語速對英中逐步口譯筆記及準確度的影響

全文

(1)國立臺灣師範大學翻譯研究所碩士論文 Graduate Institute of Translation and Interpretation National Taiwan Normal University Master’s Thesis. 語速對英中逐步口譯筆記及準確度的影響 The Impact of Fast Speech Rate on Note-taking and Accuracy in E-C Consecutive Interpreting. 戴佑安 Tai, You-an. 指導教授:汝明麗博士 Advisor: Ju, Ming-Li, Ph.D.. 中華民國一○九年二月 Feb, 2020.

(2) Acknowledgement My deepest appreciation goes to Prof. Elma Ming-Li Ju for her insightful suggestions and patient guidance throughout the completion of my thesis. My sincere thanks and gratefulness also go to my thesis committee members, Prof. Tze-Wei Chen and Prof. Albert Cheng whose meticulous comments were an enormous help to me. I would also like to thank my family, close friends, and peers for their moral support and warm encouragements. Lastly, I would like to express my gratitude to the participants in this study, without whom this thesis could not have been written.. i.

(3) Abstract This study aims to find out the impact of fast speech rate (FSR) on note-taking strategies and accuracy of student interpreters performing consecutive interpreting (CI). Twenty students studying in the interpreting track of graduate institutes of translation and interpretation were recruited to participate in a CI experiment and a retrospective interview. During the experiment, each participant interpreted a speech delivered at a slower speech rate, and one delivered at an FSR, both from English (their B language) to Chinese (their A language). The results of the quantitative and qualitative analyses reveal that: a) In the FSR setting, the student interpreters significantly wrote more language in the language/symbol ratio; fewer horizontal lines per proposition; and fewer total note units per proposition. b) In the FSR setting, the student interpreters significantly scored lower in the accuracy rating. c) The ten student interpreters who scored higher than the other half of the student interpreters in the FSR setting used more horizontal lines in their notes. d) The main strategies adopted by the student interpreters to deal with FSR were: focus on listening and understanding the logic and structure of the speech, and reduce note quantity and complexity. Based on the findings, this study concludes that a possible way for student interpreters to enhance CI performance in the event of FSR is to prioritize logic comprehension of the source speech, select important messages, and emphasize the relation between messages on notes.. Keywords: consecutive interpreting, fast speech rate, note-taking, accuracy. ii.

(4) 摘要本研究旨在探討語速對於學生口譯員進行逐步口譯時的筆記策略與準確度的影響。研究對象為二十名於翻譯研究所口譯組修習的學生口譯員（中文 A、英文 B），實驗內容為每位學生口譯員針對一篇中等語速與一篇快速語速的講者發言進行英進中逐步口譯，在實驗之後進行回溯訪談。綜合質性與量性分析，發現： (1) 學生口譯員在快速語速之下對比符號而言增加了語言的使用；減少了每命題的分隔線數；減少了每命題的總筆記數量 (2). 學生口譯員在快速語速之下準確. 度分數較低 (3) 快速語速之下表現較好的十位學生口譯員所使用的分隔線數較多 (4) 學生口譯員應對快速語速而主動採取的策略主要為：專注在邏輯理解，以及減少筆記數量及筆畫。根據研究結果發現，學生口譯員在面對快速語速時，著重在聽懂原文邏輯關聯、篩選重要訊息，並將訊息之邏輯關聯反應於筆記上，將有助於口譯表現的提升。. 關鍵字：逐步口譯、語速、筆記、準確度. iii.

(5) Table of Contents Acknowledgement........................................................................................................i Abstract........................................................................................................................ii 摘要..............................................................................................................................iii Table of Contents.........................................................................................................iv List of Tables...............................................................................................................vii List of Figures............................................................................................................viii Chapter 1 Introduction................................................................................................1 1.1 Research Background......................................................................1 1.2 The study and Thesis Organization................................................5 Chapter 2 Literature Review.......................................................................................7 2.1 Overview of Consecutive Interpreting (CI) ..................................7 2.1.1 Process of CI...........................................................................8 2.1.2 Processing capacity saturation and problem triggers............10 2.2 Fast Speech Rate (FSR).................................................................13 2.2.1 The definition of FSR............................................................14 2.2.2 FSR’s Impact on Interpreting................................................17 2.2.3 Strategies to cope with FSR in interpreting...........................19 2.3 CI Note-taking.................................................................................23 2.3.1 Prescriptive studies on note-taking skills...............................23 2.3.2 Empirical studies on note-taking preferences........................28 2.3.3 Empirical studies on note-taking preferences and interpreting performance...........................................................................30 2.4 Interpreting Performance...............................................................32 2.5 Research Questions..........................................................................34. iv.

(6) Chapter 3 Research Method......................................................................................36 3.1 Participants......................................................................................38 3.2 Speech Materials.............................................................................40 3.2.1 Speech text preparation and pilot test.....................................40 3.2.2 Speaker....................................................................................47 3.2.3 Speech Recordings..................................................................49 3.3 Procedures........................................................................................50 3.3.1 CI Task....................................................................................50 3.3.2 Retrospective interview...........................................................53 3.4 Data Preparation of CI Results......................................................54 3.4.1 Categorization of note units....................................................55 3.4.2 Accuracy Rating......................................................................58 Chapter 4 Findings.....................................................................................................61 4.1 CI Results..........................................................................................61 4.1.1 Note-taking preferences..........................................................62 4.1.2 Accuracy ....................... .........................................................66 4.1.3 Note-taking preferences and accuracy in the FSR setting.....................................................................................68 4.2 Interview Results..............................................................................72 4.2.1 FSR difficulty perception........................................................73 4.2.2 FSR challenges to note-taking.................................................73 4.2.3 Note-taking preferences..........................................................80 4.2.4 Note-taking strategies and techniques.....................................83 4.2.5 FSR’s impact on accuracy ......................................................87 Chapter 5 Discussion..................................................................................................92 5.1 Note-taking Preferences and Strategies.........................................92 5.1.1 The gap between strategy and execution................................92 5.1.2 Processing capacity saturation................................................95 v.

(7) 5.2 Accuracy and Note-taking Preferences..........................................97 5.2.1 FSR’s impact on accuracy.......................................................97 5.2.2 More segmentation for better accuracy...................................99 5.3 The Difference Between Y1 and Y2 Students...............................101 5.4 Text-related Factors of Difficulty..................................................102 5.4.1 FSR for CI versus SI...............................................................102 5.4.2 Difficulty parameters besides FSR.........................................103 Chapter 6 Conclusions.............................................................................................106 6.1 Summary of Study...........................................................................106 6.2 Pedagogical Implications................................................................109 6.3 Contributions and Limitations.......................................................110 6.4 Future Directions.............................................................................112 References.................................................................................................................114 Appendices................................................................................................................124 Appendix i. Information Sheet............................................................124 Appendix ii. Practice Speech...............................................................125 Appendix iii. SSR Speech.....................................................................126 Appendix iv. FSR Speech.....................................................................127 Appendix v. Form of Consent for Participants..................................128 Appendix vi. Basic Information Questionnaire..................................129 Appendix vii. ANOVA on the Effect of Grade and SR on Note-taking Preferences........................................................................130 Appendix viii. CI Results- SSR setting.................................................132 Appendix ix. CI Results- FSR setting...................................................134. vi.

