much in morphological complexity and economy of
writing, the native/non-native language status plays a leading role in determining the
language choice: interpreters prefer the native language. When the interpreting
competence is inadequate (e.g., student interpreters at a beginning stage), the language

163
choice is mainly subject to the source/target language status, with source language being
the dominant language in note-taking.

7.1.2 Note-taking and cognitive load in CI

Literature suggests that the main concerns in note-taking are the time needed for writing,
the effort of the hand and the cognitive costs (Alexieva, 1994; Allioni, 1989; Gile, 2009,
p. 178). In this study, the temporal, physical and cognitive demands of varied note-
taking choices were inspected via pen recording (see Chapter 5). The data suggest that
interpreters’ note-taking choices in Phase I of CI mainly correlate with the cognitive
demand, rather than the temporal or physical demands.
With regard to the choice of form, the cognitive demand of language notes was
found to be lower than that of symbol notes, and interpreters prefer language to symbol
even though the physical and temporal demands of language notes were higher.
Interpreters do not have a preference between abbreviations and full words, which are
associated with similar levels of cognitive load, despite the fact that the physical and
temporal demands of abbreviations are lower.
In terms of the choice of language, the cognitive demand of English notes was
found to be lower than that of Chinese notes when interpreting from Chinese to English.
In this case, interpreters prefer English to Chinese despite the facts that notes in the two
languages are similar in their physical and temporal demands and that English is the
non-source and non-native language. When interpreting from English to Chinese, no
significant difference was found between the cognitive demands of the two languages.
In this case, English seems to have become the preferred choice because of its source-
language status.
In Phase II of CI, the cognitive demands of note-taking choices were inspected via
eye tracking (see Chapter 6). The data suggest that language notes are more cognitively
demanding to read than symbol notes. Notes in these two forms differ not only in early
and fast cognitive processing such as recognising and identifying the note and its
meaning, but also in slower operations such as linking the note to the context and
producing a target speech around it. Moreover, eye movement measures including
average fixation duration, number of revisits, and skip rate suggest that reading
language notes and reading symbol notes can almost be treated as different tasks: the
differences between them are greater than those found between different reading types
(e.g., reading for comprehension and reading for translation).

164
 The study also uncovered that full words are processed for longer times than
abbreviations, not because they are more difficult to recognise or comprehend, but
mainly because they are longer. In other words, the difference in visual attention
received by abbreviations and full words is a result of the length factor. However,
considering that time is a valuable resource in interpreting and that full words do cost
more time to process during note-reading, abbreviations seem to be a more optimal
choice because they attract less visual attention and take less time to process.
In addition, data in this study show that note-reading in Chinese is less cognitively
demanding than note-reading in English when the direction of interpreting is from
English to Chinese. In this case, two factors could have played important roles. First,
Chinese is the native language of the interpreters, hence reading Chinese notes (native
language processing) is likely to be less demanding than reading English notes (foreign
language processing). Second, when interpreters read Chinese notes and produce a
target speech in Chinese, reading and speech production are in the same language. This
within-language processing is likely to be less cognitively demanding than between-
language processing, i.e., reading one language and speaking in another language. This
helps explain why the difference in cognitive demands between Chinese and English
notes was not significant in the Chinese to English direction. When the direction of
interpreting was reversed, the cognitive cost of reading Chinese notes was raised by
between-language processing (from Chinese notes to English speech).

7.1.3 Note-taking and interpreting performance

Previous studies on the relationship between note-taking and interpreting performance

have reached inconsistent conclusions (see Chapter 2). Some claim to have detected
evidence pointing to a potential link between certain note-taking features and the quality
of interpreting (Cardoen, 2013; Dam, 2007; Dam, Engberg, & Schjoldager, 2005; Her,
2001) while others fail to detect any such relationship (Dai & Xu, 2007; Wang et al.,
2010).
This study inspected the relations between interpreting performance and a series of
note-taking related factors, including the quantity of notes, the distribution of notes over
form and language, and the time spent on multi-tasking (listening/analysing and note-
writing at the same time) (see Chapter 4 and Chapter 5). In both directions of
interpreting, a higher percentage of time spent on multi-tasking correlated with a better
interpreting performance, whereas a higher percentage of English notes correlated with

165
a worse interpreting performance. In addition, some correlations that were present in
one direction of interpreting were absent in the other. In English to Chinese interpreting,
the performance quality was higher when interpreters used more symbol notes and less
language notes. In Chinese to English interpreting, the interpreting performance was
better when the quantity of notes was higher. It has been made evident in this study that
although some findings on the relation between note-taking and interpreting
performance are consistent in both directions of interpreting, there are also critical
differences between the two directions.