(8) List of Tables Table 2.1 Five tasks of CI proposed by Liu (2008).......................................................7 Table 2.2 The Efforts in Gile’s Effort Model of CI (2009)...........................................9 Table 3.1 The difficulty level of version-one and version-two speech materials........43 Table 3.2 Difficulty assessment of the speech materials.............................................46 Table 3.3 Basic information of the speaker.................................................................48 Table 3.4 A 4-point scale for accentedness judgments................................................48 Table 3.5 Treatment protocols.....................................................................................51 Table 3.6 Categories of note units (Parts were adapted from Chen, 2017) ................56 Table 4.1 ANOVA on the effect of SR on language/symbol ratio...............................63 Table 4.2 Descriptive statistics of language/symbol ratio...........................................63 Table 4.3 ANOVA on the effect of SR on full word/incomplete word ratio...............64 Table 4.4 ANOVA on the effect of SR on English/Chinese ratio................................64 Table 4.5 ANOVA on the effect of SR on horizontal lines per proposition.................65 Table 4.6 Descriptive statistics of horizontal lines per proposition.............................65 Table 4.7 ANOVA on the effect of SR on total note units per proposition..................66 Table 4.8 Descriptive statistics of total note units per proposition..............................66 Table 4.9 ANOVA on the effect of grade and SR on accuracy....................................67 Table 4.10 ANOVA on the effect of SR on accuracy...................................................67 Table 4.11 Descriptive statistics of accuracy...............................................................68 Table 4.12 ANOVA on the effect of FSR score-group on language-symbol gap........69 Table 4.13 ANOVA on the effect of FSR scoring group on full-incomplete word gap..............................................................................................................70 Table 4.14 ANOVA on the effect of FSR score-group on English-Chinese gap..........71 Table 4.15 T-tests on the effect of FSR scoring group on all categories......................72. vii.

(9) List of Figures Figure 3.1 Categorization of note units........................................................................55. viii.

(10) Chapter 1 Introduction 1.1 Research background Speech rate (SR) is defined as “the speed of message delivery” (Pöchhacker, 2004, p.129). SR is a determinant factor of listening comprehension of a speech (Kopparapu, 2013; Rodero, 2012). That is, SRs that are either too fast or too slow can both adversely impact listeners’ understanding. While slower SRs generally provide more time for listeners to process information (Megehee, Dobie, & Grant 2003), SRs that are too slow are hard to retain attention on (Rodero, 2012) and cause difficulty in comprehension (Peltekov, 2017). On the other hand, very fast SRs are hard to keep up, too, because too many words are presented within a short amount of time without enough pauses for listeners to process and understand the message the speaker wants to convey (Nikitina, 2011; Rodero, 2012). To enable listeners to fully understand the message, it is best for speakers to speak at an optimal speech rate (Munro & Derwing, 2001). The optimal rate of public speaking in English, according to Arina Nikitina, is around 140-150 word per minute (wpm) (Nikitina, 2011). To conference interpreters, fast speech rate (FSR) is known to be especially challenging. In the Statistical World Report 2009 by AIIC, fast speeches were identified as the biggest stressor felt by interpreters (Neff, 2011). FSR is difficult to cope with because it increases processing capacity requirements for interpreting (Gile, 2009). In Gile’s Effort Models, different Efforts were introduced as processing capacity demands on interpreters. Each Effort has its requirements on interpreters’ attention, and thus interpreters have to allocate their attention among them. Gile (1995) also believes that 1.

(11) interpreters only have a limited supply of processing capacity when performing interpretation; therefore, when one or more of these Efforts require greater capacity than available, the total capacity at the interpreter’s disposal may be insufficient, thus resulting in a regression in the interpreter’s output performance. Exceedingly fast SR means that more information will have to be processed in the same amount of time, which will lead to the afore-mentioned elevated capacity requirements. Since an interpreter’s attention is allocated to meet the requirements of each Effort, which is different from only listening and understanding a speech, the optimal rate of input for interpretation is lower than that for just listening and comprehending (Vančura, 2013). This is probably why according to Setton and Dawrant (2016b), 120-140 wpm is considered a “moderate” SR for interpreters, while 140-160 wpm is considered “challenging.”1 If SR exceeds a certain limit, interpreters will not be able to render the message completely (Li, 2010). This can happen to interpreters performing simultaneous and consecutive interpreting alike (Chang, 2009; Gerver, 2002 [1969]; Ribas, 2012; Wu et al., 2010). In the context of consecutive interpreting (CI), a few experimental studies show that when interpreters encounter a FSR, accuracy decreases, omission increases, and summarization is used more (Ribas, 2012; Wu et al., 2010). This suggests that FSR makes it challenging for CI interpreters to keep every message full and complete until output production. Some researchers have categorized CI into several phases. For example, Giles’ Effort Model, when applied for CI, contains a “listening and 1. Setton and Sawrant (2016b) pointed out that measuring SR by wpm may yield different results for different languages. Though it was not specified in Setton and Sawrant to which languages do these measurements apply, it can be inferred that they can at least be applied to English. 2.

(12) note-taking” phase and a “target-speech production” phase (2009, p. 175). In the first phase, the interpreter has to exert Efforts in “listening and analysis,” “note-taking,” “short-term memory operations,” and “coordination.” In the second phase, Efforts include “remembering,” “note-reading,” and “production.” Liu (2008, p. 19) proposed a similar model, categorizing the procedures of CI into “listening,” “analysis and comprehension,” and “short-term memory operations and/or note-taking” in the first phase, and “remembering and/or note-reading” and “production” in the second phase. Wu et al. (2010) stated that since every step in CI is connected, a fast speaker will have a chain effect, resulting in the interpreter’s failure to render complete and/or accurate messages. One of the important parts in CI is note-taking. According to Gile (2009), different Efforts compete with each other for the limited processing capacity available in interpretation. Note-taking and listening are both performed in the first phase of CI, in which focus of attention shifts between listening and note-taking at a very fast pace (Liu, 2008). Taking notes occupies a certain amount of processing capacity; therefore, it will more or less affect the listening Effort (Liu, 2008). However, note-taking can be an effective tool in CI if used effectively, because for long CI, the length of each interpreting segment is usually longer, which makes it difficult for interpreters to memorize all the message (Liu, 2008). CI notes help interpreters to remember messages from the speech, serving as “memory reinforcers” (Gile 2009, p. 178) or “visual cues” (Liu, 2008, p. 27). Effective notes may boost “remembering” operations and lower required processing capacity for “remembering” in the second phase of CI (Gile, 2009, 3.

(13) p. 176). However, poor notes will hamper Efforts needed in the second phase of CI. Thus, note-taking effectiveness can potentially make or break a CI performance. According to Daniel Gile’s observations, CI is commonly used in Asia’s interpreting markets (Gile, 2001). In Taiwan, in particular, CI has been one of the most common modes of interpretation for both freelance and in-house interpreters. According to a survey conducted by Government Information Office (GIO), Executive Yuan (2004), among all the interpretation work done by freelance interpreters in Taiwan in 2001, the ratio of CI to SI was 38:57, while whispering was much less frequent than the other two modes. In addition, a survey on freelance translators and interpreters in 2012 shows that among all the interpreting assignments, CI was the most frequently used mode of interpretation, and 27.63% of the surveyed interpreters claimed that CI accounted for over 81% of all of their interpreting assignments (Chen et al., 2012). As for in-house interpreters, among all the interpreting assignments, whispering and CI are reported to be the most common (Li, 2016). In view of the important role note-taking plays in CI as previously mentioned, as well as the prevalent use of CI in Taiwan, it makes sense to investigate into note-taking strategies to cope with FSR. Nevertheless, research on this issue is sparse. Past research has focused on CI note-taking skills and strategies, or strategies to deal with the challenge of FSR in SI. The former, for instance, includes the early study by Rozan (1956) in which a note-taking system was proposed, and more recently, studies by Gillies (2005) and Chuang (2008), in which note-taking guidelines and techniques were introduced. Recent studies of CI note-taking have focused on the relation between note-taking preferences 4.