7.2 Note-taking, cognitive load and interpreting performance: the interplay and
implications

The interplay between note-taking, cognitive load and interpreting performance are
summarised in Table 7.1. Taken together, the data show that interpreters take proactive
actions to reduce cognitive load during Phase I of CI, but the choices they make
sometimes leave the extra burden to Phase II and are at the expense of interpreting
quality. This means that the choices made by interpreters with the intention to ease the
cognitive system are not always optimal when all factors are taken into account.

Table 7.1 The interplay between note-taking, cognitive load and interpreting
performance

Choice of form Choice of language

Language vs. Abbreviation vs. Chinese vs.
symbol full word English
Physical and Language > Abbreviation < Chinese ≈
temporal demands Symbol Full word English
Cognitive load in Language < Abbreviation ≈ Chinese >
Phase I Symbol Full word English1
Preferred note-
Language No preference English
taking choice
Cognitive load in Language > Abbreviation < Chinese <
Phase II Symbol Full word English2
Relation with Less language, more Less English, better
No relation
performance symbol, better performance3 performance4
Notes:
1. This result is only significant in the C-E task;
2. This result is only significant in the E-C task;
3. This result is only significant in the E-C task;
4. No significant correlation is found between the choice of Chinese notes and the interpreting
performance.

166
 When choosing the form of note-taking, interpreters would benefit more from using
more symbols considering that they can save effort both in note-writing and note-
reading, and that they have a positive contribution to interpreting performance.
Interpreters could also use more abbreviations, especially considering that abbreviations
save time during note-writing and that they are not more difficult to process than full
words in note-reading.
In interpreter education, more emphasis should be put on teaching the students how
to effectively use symbols and abbreviations. The training should have two outcomes: a
decreased level of cognitive cost and an increased amount of usage. It is not enough to
merely point to the highly individualised nature of symbols and abbreviations in note-
taking and leave it to the students to develop a system of their own. The trial and error
process is usually overly painstaking and students could potentially set out on the wrong
track, even leading to under-developed note-taking skills towards the end of training. It
is therefore advisable to provide a seminal set of symbols and abbreviations rich enough
for the students to start with, together with rules on how to develop new symbols and
abbreviations based on the given sample. The students can then spend more time on
practising using the symbols and abbreviations instead of (re)inventing them.
When choosing the language of note-taking, many factors are relevant, such as the
morphological complexity, economy of writing, native/non-native language status,
source/target language status, and interpreting competence. There is no absolute correct
answer when it comes to which language should be used. Specific strategies should be
prepared to cater to different language pairs and stages of competence development.
What has also been made evident in this study is that there are some differences
between the two directions of interpreting. Teachers, students and professional
interpreters should mind this factor in their pedagogical planning, practices, and
continued professional development. Diversified approaches should be taken to cope
with the specific challenges brought by different interpreting directions.

7.3 Strengths and limitations of the study

Overall, the research has three main strengths. First, the research enhances the
theoretical framework of studying the cognitive aspects of interpreting by presenting an
in-depth discussion on the construct of cognitive load in interpreting and its
measurement. Building on both translation and interpreting studies and the research on