(14) and CI performance. Most of them investigated into note quantity, symbols and abbreviations, or source and target languages or A and B languages in general CI occasions. As for how to deal with FSR, Barghout, Rosendo, and García (2015), Li (2010), and Setton and Dawrant (2016b) listed various approaches in the context of SI. Some strategies, such as summarization, can possibly be applied to CI and note-taking. However, other strategies such as speeding up in output production and switching off the microphone are not entirely applicable to CI without some modifications. Overall, to date, no study has been done specifically on finding possible note-taking strategies for coping with FSR in CI. In this study, the researcher intends to fill the gap, and propose some possible note-taking strategies that could be adopted in coping with FSR.. 1.2 The study and Thesis Organization This study aims to analyze and compare the CI notes and accuracy of student interpreters in the first or second year of translation and interpretation graduate programs. The focus is on how they deal with a slow SR (SSR) and a FSR and what strategies may be applied to coping with FSR when performing CI in the note-taking phase. The specific research questions addressed by the study are: 1. How does FSR affect student interpreters’ note-taking preferences in CI? 2. How does FSR affect student interpreters’ note-taking strategies in CI? 3. How does FSR affect student interpreters’accuracy in CI?. 5.

(15) 4. Do certain note-taking preferences correlate with better CI accuracy in FSR occasions? The remainder of this thesis is structured as follows. Chapter 2 reviews the existing literature on CI note-taking, the impact of FSR on conference interpreting, and interpreting performance. Next, Chapter 3 introduces the methods and procedure of the data collection, which includes a CI experiment and a retrospective interview. After that, Chapter 4 presents the findings of the data collection, followed by a discussion on the findings in Chapter 5. Finally, Chapter 6 concludes the thesis and provides suggestions for future work.. 6.

(16) Chapter 2 Literature Review This chapter provides a review of previous studies that are relevant to the purpose of this thesis. It serves as the theoretical foundation for both the study design and the analytical approach of the thesis. Beginning with an introduction to consecutive interpreting (CI) and its problem triggers, this chapter than looks at the impact of fast speech rate (FSR) on information processing and rendering. Since only a handful of studies to date have touched upon the impact of FSR on CI, the researcher included a review of various studies of the impact of FSR on SI or on interpreting in general as well. A review of previous studies on CI note-taking and interpreting performance and quality assessment was presented as a reference for assessing interpreting performance in this thesis. Finally, this chapter ends with proposed research questions aimed at closing the research gap. 2.1 Overview of Consecutive Interpreting (CI) Consecutive interpreting (CI) is one of the most common modes of interpreting. When interpreters perform CI, they first listen and store the message they hear by remembering or taking notes, and deliver the message the speaker just conveyed in another language once the speaker finishes a segment. There are two modes of CI, short consecutive and long consecutive, which differ from each other by whether it requires interpreters to take notes (Liu, 2008). In performing short consecutive, interpreters usually do not have to resort to note-taking to help them recall the message later except for writing down information such as names and numbers, because they only have to 7.

(17) interpret one to three sentences at a time. In performing long consecutive, interpreters generally have to deal with segments with more than three sentences, hence the need to take notes. The mode under investigation in this thesis is long consecutive, and the phrase “consecutive interpreting” or “CI” will be referring to long consecutive by default in the remaining part of the study for the purpose of being concise. 2.1.1 Process of CI The process of CI has been suggested by Liu (2008) and Gile (2009). In a nutshell, it is generally observed that there are two main phases of CI; under the two phases, several tasks need to be performed to produce a successful CI performance. Liu (2008, p. 19) believed that there are five tasks in CI, among which three tasks are performed at the same time, and the other two are also performed simultaneously, as shown in Table 2.1. Table 2.1 Five tasks of CI proposed by Liu (2008) Phase. Task. Description. 1: Listening. 1. Listening. 2. Analysis and comprehension. 3. Short-term memory operations and/or note-taking. 2: Interpreting. 4. Remembering and/or note-reading. 5. Production 8.

(18) In particular, Liu pointed out that although the tasks appear to be performed simultaneously in phase one, the focus of attention actually moves at a very fast pace between the tasks. That is, in the listening phase, the focus of attention of interpreters is always moving very quickly from listening, understanding the message, and taking notes. In the Effort Model of CI proposed by Gile (2009), the process of CI is categorized into two main phases as well, as shown in Table 2.2. Table 2.2 The Efforts in Gile’s Effort Model of CI (2009) Phase. Effort. Description. 1: Comprehension. 1. Listening and analysis. 2. Note-taking. 3. Short-term memory operations. 4. Coordination. 5. Remembering. 6. Note-reading. 7. Production. 8. Coordination. 2: Reformulation. Gile used “Effort” to refer to each of the operations during interpretation, as can be seen in Table 2.2. In every phase of any mode of interpretation, there are several Efforts that need to be performed in order for a successful interpretation. For instance, in Table 2.2, Efforts 1, 2, 3, and 4 all need to be performed to complete Phase one.. 9.

(19) In addition to the tasks Liu proposes, Gile’s model includes a Coordination Effort to both the first and second phase. The function of the Coordination Effort is to allocate the total mental energy of an interpreter to the other Efforts in the same phase. Both Liu and Gile agreed that CI is a form of multitasking in both phases, maintaining that attention shifts between different tasks in the same phase. Gile (1995) even believed that sometimes overlapping happens. For instance, in the second phase, according to Gile’s argument, the Efforts might work like this: Listening and analysis Effort. Segment A Segment B Segment A Segment B. Short-term memory operations Effort. Segment A Segment B. Note-taking Effort. This shows that while the Note-taking Effort is still working on Segment A, the Listening and analysis Effort is already dealing with Segment B. In sum, CI is a two-phase process which takes up interpreters’ mental energy. On the surface, interpreters listen, take notes, and later deliver the interpretation. However, in the cognitive aspect, interpreters have to perform several tasks within the same time frame. Due to the nature of CI containing more than one tasks in both phases, interpreters have to attentively allocate their mental energy to each task in order for a successful interpreting performance. 2.1.2 Processing capacity saturation and problem triggers A successful management of mental energy during interpreting is, however, not always the case. When the demand of mental energy for interpreting surpasses what 10.