167
mental workload and Cognitive Load Theory, this study gives an operationalised
definition to cognitive load in interpreting. A detailed discussion on the relevant
variables and their interactions was presented. This discussion can not only help identify
which factors potentially impact cognitive load in interpreting, but also provide a
guideline in terms of which variables to control for to create better working conditions
for interpreters and to obtain more valid experimental results in research. This study
also elaborated on the measurement of cognitive load, especially how to measure the
cognitive load in CI and note-taking, providing a good reference for future work to be
carried out on the topic.
Second, this study makes a methodological contribution to language processing
research (especially process-oriented translation and interpreting research) by
triangulating the methods of pen recording, eye tracking, and voice recording. Through
a careful design, diverse sources of data during the entire interpreting process were
recorded in a synchronised way. This novel approach is likely to be beneficial to future
investigations on the cognitive aspects of translation and interpreting, and more
generally to language processing research.
Third, this research has contributed ample empirical data to reveal traces of
processing efforts in note-taking and CI. It proposes a unique array of indicators for the
measurement of the physical, temporal and cognitive demands of note-writing and note-
reading, indicators that can be useful in future studies on related topics.
Despite the strengths outlined above, the research has several limitations. First,
although a total of 26 professional interpreters participated in the study, valid data was
collected from only 22 participants during pen recording (15% data loss) and 18
participants during eye tracking (31% data loss1). Purposive sampling was carried out
through the professional network of the researcher, giving rise to exclusion bias and
placing limits on the generalisation of the findings.
Second, this research only examined the correlations between the note-taking
choices and a series of indicators of cognitive load. The current design does not allow
any causal relations to be tested between the variables. Even though the findings are
helpful for generating hypotheses for future studies, they cannot be taken as evidence of
how cognitive load changes because of note-taking or the other way around.

1
This kind of data loss is consistent with eye tracking studies of this nature (see, for example, O'Brien
(2009) and Section 6.5.1).

168
 Third, this study only involved one language pair (Chinese and English) and one
type of interpreter (professional interpreters). It remains to be tested whether the
findings can be generalised to other language pairs and to other types of participants.
Fourth, the eye tracker used in this study has a relatively low sampling rate (60 Hz)
and is not compatible with the Eye and Pen software that is used to collect pen
recording data. The synchronisation of the data requires laborious manual work.
Although this particular eye tracker has been selected for a number of reasons (e.g.,
allowing free head and hand movement, high ecological validity, availability, see
Chapter 1), the study would have benefited from an eye tracker with a higher sampling
rate and compatibility with the Eye and Pen software.

7.4 The way forward

The findings and limitations of this research together have pointed to some interesting
directions for future research. To begin with, it would be interesting to see how
replicable this study is. A promising approach is to test whether the same results can be
obtained using a different language combination or a different type of participant. A
further step can be taken to see how the results change when tested with a cluster of
participant groups with varying degrees of experience.
Another appealing topic is the causal relations between the variables investigated in
this research. The observed correlations in this study form a good starting point for
hypotheses to be formed. On the one hand, how does a change in note-taking choice
affect the level of cognitive load? On the other hand, how will interpreters’ note-taking
choices change as a result of increased or decreased cognitive load?
A third issue that is worth investigating is other feasible cognitive load measures.
Within the scope of a PhD research project, this study has tested the usability of a series
of cognitive load measures. However, there are many other measures that are potentially
useful for translation and interpreting studies. Explorations on whether certain measures
are effective and how to apply the measures can be a productive line of research.
A fourth valuable avenue of research involves testing other apparatus and designs.
Are there any other types of pen recording and eye tracking apparatus that would be
useful in this kind of research? Can a different design increase the quality of data,
reduce data loss, and cut down manual work in data analysis? Answers to these

169
questions can help consolidate the foundations for experiments to be carried out on
similar topics in the future.
Research on the cognitive aspects of interpreting is experiencing an exciting start.
Hopefully more interested researchers will join the effort because this branch of
interdisciplinary research can not only benefit the field of interpreting but also help
enhance our understanding of language processing and the human mind.

7.5 References

Abuín González, M. (2012). The language of consecutive interpreters’ notes:

Differences across levels of expertise. Interpreting, 14(1), 55–72.
doi:10.1075/intp.14.1.03abu
Alexieva, B. (1994). On teaching note-taking in consecutive interpreting. In C. Dollerup
& A. Lindegaard (Eds.), Teaching translation and interpreting 2: Insights, aims,
visions (pp. 199–206). Amsterdam: John Benjamins.
Allioni, S. (1989). Towards a grammar of consecutive interpretation. In L. Gran & J. M.
Dodds (Eds.), The theoretical and practical aspects of teaching conference
interpretation (pp. 191–197). Udine: Campanotto.
Andres, D. (2002). Konsekutivdolmetschen und notation [Consecutive interpreting and
note-taking]. Frankfurt: Peter Lang.
Cardoen, H. (2013). The Effect of Note-taking on Target-text Fluency. In G. González
Núñez, Y. Khaled, & T. Voinova (Eds.), Emerging Research in Translation
Studies: Selected Papers of the CETRA Research Summer School 2012 (pp. 1-
22). Leuven: CETRA.
Dai, W., & Xu, H. (2007). 汉英交替传译过程中译员笔记特征实证研究——以职业
受训译员和非职业译员为例 [An empirical study of the features of interpreters'
notes in Chinese-English consecutive interpreting: The examples of
professionally trained and unprofessional interpreters]. Foreign Language
Teaching and Research, 39(2), 136–144.
Dam, H. V. (2004a). Interpreters' notes: On the choice of form and language. In G.
Hansen, K. Malmkjær, & D. Gile (Eds.), Claims, changes and challenges in
Translation Studies (pp. 251–261). Amsterdam: John Benjamins.
Dam, H. V. (2004b). Interpreters' notes: On the choice of language. Interpreting, 6(1),
3–17. doi:10.1075/intp.6.1.03dam

170
Dam, H. V. (2007). What makes interpreters’ notes efficient?: Features of (non-
)efficiency in interpreter's notes for consecutive. In Y. Gambier, M. Shlesinger,
& R. Stolze (Eds.), Doubts and directions in Translation Studies: Selected
contributions from the EST Congress, Lisbon 2004 (pp. 183–197). Amsterdam:
John Benjamins.
Dam, H. V., Engberg, J., & Schjoldager, A. (2005). Modelling semantic networks on
source and target texts in consecutive interpreting: A contribution to the study of
interpreters’ notes. In H. V. Dam, J. Engberg, & H. Gerzymisch-Arbogast (Eds.),
Knowledge systems and translation (Vol. 7, pp. 227–254). Berlin: Mouton de
Gruyter.
Gile, D. (2009). Basic concepts and models for interpreter and translator training
(revised edition). Amsterdam: John Benjamins.
Her, H. (2001). Notetaking in basic interpretation class: An initial investigation. Studies
of Translation and Interpretation, 6, 53-77.
Herbert, J. (1952). Manuel de l'interprète: comment on devient interprète de
conférences [Interpreter's manual: How to become a conference interpreter].
Geneva: Georg.
Jones, R. (1998). Conference interpreting explained. Manchester: St. Jerome.
Kirchhoff, H. (1979). Die notationssprache als hilfsmittel des konferenzdolmetschers im
konsekutivvorgang [The language of note-taking as a tool for the conference
interpreter in consecutive interpreting]. In W. Mair & E. Sallager (Eds.),
Sprachtheorie und sprachpraxis [Language theory and language practice] (pp.
121-133). Tübingen: Gunter Narr.
Lim, H.-O. (2006). A post-mortem of note-taking. Forum, 4(2), 89–112.
Liu, J. (2010). 英语专业本科学生汉英交传笔记特征——一项基于学生交传笔记的
实证研究 [Note-taking characteristics of English majored undergraduates in
Chinese-English consecutive interpreting: an empirical study based on students'
consecutive interpreting notes]. Foreign Language World, 2, 47–53.
Lung, R. (2003). Taking "notes" seriously in the interpretation classroom. In Á.
Collados Aís, M. M. Fernández Sánchez, & D. Gile (Eds.), La evaluación de la
calidad en interpretación: investigación [The evaluation of quality in
interpretation: Research] (pp. 199–205). Granada: Comares.

171
O'Brien, S. (2009). Eye tracking in translation process research: Methodological
challenges and solution. In I. M. Mees, F. Alves, & S. Göpferich (Eds.),
Methodology, technology and innovation in translation process research: A
tribute to Arnt Lykke Jakobsen (pp. 251–266). Frederiksberg: Samfundslitteratur.
Rozan, J.-F. o. (1956/2002). Note-taking in consecutive interpreting (A. Gillies, Trans.).
Cracow: Tertium Society for the Promotion of Language Studies.
Wang, W., Zhou, D., & Wang, L. (2010). 口译笔记特征与口译产出质量实证研究
[An empirical study of note-taking characteristics and output quality in
interpreting]. Foreign Language World, 4, 9–18.