(20) interpreters possess, it will be difficult for interpreters to keep up. Gile (1995) referred to the mental energy or attention needed to cover the requirements of interpreting as processing capacity. According to Gile, each Effort in the first phase of CI has its capacity requirements on interpreters. To achieve CI successes, there are several conditions in terms of processing capacity that need to be fulfilled. First, the total processing capacity owned by interpreters needs to be equal or greater than the total requirements, which is the add-up of all the capacity requirements of each Effort. Should this condition not be met, processing capacity saturation may happen, impacting interpreting performance. Moreover, the processing capacity ready for use to meet the requirements of each of the four Efforts needs to be equal or greater than the capacity requirements of each Effort. For instance, the capacity at interpreters’ disposal to perform the Note-taking Effort needs to be equal to or more than the capacity requirement of that Effort. Even if there is sufficient total processing capacity available, a lack of processing capacity for any one of each Effort may cause interpreting failures. According to the above mentioned statements, it can be seen that failures to meet the processing capacity requirements may lead to quality deterioration in interpreting performance. This entails that interpreting failures may not result only in “insufficient linguistic or extralinguistic knowledge or mistakes,” but also in “cognitive tension between processing capacity supply and demand” (Gile, 2009, p. 182). In fact, cognitive saturation problems have been listed as a common reason for interpreting difficulty by researchers. Gile (2009) and Setton and Dawrant (2016b) believed that interpreting obstacles, if not brought by insufficient cognitive skills or knowledge needed to perform 11.

(21) interpreting, mainly result from three areas, including substandard working conditions and environment, interpreters’ personal conditions, and certain input features. These three factors can all lead to processing capacity problems (Gile, 2009). First, environmental factors such as noise and poor ventilation impact interpreters’ cognitive efficiency. Lack of background information and not being able to see the speaker may also detract interpreters from working at their best. Moreover, subjective factors relating to interpreters’ own conditions, such as high levels of stress and lack of time to prepare, may lead to a disorganized management of processing capacity. Also, particular factors in the source speech that are hard to deal with may increase processing capacity requirements. What the researcher aims to investigate in this present study is closely related to the third area, difficult factors in the source speech. For example, an information-dense speech may significantly increase the requirement of Listening and Analysis Effort. The density of speech may come from its fast delivery rate without enough pauses between sentences, or the density of information it contains. Moreover, external features of the speech, such as unfamiliar grammatical structures, accents, and linguistic style, can all lead to an increase in the processing capacity requirements in the Listening and Analysis Effort as well. On the other hand, certain input factors can also increase the requirements of the Short-term Memory Operations Effort. For instance, long names unheard of by interpreters may put burden on the Short-term Memory Operations Effort. This is especially so when the syntactical structure between the source and target speech is very 12.

(22) different, such as that of Chinese and English, as interpreters will have to store a chunk of information due to syntactical differences before rendering the interpretation into the other language. In addition, speeches that are hard to anticipate due to the speakers’ feature of speech may also increase the requirements of the Short-term Memory Operations Effort, because interpreters have to store more information before clearly knowing what the speaker intends to say. Last, the above-mentioned information density can also increase the requirements of the Production Effort. In the case of CI, the Production Effort refers to note production in the first phase. When the speech is fast, interpreters will have to take notes more quickly; when the speech has a high information density, interpreters will have to write down more information within a period of time. In all, there are a large number of factors, besides interpreters’ knowledge of the subject matter and linguistic skills, that may lead to a deterioration in interpreting performance. These problem triggers often contribute to increased processing capacity requirements on one or more Efforts. Once individual or total capacity requirements surpass the processing capacity available, quality interpretation can no longer be guaranteed. 2.2 Fast Speech Rate (FSR) As mentioned in the preceding section, particular factors in the source speech may prove challenges for interpreters. Speech rate (SR) can be one of them. SR means “the speed of message delivery” (Pöchhacker, 2016, p. 124). Among the contributing factors to elevated processing capacity requirements, fast speech rate (FSR) is viewed as a 13.

(23) leading interpreting challenge. Pöchhacker states that FSR is itself one of the influential factors to interpreting performance. Gile (2009, p. 192) even suggests that FSR might be “the most frequent source of interpreting problems,” echoed by a 2009 AIIC report in which FSR was considered by respondents to be the biggest source of stress (Neff, 2011). 2.2.1 The definition of FSR First of all, to report SR, it is best to count the “units of speech output per unit of time” produced by a speaker (Carroll, 1966, p. 2). There are two common measuring units: words per minute (wpm) and syllables per second. Researchers have proposed that using wpm to measure SR is not perfect, because two different words can vary in length greatly even in the same language, let alone in different languages. In addition, in different languages, the definition of a word may be different (Setton & Dawrant, 2016b). However, measuring wpm has been the conventional way of SR measurement (Carroll, 1966), perhaps because words are easier to count than syllables (Setton & Dawrant, 2016b), and syllable structures may also vary to a great extent in different languages (Roach, 1991). Plenty of research on the impact of FSR on interpreting performance used wpm to measure SR. This serves as the reference for this thesis, in which wpm was chosen to be the measuring unit for counting SR. FSR has its impact on listening comprehension and interpreting. In Rodero’s (2012) research on radio news presentation, she pointed out that when a speech is too fast, it becomes hard for listeners to understand the message. Rodero stated that a speaker should use a slower SR to enable listeners to better understand the speech by giving 14.

(24) them more time to process a message. There is no consensus on what rate is considered optimal for all types of speech settings. For instance, a serious matter may be better delivered with a slower SR, and for a light-hearted speech, a faster SR may be more suitable (Osborn, Osborn, & Osborn, 2009). Moreover, good public speakers can vary greatly in the pace of their speech (Zarefsky, 2013). That being said, views have been expressed on the typical SR for public speaking. For example, the average SR, according to Osborn, Osborn, and Osborn, is 125 wpm; it is 120-150 wpm according to Zarefsky (2013), and 140-150 wpm according to Nikitina (2011). Overall, a SR under 150 wpm is generally considered suitable for listening comprehension in a public speaking setting. For a speech setting where interpreting service is provided, the speaker’s delivery needs to be slower than how an average person normally speaks, because in every segment the interpreters would have to process more information than a person who is only listening (Vančura, 2013). This is due to the multitasking nature of the interpreting task, in which there are processing capacity requirements on the interpreters (see 2.1.2). To maintain good interpreting quality, therefore, the ideal SR for SI is claimed to be around 100-120 wpm2 if the speaker is not speaking from a script (Gerver, 2002 [1969]; Li, 2010; Pöchhacker, 2004). If the SR exceeds 120 wpm, interpreters will start to fall behind the speaker, and there will be more frequent errors (Gerver, 2002 [1969]). When the SR exceeds 140 wpm3, it is considered fast for interpreting (Communicate,. 2. For illustrative purposes, one may listen to “The most important lesson from 83,000 brain scans” by Daniel Amen, which was delivered at 120 wpm on average. (https://www.youtube.com/watch?v=esPRsT-lmw8) 3 Another illustrative example is “Why 30 is not the new 20” by Meg Jay, which was delivered at 148 15.

(25) 1999, as cited in Li, 2010). This roughly corresponds to the Speech Difficulty Index proposed by Setton and Dawrant (2016b), which shows that for interpreting, 100-120 wpm is considered easy; 120-140 wpm moderate; 140-160 wpm challenging; and >160 wpm4 difficult. The word challenging is referred to a situation where “professional conference interpreters can keep up, give a (mostly) complete if compressed version but may find the pace uncomfortably fast and tiring” (Setton & Dawrant, 2016b, p. 52); and the word difficult means “even skilled professionals find it difficult to keep up, […]” (Setton and Dawrant, 2016b, p. 52). In a nutshell, it is asserted that if the SR exceeds a certain limit, interpreting quality will be affected. A number of studies have been carried out on the effect of SR difference on interpreting performance in interpreting, mainly SI, from the early study conducted by Gerver (2002 [1969]) to more recent studies (Barghout, Rosendo & García, 2015; Chang, 2009; Korpal, 2012; Pio, 2003; Wu et al., 2010). In these studies, speeches with different SRs have been chosen as FSR and slow SR (SSR) speech materials in SI experiments to observe the impact of SR on interpreting output. In a number of these studies, the SRs were measured by wpm. The SSRs were often set at around 100-130 wpm, which is “easy” or “moderate” for interpreting according to Setton and Dawrant (2016b); the FSRs were around 145-200, which fall into the category of “challenging” or “difficult”. For instance, in Gerver (2002 [1969]), the fastest SR was set at 164 wpm, and the slowest SR was set at 95 wpm. In Pio (2003), the SSR was set at 108 wpm; FSR. wpm on average. (https://www.youtube.com/watch?v=vhhgI4tSMwc) 4 Yet another example is “Your body language may shape who you are” by Amy Cuddy, which was delivered at 177 wpm on average. (https://www.youtube.com/watch?v=Ks-_Mh1QhMc) 16.

(26) set at 145 wpm. In Korpal (2012), the SSR was set at 130/130 wpm; FSR at 177/180 for the two groups in his experiment. In Barghout, Rosendo, and García (2015), the SR was designed to change between 120, 160 and 200 wpm in the speeches in their experiment. In all, research has suggested a range of comfortable and challenging SR for interpreting. Though the numbers of FSR and SSR vary from study to study, it is clear that there is a limit over which interpreting will become difficult. 2.2.2 FSR’s Impact on Interpreting The studies about the effect of SR difference on interpreting performance in interpreting have yielded a consistent pattern of results regarding how SR variance affected interpreter’s performance regarding the number of omission, errors, and ear-voice span (EVS), which is the lag time that ranges from the interpreter’s perception of a message in the source text and the beginning of the target language presentation (Chang, 2009). First, as SR increases, interpreters tend to omit more information (Barghout, Rosendo & García, 2015; Chang, 2009; Gerver, 2002 [1969]; Korpal, 2012; Pio, 2003; Ribas, 2012; Wu et al., 2010). While sometimes an omission is a conscious decision, which makes it a coping strategy (Ribas, 2012), omissions may also happen because of other factors in the interpreting process (Barghout, Rosendo & García, 2015). Also, as SR increases, interpreters tend to make more errors in their TL output (Chang, 2009; Gerver, 2002 [1969]; Pio, 2003; Wu et al., 2010), and lag further behind the speaker (Chang, 2009; Gerver, 2002 [1969]; Pio, 2003).. 17.

(27) In the aforementioned studies, it was implied that an interpreter makes more omissions or errors as SR picks up because the processing capacity requirement for interpreting is higher than the processing capacity available when SR reaches a certain limit (Barghout, Rosendo & García, 2015; Chang, 2009; Gerver, 2002 [1969]; Korpal, 2012; Pio, 2003; Wu et al., 2010), which corresponds to the processing capacity saturation described by Gile (see 2.1.2). Chang (2009) suggested that processing capacity overload caused by FSR may happen in any Efforts in the SI process, including the Listening and Analysis Effort, the Memory Effort, and the Production Effort. In terms of the impact of FSR on the Listening and Analysis Effort, Barghout, Rosendo and García (2015) and Wu et al. (2010) suggested that sometimes the interpreter does not even perceive a message in the speech, because the interpreter cannot keep up with the speed the incoming messages are presented. If fewer messages are perceived, fewer messages can be further processed to be understood (Wu et al., 2010). Even if a message is perceived, the interpreter may not be available to process the message as the SR is too fast, making it difficult to have the message understood (Wu et al., 2010). For instance, Pio (2003) speculated that the interpreter lacks enough time to properly analyze, segment, and connect the messages in an FSR occasion, which leads to more errors related to time sequences. Sometimes, not being able to either perceive or analyze a message may result from the interpreter still processing the previous segment because the SR is too fast (Pio, 2003). In terms of the impact of FSR on the Memory Effort, a message needs to be first understood in order to be stored in short-term memory (Wu et al., 2010). Interpreters 18.

(28) may encounter difficulties in storing messages in memory in the face of FSR (Barghout, Rosendo and García, 2015), due to more information requiring processing capacity in the Memory Effort from the interpreter (Wu et al., 2010). Due to lack of enough processing capacity, messages in short-term memory “accumulate and deteriorate faster than the interpreter can cope” (Gerver, 2002 [1969], p. 64). In terms of the impact of FSR on the Production Effort, Barghout, Rosendo and García (2015) believed that FSR negatively affects the interpreter’s ability to reformulate messages in SI. On the contrary, in Ribas’s 2012 study, student interpreters doing CI reported finding FSR not a major problem to production; it was instead reported to be a major problem to listening and note-taking. This demonstrates the different nature between SI and CI. In CI, production takes place in the second phase (see 2.1.1). In other words, CI does not require the interpreter to produce the TL output while new messages keep coming at the same time as in SI. Therefore, this may be the reason why the participants in Ribas observed that their TL output was not pressured by the SR of the speaker. Nevertheless, in both CI and SI, message losses in perception, comprehension, and memory due to FSR may lead to an incomplete TL output in the end (Wu et al., 2010). 2.2.3 Strategies to cope with FSR in interpreting It is recognized by researchers that FSR is a source of difficulty in both SI and CI. Researchers have identified several strategies to deal with this difficulty. First of all, Setton and Dawrant (2016b) suggested interpreters approach the speaker directly and ask them to maintain a moderate SR beforehand, or to slow down during the talk. 19.

(29) However, Setton and Dawrant acknowledged that this method often a has limited effect, since the speaker may speak fast again after only a short while. Therefore, the interpreter may need to resort to other strategies when this is the case. For instance, Setton and Dawrant pointed out that the interpreter may react to FSR by speeding up themselves when they deliver the output. Nevertheless, this strategy will also be difficult to implement if the SR becomes too fast, since it is not possible to increase the output SR indefinitely (Setton & Dawrant, 2016b). When SR becomes too fast, the interpreter would be left with little choice but to use other strategies to try to get as much information as they could. For instance, Gile (2009) stated that the conference interpreting community generally agree that meaning-based interpreting can provide better quality, since it allows the interpreter to better understand the speaker’s intentions and produce the TL output without having too much linguistic interference from the SL. However, when processing meaning-based segments becomes difficult due to FSR, in order to still be able to render an acceptable amount of information, the interpreter may need to resort to a form-based strategy (Gile, 2009; Massaro & Shlesinger, 1997). This means that the interpreter will be following “the surface form of the source text” when they produce the TL output, instead of detaching themselves from the source text and translating based on the meaning of the source text (Dam, 2001, p. 27). Interpreters would need to take the risk of being less clear and idiomatic, however, if they do form-based interpreting (Gile, 2009).. 20.