172
Appendix A & B of this thesis have been removed as they may contain sensitive/confidential content
Appendix C: Demographic questionnaire

* 1. Participant ID

* 2. Professional background

What is(are) your full-time job(s)

What is(are) your part-time job(s) (if any)

Which city do you most frequently work in as a consecutive interpreter? (Please name

only one city)

How many NAATI working language combinations do you have? Please specify. (e.g.

Two, Mandarin/English, and Cantonese/English)

How many years have you been practicing as a professional consecutive interpreter?

(e.g. about 4 years)

Please give an estimate of the number of occasions you've provided consecutive

interpreting service in the past 12 months. (e.g. about 12 occasions)

* 3. What is the highest level of education you have COMPLETED?

Bachelor's degree

Master's degree

Doctoral degree

Other (please specify)

* 4. What kind of interpreting training and education have you received? (You
may indicate more than one)

Intensive interpreting training course

Interpreting diploma

Postgraduate-level interpreting degree

Other (please specify)

179
* 5. What time of the day do you function the best?

180
Appendix D: Post-experiment questionnaire

Participant ID:

1. On a scale from 1 (extremely bad) to 7 (extremely well), how well rested are you
before coming to the experiment? _____
Extremely Extremely
bad well
1 2 3 4 5 6 7

2. On a scale from 1 (never heard of) to 7 (extremely familiar), please rate how familiar
you are with the topic of the source speeches. By familiarity, we mean how often you
have come across the topic, as well as how well you know the topic and what it is about.
Topic of the Chinese speech: _____
Topic of the English speech: _____
Never Extremely
heard of familiar
Topic of the Chinese
speech: How to
1 2 3 4 5 6 7
purchase property in
Australia
Topic of the English
speech: How to
1 2 3 4 5 6 7
register a business in
Australia

3. On a scale from 1 (strongly disagree) to 7 (strongly agree), do you agree that the
digital tablet and pen are sufficiently similar to real pen and paper, and therefore did
NOT affect your note-taking behaviour? _____
Strongly Strongly
disagree agree
1 2 3 4 5 6 7

4. Have you received specialised training on note-taking? If so, do you think it has
affected your note-taking style (e.g. use of symbol, choice of language)?

181
5. What is your preference on the form of note-taking? That is to say, do you usually
use more language or symbol? Do you usually use more abbreviations or full words?
Why?

6. What is your preference on the language of note-taking? That is to say, do you

usually use more source or target language? Do you usually use more Chinese or
English? Why?

7. Is the eye tracker (the glasses) comfortable? Do you think wearing the eye tracker has
affected the main task of interpreting?

8. Do you think the instructions of the experiment are clear enough?

9. Do you have any comments on the interpreting tasks (e.g. topic, length, segmentation,
difficulty, quality of audio)?

10. Do you have any other comments about today’s experiment?

Thank you very much!

182
Appendix E: Scripts for the two interpreting tasks

Chinese to English task

Topic: 如何在澳大利亚买房
Segment 1
许多想在海外投资的人如今将澳大利亚视为排名第一的投资地。更好的气候，
更清新的空气，买得起的房子以及大量的投资机遇带来的诱惑让人血脉偾张。
如果你已经决定要在澳洲买房，那么恭喜你。买房可以为你和你的家人提供
未来的资产保障并创造财务上的自由。今天我将针对在澳洲买房谈一谈我的几点
建议。
第一点建议，在买房前先要解决好法律上的问题。如果不是澳洲国籍，或者
永久居民，你很可能要从澳洲的外国投资审查委员会获得批准才能购房。如果不
想申请批准，不妨试试购买新建的房产。买新房不太需要经过审查，但下手要快，
因为新房非常抢手。
Segment 2
第二点建议，找出澳洲哪里的房产最受欢迎。大城市比如悉尼、珀斯、布里
斯班和墨尔本，这些地方的房价升幅最大。这些城市还可以提供最佳的工作机遇
和较高的租房需求。除此之外，沿海地区由于旅游业的蓬勃发展，也可以提供获
得收入的良好机会。这也是很多中国移民选择在黄金海岸和凯恩斯这样的热门旅
游地做起生意的原因。如果你想让资金更细水长流一些，就可以看看更靠近郊区
的地方。但要记住，郊区在澳大利亚基本就等同于偏远的地方，因此在着手购买
前一定要做好研究。
第三点建议，尽可能多看一些房产。通过在多个地区看多套房产，你可以对
各个地区及其房产的实际价值有更好的理解。你可以去网上注册几个房产网站。
当你感兴趣的地区有房产可买的时候，这些网站就会提醒你。你可以重点关注那
些已经在市场上销售超过六个星期的房产，或者是已经降价的房产。这表示购买
这些房产的竞争压力比较低，并且你可能有讨价还价的余地。在你看过的房产中，
你可以对其中的几个报价。可能你的报价中只有两三个会被接受，然后你就可以
从中挑选你最满意的房产进行购买了。许多著名的房产投资人会为了购买一处房
产而看一百多套房。
Segment 3