(30) When it is difficult to process fuller units due to FSR, another strategy interpreters could use is to try to shorten their EVS in SI, which means to shorten the lag behind the speaker (Chang, 2009). The purpose of this strategy is to alleviate the pressure on the Short-term Memory Effort. This way, interpreters will not have to store too many messages in short-term memory, allowing more mental energy for the other Efforts. However, it may also be risky to shorten EVS (Chang, 2009), as maintaining a certain length of EVS is a way to guarantee comprehension of the source speech, because interpreters process the source speech by units larger than single words. Foulke and Sticht (1967) believed that while single words themselves may be intelligible when delivered at a fast pace, the message the words imply may not be comprehensible, since comprehension requires a more complicated processing from the listener, which may be inferred by FSR. Gerver (2002) believed that since interpreting requires the comprehension and analysis of messages instead of perceiving isolated words, a certain length of EVS needs to be maintained to enable the interpreter to process larger grammatical units, which was echoed by Liu (2008). In Barghout, Rosendo and García (2015), it was mentioned that sometimes the interpreters missed information in their TL output when the speech was fast because they followed too closely to the speaker’s sentence structure and clauses. It was possible that the interpreters did not have enough processing capacity to lag further behind the speaker, listen for a while, and fully understand the messages before producing the target, because the new messages keep on coming (Chang, 2009).. 21.

(31) Similar to EVS in SI, ear-pen span (EPS) in CI, which is the lag time that ranges from the interpreter’s perception of a message in the source text and the moment a note is written down to represent that message (Chen, 2017), has also been investigated. Though no prior studies have studied FSR and EPS in CI, a few studies did discover the correlation between longer EPS and messages that take more time to be understood (Andres, as cited in Setton & Dawrant, 2016b; Setton & Dawrant, 2016a), and the correlation between longer EPS and higher cognitive load (Chen, 2017). When the speaker goes beyond a certain SR, in order to keep the essentials, omission is a possible strategy to adopt. While omission has often been viewed as a type of translation error due to cognitive overload (Chiu, 2017), studies have shown that omission can sometimes be a deliberate strategic move. For example, Shlesinger’s study on professional interpreters (2003) revealed that interpreters tend to consciously make the decision of omitting some of the modifiers. The results in Barghout, Rosendo, & Garcia (2015, p. 328) also suggested that omission might be a “justifiable strategy.” In fact, Barghout, Rosendo, & Garcia argued that based on the main purpose of interpreting, which is to enable comprehension by the listeners, it is reasonable for interpreters to resort to omission as a strategy in FSR occasions, as long as it does not affect the listeners’ comprehension. Another way to safeguard the essential points in the source speech is through summarizing. According to Setton and Dawrant (2016b), professional interpreters are equipped with the ability to use fewer words to summarize the message in the source speech. This skill can be a useful strategy when FSR makes it difficult to convey every 22.

(32) message. However, Setton and Dawrant also pointed out that summarizing due to FSR can be very tiring to the interpreting, and often cannot be used over a number of speeches in a row. If all strategies fail due to an impossibly fast SR, interpreters may need to stop interpreting to avoid the risk of distorting the messages (Setton & Dawrant, 2016b). Although switching off the microphone in this case is probably better than providing very poor interpreting service, Gile (2009) believed that nowadays clients seldom accept this extreme strategy. Because of this, Gile pointed out that usually interpreters just keep trying their best after informing the listeners that the situation prevents them from delivering good interpreting. 2.3 CI Note-taking CI notes, according to Liu (2008), can be written in any language or form, serving as visual cues to help the interpreter fully convey the messages. Note-taking is an essential part in CI, as notes serve as memory aids for interpreters to render the messages in the original speech, which can last longer than what an average person could remember (Chuang, 2008; Gillies, 2005; Lu, 2013; Mahmoodzadeh, 1992; Pöchhacker, 2016). 2.3.1 Prescriptive studies on note-taking skills No exact rules determine what good notes are, since every interpreter may have their own preferences in note-taking to help them render the messages faithfully in the target speech (Gillies, 2005; Kohn & Albl-Mikasa, 2002; Lu, 2013). Besides, it often 23.

(33) takes more than things written down to deliver a good CI performance. For instance, the interpreter needs to rely on information not only written on the notes but also stored in their short-term memory to interpret well (Chen, 2017). Also, under some circumstances, writing notes may not necessarily be more helpful than just listening (Liu, 2008). Nevertheless, researchers have elaborated on some basic elements of the interpreter’s notes in prescriptive studies to provide solutions to common problems. The strategies that have been proposed can be mainly categorized into five categories: concepts instead of words, summarizing, spacing arrangements, links, and symbols and abbreviations. First and foremost, the interpreter should aim at writing down concepts from the speech instead of transcribing the exact words (Gillies, 2005; Liu, 2008; Rozan, 2003 [1956]). “Concept” in this sense means “the underlying meaning of a word used” in the source speech (Gillies, 2005, p. 53). It was advised that the interpreter try to write shorter synonyms to not be trapped in the exact word the speaker says (Gillies, 2005). If the interpreter is able to do this, it shows that they have analyzed the message before writing notes (Liu, 2008). In fact, in many cases, very few notes are needed to represent a complicated concept if the interpreter had understood the ideas. Too many notes may even be a hindrance to decoding in some situations (Liu, 2008). In addition, sometimes the interpreter’s background knowledge may enable them to take only a few notes, because the notes are sufficient to remind them of things they already know (Gillies, 2005; Kohn & Albl-Mikasa, 2002). If the interpreter is able to grasp the concept in the source speech, it should not matter which language they write in, and how many notes 24.

(34) they take, as long as the notes are effective to the interpreter (Liu, 2008). Nevertheless, there are some things that have been suggested to write down upon hearing, including numbers and proper names. Contrary to the method of maintaining an amount of time span between the speech and note-taking to allow comprehension, it has been suggested that numbers and proper names be written down as soon as the interpreter hears them (Gillies, 2005; Liu, 2008). The reason is that a number or a proper name can contain only one meaning (Liu, 2008), and is difficult to work out based on the context if not written down (Gillies, 2005). In regards to numbers, Liu (2008) emphasized the importance of understanding the content of the speech around numbers, because being able to render what numbers mean is often more important than getting the exact numbers right. While writing down concepts is important in note-taking, it may not be wise for the interpreter to write down every detail the speaker has said. The ability to transfer the source text into organized and simplified notes has been considered important. In Ribas (2012) it was found that the main strategy advanced students resorted to in note-taking was summarizing. Albl-Mikasa (2008) also found that interpreters reduce their notes mainly using two strategies, one being selecting what should be noted, and noting the most important content words. Thus, identifying keywords is an important skill (Hsiao & Chang, 2014), and the interpreter should decide which messages should be noted (Liu, 2008; Tang, 2010). After all, it is often challenging, if not impossible, to note every exact detail. Einstein, Morris, and Susan Smith (1985) found that students who performed better wrote down more important propositions, which reveals that noting the 25.

(35) right ideas may lead to better interpreting performance. A good way to showcase different levels of importance in the speech on notes is making spacing arrangements (Gillies, 2005). According to Liu (2008), effective spacing arrangement is the most important note-taking principle. Notes are supposed to be a visual cue to the interpreter, reflecting the logic of the speech and the relationship between messages with the space arranged by the interpreter (Gillies, 2005; Liu, 2008). A simple way would be, for example, in Chuang (2008, p. 99), to “broke down sentences into subjects and predicates” in notes. There are also several ways to arrange spacing that has been proposed. Rozan (2003 [1956]) proposed applying verticalization to note-taking, which was echoed by Liu (2008), believing that it enables the interpreter to see several lines of notes at a glance, and showcases the relationship between messages better. Secondly, indentation of notes can further emphasize the relationship between messages, while giving a clear distinction of different meaning units. Notes which belong to the same meaning unit can also be indented to the same level to show that they have the same level of value (Liu, 2008). Thirdly, drawing a horizontal line where an idea ends can visually remind the interpreter where a short segment is finished (Liu, 2008). Lastly, Gillies (2005) and Rozan (2003 [1956]) both suggested using brackets to show that a few words belong together in a clause that is not the most important part of an idea. Apart from ideas, links have also been recommended to be written down (Gillies, 2005; Liu, 2008; Rozan, 2003 [1956]). Gillies (2005) believed that the ideas in the speech and the links connecting the ideas are what a speech is made up of. The 26.