183
 第四点建议，跟你的邻居聊聊天。多去走走，敲敲你邻居的门，跟他们说说
话，从他们口中尽可能地获取当地的相关信息。有的时候这并不是要看邻居们具
体说了什么，而是要看看你的邻居是谁。通过跟住在当地的人交流，你可以挖掘
出很多当地街道和社区的信息，并了解到附近住的是什么样的人。
在澳洲买房尽管花费很高，但投资潜力是非常好的，尤其是在市区买房。举
个例子，悉尼的房价在 2000 年一年增长了 50%。而东海岸在 1997 至 2003 年间
的总增长更是达到了惊人的 112%。当然现在涨势已经没有那么凶猛了，但投资
者不应因此丧失积极性。在西海岸的珀斯，房价仍然以每年 15%的速度在增长。
如果打算将房子租出去，你的预计收入可能达到每年 3%到 6%。你还在等什么呢？

184
English to Chinese task
Topic: How to register a business in Australia
Segment 1
Today I'd like to share some information about registering a business in Australia.
The number of Australians starting their own businesses is rocketing, and this
entrepreneurial spirit looks set to continue in the near future. More than 520,000 new
businesses have been registered during 2013, a rise of 8% on 2012 and a record high.
The vast majority of those, at 95%, were small ventures launched by individuals. If you
want to join them, today is your best opportunity to learn some useful information.
However, the first thing you need to know is that not everyone is entitled to register
a business. Therefore, you must assess whether you are eligible to do so. These are
some of the questions you need to ask yourself. Do you intend to make a profit from
your activities? Is the activity of your business of a commercial nature? Are you
committed to running a business? Your answer to these questions needs to be "yes"
before you can register a business.
Segment 2
The second thing to know is the government organization that you'll deal with. The
Australian Securities and Investments Commission, or ASIC, is an independent
government body that regulates Australia's corporations, markets and financial services.
It is the government body that you'll deal with the most when registering your business.
The ASIC website is very convenient to use for various applications, and you will learn
about many of its functions today.
To begin registering your business, you need to apply for an ABN. An ABN is
short for Australian Business Number. It is a unique nine digit number, which helps
government organizations identify your business. An ABN allows your trading partners
to easily confirm your company's details for ordering and billing. This process can be
done on your own, through a corporate service provider, or by an accountant on your
behalf. If you decide to register on your own, you can do this either online or by lodging
a form. If you register online, you usually receive your ABN immediately. However,
some approvals may take up to twenty-eight days.
Once you have an ABN, you should do some business name research. Check the
ASIC website to see if your desired name is available, because duplicate names will not
be approved. For example, if you wish to register "Garden Services" as your business

185
name, and this is already taken, you may need to try a slightly different name like
"Jack's Garden Services".
Segment 3
The next step is to do some trademark homework to ensure that you are not
infringing on anyone's intellectual property. Remember, the fact that a name is not
registered doesn't mean it's not trademarked. It's also a good idea to check if your
preferred domain name is available. You wouldn't want to have a web address too
different from your business name.
Once you've done your research and decided on your desired name, it is time to
apply. To apply for your business name, you will need to provide your ABN, your
personal details such as date of birth and country, your email address, residential
address, and business address. Business name registration is relatively simple. You can
do it on your own through the ASIC website. Or if you don't like to deal with
government organizations, you can have an authorized third party apply for you.
The turnaround time for business registration is about five working days. When
your name is confirmed by the ASIC, you'll receive a letter of confirmation, which
signals that you can register your domain, start creating your logo and business cards,
and anything else you may need to do to get your new business up and running.
Hopefully you are now ready to take the next step. Good luck and I hope your new
venture is successful.

186