(36) functions of links include, first, to highlight the logic of the speech, and secondly, to add color to a speech (Liu, 2008). Gillies (2005) and Liu (2008) both recommended using symbols or abbreviations that are familiar to the interpreter to put down links to save time, and writing links in the left margin on the paper to emphasize them. Symbols and abbreviations can be useful in jotting down not only links but also ideas. Both symbols and abbreviations are tools to represent something in a more succinct manner. Using symbols and abbreviation strategically can be effort-saving (Hsiao & Chang, 2014; Liu, 2008; Tang, 2010). However, symbols should not be viewed as the most important part of note-taking skills. In fact, different interpreters vary greatly in the number of symbols they use (Liu, 2008). In addition, Rozan (1956, cited in Liu, 2008) believed that 20 symbols are enough for note-taking, among which only 10 are indispensable. The key is to focus on forming a sound note-taking system in the first place (Gillies, 2005), and perhaps develop a set of effective symbols that have a consistent meaning to the interpreter (Gillies, 2005; Liu, 2008). It has been advised that the interpreter do not improvise symbols (Gillies, 2005). For instance, Chuang’s (2008) experiment showed that inventing new symbols on the spot may lead to difficulty in the decoding stage. In sum, familiar and consistent symbols and abbreviations can be helpful tools if used efficiently to save time. Based on the note-taking strategies listed above, this thesis wishes to analyze the strategies subjects used in the experiment, and to investigate whether FSR impels interpreters to come up with particular strategies in note-taking.. 27.

(37) 2.3.2 Empirical studies on note-taking preferences While prescriptive studies offered fruitful suggestions for note-taking skills, researchers were inquisitive about how interpreters’ notes actually look in authentic interpreting situations. Thus, researchers have conducted empirical studies on interpreters’ note-taking preferences, attempting to generalize trends. Note-taking preferences are the choice of form and language the interpreter resort to in note-taking (Chen, 2017). The choice of form most commonly includes language versus symbol, and abbreviations versus full words, while the choice of language refers to source versus target language or A versus B language. Furthermore, researchers were curious whether certain note-taking preferences may lead to better interpreting performance, on which empirical studies have also been conducted. Although a large volume of data has been accumulated on this subject matter including Andres (2002), Cardoen (2013), Chen (2017) Dai & Xu (2017), Dam (2004, 2007), Her (2001), Liu (2010), Szabo, (2006), and Wang et al. (2010), just to name a few, the results of different studies have been diverse. The different results may have resulted from the different design of the studies, such as inclusion criteria of participants, interpreting direction, and language combination of participants (Marani & Tabrizi, 2018). The following is a review on the studies. In regard to the choice between language and symbol, past studies seemed to reveal a proclivity for language. In Chen (2017), Dai and Xu (2007), and Wang et al. (2010), the interpreters preferred writing more words and less symbols. This result contradicted the result in the early study conducted by Dam (2004), in which the interpreters wrote more 28.

(38) symbols than words and abbreviations. The different results did not seem to result from the criteria of the participants, considering the fact that Chen and Dam both used professional interpreters, while Dai and Xu and Wang et al. used student interpreters. In other words, we may assume that both professional and student interpreters seem to have this preference according to the studies. While language seemed to be used more than symbols, studies have yielded mixed results on the choice between full words and abbreviations. On the one hand, some studies (Dai and Xu, 2007; Dam, 2004; Liu, 2010) demonstrated a dominance of full words over abbreviations, but on the other hand, others (Chen, 2017; Wang et al., 2010) suggested the opposite. In this case, the mixed results did not seem to be due to inclusion criteria differences either. The third note-taking preference, the choice of language, is perhaps the most complex issue among the three, as it can be investigated from the angle of interpreting direction or the interpreters’ A and B language. From the point of view of A and B language, no consistency has been sought. While some studies (Andres, 2002; Chen, 2017; Dam, 2007; Li, 2012; Wang et al., 2010) revealed a preference for more B language in notes, an almost equal amount of studies (Dai & Xu, 2007; Dam, 2004; Liu, 2010) reported a preference for A language over B language. On the other hand, among these studies, a majority of them yielded the result of a preference for source language over target language. One exception is the early explorative study Dam (2004). In Chen (2017), although the interpreters wrote more in target language when the target text is in their B language, they wrote more in source language when the source text is their B language, 29.

(39) and the tendency was stronger in the latter experiment. The results of the studies, once again, were not clearly associated with whether the researchers used professional or student interpreters. 2.3.3 Empirical studies on note-taking preferences and interpreting performance From the studies about note-taking preferences reviewed above, there is no denying that note-taking preferences is a highly complex and subjective issue. Since note-taking preferences have appeared to be quite diverse, researchers were interested in finding out the note-taking preferences that could lead to a better interpreting performance. In this area, studies can be mainly categorized into the investigation into note quantity, note form, and language. Among the studies about the relation between note quantity and interpreting performance, some (Chen, 2017; Dam, 2007; Li, 2012) have suggested more notes can lead to better performance. In particular, Chen proposed that more notes of good quality can possibly help the interpreter enhance their performance. Similarly, Her (2001) also pointed out the relation between note quality and interpreting performance. Nevertheless, there are different results from other studies. In Dai and Xu (2007) and Wang et al. (2010), no correlation was found between note quantity and interpreting performance. In addition, Cardoen (2013) even found that the interpreters took fewer notes in the parts where they performed more fluently. Among the three studies revealing the correlation, two used professional interpreters as subjects; while all the three studies that did not show the correlation used student interpreters. It is possible that since student interpreters are not yet skilled in note-taking, more notes do not necessarily lead to better performance. 30.

(40) In terms of the form of notes, while prescriptive studies have recommended the use of more symbols and abbreviations, no consistency has been reached in empirical studies. While Dam (2007) discovered the relation between more abbreviations and better interpreting performance, Cardoen (2013) found the relation between more full words and better interpreting performance. Apart from that, Liu (2010) found out that more symbols seemed to lead to better interpreting performance. Liu (2010) also discovered that using more separation marks seemed to help boost performance. A separation mark, according to Liu, can consist of a line, two lines, an end mark, or any other clear marks, serving the purpose of efficiently reminding the interpreter of the start and end of a segment of target text. Separation marks can be a note-taking preference on its own, while few research has addressed this particular variable. Lastly, studies (Dam, 2007; Wang et al., 2010) have yielded no correlation between language and interpreting performance. It seems that the choice of language is indeed highly subjective. Good performance does not seem to be associated with writing more in A or B language, or source or target language. In sum, while rich literature has been developed on note-taking preferences and their relation with interpreting performance, inconsistencies among studies still exist in quite a few areas. Moreover, very few of the studies have combined the quantitative data on note-taking preferences with qualitative data on note-taking decision-making process. As a result, this study attempts to explore note-taking preferences and interpreting performance using both quantitative and qualitative analyses in order to probe into the reason behind a particular choice. Also, based on Liu (2010)’s result on separation marks, 31.

(41) this study wishes to test the relation between the use of segmentation in notes and interpreting performance. 2.4 Interpreting Performance In this thesis, the interpreting performance of interpreters will be assessed to shed light on the relations between notes and performance in an FSR vs. SSR setting. To assess performance of interpreters, studies on the quality of interpreting first needs to be reviewed. Quality has long been a focus of research in the field of interpreting. While researchers in the past have investigated this issue from different angles, there is currently still no consensus on one set of objective criteria of good quality in interpreting (Ding, 2017; Pöchhacker, 2016). However, fidelity (or sense consistency with the source text) has generally been considered the most important criteria in judging interpreting quality (Buhler, 1986; Kurz, 1993; Pöchhacker, 2016; Strong & Rudser, 1985). Harris (1990, p. 118) maintained that interpreters should “re-express the original speakers' ideas and the manner of expressing them as accurately as possible and without significant omissions;” while Hale (1997, p. 202) argued that interpreters should “not only convey the ‘general idea’ of the message, but also the speaker’s intention.” Though there is no definite answer for what interpreting a text faithfully clearly entails (Moody, 2011), it is commonly agreed upon that the concept of fidelity is not equal to word-to-word rendition. Viaggio (1992) wrote that understanding the sense in the source text can enable one to interpret it without sticking to the words. Moreover, interpreters sometimes even have to resort to omitting some of the concepts in the source text to convey the main ideas and meaning in the face of processing difficulty 32.

(42) due to different factors, such as fast delivery of the speaker (Barghout, Rosendo, & García, 2015). From the studies reviewed above, it is apparent that maintaining sense consistency without staying too close to the source text has been suggested as a way to guarantee quality. Past research has accumulated sets of criteria to measure interpreting quality. Lee (2008) pointed out that research on interpreting quality assessment is scarce, possibly because performance assessment is still rather subjective and beyond simply checking correct answers. Yeh and Liu (2006) also maintained that evaluators may be subjective when there is not enough time to examine the test takers’ performance or when the criteria are not clear enough. Other researchers have also touched upon the difficult nature of conducting an objective assessment (Choi, 2006; Kalina, 2005). Nevertheless, performance assessment is extremely important because it helps interpreters or interpreting students understand places where improvements are needed (Choi, 2006; Schjoldager, 1996). While researchers have generally agreed upon the need of assessing interpreters’ performance, there have been different sets of criteria listed by various researchers. For instance, Schjoldager (1996) listed four criteria for evaluating performance in simultaneous interpreting used in her classroom setting, namely, comprehensibility and delivery, language, coherence and plausibility, and loyalty. In Robert’s (2000) guidelines on evaluating professional community interpreting performance, it was suggested that interpreters be scored based on language skills and interpreting skills. Pöchhacker (2016) also proposed a model for quality assessment, including accuracy, 33.

(43) adequate target-language expression, equivalence, and successful communicative interaction as elements. Lee (2008) adjusted Pöchhacker’s model and selected accuracy, target language quality, and delivery as the criteria on which students’ performance assessment should be based. Yeh and Liu (2006) proposed scoring rubrics for assessment, selecting “fidelity” and “intelligibility” as criteria, each is scored in bands from 1 to 5. Similarly, the CI exam held by the Ministry of Education in Taiwan assesses test-takers’ performance based on accuracy and delivery, each is scored in bands from 1 to 5 as well. Overall, researchers have proposed different sets of criteria for quality assessment, each of which contains more than one assessment criteria to guarantee a more holistic judgment. Based on the literature reviewed above, delivery, target language skills and accuracy are the three main criteria that have been commonly listed. Among the three main criteria to assess interpreters’ performance, Liu and Chiu (2009, p. 251) specified that “accuracy [alone] is a more clear-cut criterion for determining the effect of input difficulty on the interpreting performance.” 2.5 Research Questions FSR is recognized in the existing literature as one of the challenging factors in the source text. To successfully maintain interpreting quality in the event of FSR, student interpreters will have to develop a set of coping strategies. From the literature reviewed, the researcher discovered that there is abundant research dedicated to note-taking skills, but few has provided suggestion particularly for coping with FSR. Based on the gap, the research questions are as follows. 34.

(44) 1. How does FSR affect student interpreters’ note-taking preferences in CI? 2. How does FSR affect student interpreters’ note-taking strategies in CI? 3. How does FSR affect student interpreters’ accuracy in CI? 4. Do certain note-taking preferences correlate with better CI accuracy in FSR occasions? The first two questions were proposed with the aim of identifying note-taking preferences for FSR occasions and the motives behind the decision. The third and forth questions were designed to observe whether certain note-taking strategies could lead to better CI accuracy despite the challenges brought about by FSR. It is hoped that by answering the four questions, more insights will be generated on how FSR may impact student interpreters’ note-taking and interpreting quality.. 35.

(45) Chapter 3 Research Method To answer the research questions proposed, a within-subjects experimental design was adopted in this study, in which all the subjects performed consecutive interpreting (CI) and took notes in the practice and formal sessions, and took part in a semi-structured interview. As reviewed in Chapter 2, there are five phases of CI, two of which involve interpreters’ notes. Interpreters’ notes reflect how information is processed and interpreted by the interpreters (Chen, 2017; Garretson, 1981). Therefore, investigation into CI notes is an effective way to observe the decision-making process interpreters tend to engage in. Introspective and retrospective interviews are often used in research for complementing quantitative results. An introspective research method requires participants to say what they are thinking during the process of completing a task. After the introspective verbal output was collected, the researcher may conduct a retrospective interview for the purpose of triangulation (Charters, 2003). In an interpreting experiment, however, it is impossible for subjects who are interpreting to talk about their thought and decision process while they are interpreting. Therefore, retrospective interview alone was often chosen as the qualitative tool in this kind of circumstance, which is also what the researcher chose to adopt in the study. Conducting retrospective interviews is an effective way to elicit perspective from the subjects, to complement the quantitative data, and to discover an explanation for the process and results of the task the subjects have performed (Chen, 2017; Kuo, 2012). 36.

(46) However, retrospective interview does have its drawbacks. Vik-Tuovinen (2002) believed that the data generated from retrospective interviews may be incomplete because interpreters may not be aware of or remember their own decisions during the task, and they may not share everything they are conscious of, or rationalize their performance. Yet, Vik-Tuovinen also supported the view that eliciting participants’ preferences are suitable for studies about interpreting strategies. Since one of the purposes of this study is to find out the subjects’ note-taking strategies, the adoption of retrospective interview to understand the process of decision making is quite appropriate for this study. In this study, the retrospective verbal reports were used to reveal the subjects’ own perspective on their notes, helped the researcher understand what happened during unusual interpreting presentation, and assisted the researcher to decipher the subjects’ notation. During the interviews, the researcher also asked the subjects standardized questions to understand their view concerning the research questions proposed. In terms of retrospection procedures, Ivanova (2000) experimented on the two retrieval cues to prompt the subjects’ memory discussed by Ericsson and Simon, and concluded that the more effective one was “the presentation of the original stimuli in conjunction with the use of observational notes” (p. 33). This memory-prompting method decreases the chance of the target text transcripts or recordings interfering with what the subjects think during the interpreting process, and spares the subjects from feeling embarrassed about the parts of the interpretation that they are not satisfied with. In this study, the researcher also provided access to the subjects to the transcripts and 37.