視譯閱讀理解歷程之眼動研究

全文

(1)A Thesis Presented to the Graduate Institute of Translation and Interpretation National Taiwan Normal University 國立臺灣師範大學翻譯研究所碩士論文. Thesis Advisor: Dr. Jie-li Tsai, Dr. Tze-wei Chen 指導老師：蔡介立博士、陳子瑋博士. Tracking Eye Movements in Sight Translation – the comprehension process in interpreting 視譯閱讀理解歷程視譯閱讀理解歷程之眼動研究. Advisee: Chih-Chieh Huang 研究生：黃致潔. January, 2011 中華民國一○○年一月.

(2) Acknowledgement 這一年多來，我總是在口譯工作以及論文之間手忙腳亂著，常常覺得自己好像泰國人妖秀裡變男變女變變變的演員，一轉過頭，換了一張面具，呈現的就是另一種面貌。有好多個日子，口譯工作結束後，我穿著套裝踩著高跟鞋，奔到學校去做論文，然後覺得自己格格不入；也有好多時候，我從學校搭上計程車，勿忙趕去工作或教書，好像灰姑娘搭上南瓜馬車一樣。終於，論文完成了，不用再換面具，即將完成師大翻譯所將近五年的學業，心中滿滿的，五味雜陳。這本論文得以完成，最要感謝我的兩位指導教授：政大心理系蔡介立老師及師大翻譯所陳子瑋老師。蔡介立老師學養深厚、風度謙和，總是把學生的事當成自己的事，不辭辛勞、盡心盡力，是實驗室公認的守護天使。我的行程忙碌，想必帶給老師不少困擾與不便，但老師始終包容著我，讓我心中時常非常激動與感恩。陳子瑋老師聰明睿智、格局遠大，不僅在口譯工作上對我提攜照顧，更在身兼數職、百忙之中幫助我確立研究的方向與論文主題，帶領我完成這個我原本認為不可能的任務。我喜歡口譯工作的自由及南征北討的刺激，卻也時常渴望上班族與人長期相處的機會，以及隸屬於某個團體的安定感。謝謝政大眼動與閱讀實驗室所有的伙伴，讓我擁有這樣的歸屬感，對於一個自由工作者來說，這種機會實在彌足珍貴。特別感謝我的良師益友翠屏，一路以來情義相挺，這本論文沒有妳，絕對不可能完成。謝謝妙璇學姊及家興學長的耐心指導，這本論文就是這樣一次一次又一次，在實驗室裡挨在你們身邊，一點一滴學習、累積出來的。謝謝洞穴裡的熊跟雅嵐，你們是最好的談心對象，與你們度過好多快樂的日子。也謝謝實驗室加油團的所有成員，給我信心及勇氣。謝謝所有參與實驗的同學及學弟妹，更謝謝師大翻譯所這麼多年來共同學習及玩樂的好同學們。你們個個聰明認真、才華洋溢，與你們當同學是我的榮幸。我們歷經了兩年的星光大道、一年的生活在他方，以及最後的工作論文兩頭燒，一起完成翻譯所學業這項壯舉。謝謝柏成，總是犧牲自己的假日，許多週末陪伴我一起在書桌前敲鍵盤。謝謝你由衷的欣賞我，讓我做最自在的自己，給予我最多的鼓勵和力量，也謝謝你的溫柔體貼及照顧疼愛，讓我沒有後顧之憂、勇往直前。感謝我的親愛的爸爸、媽媽及家人，給我這麼好的成長環境，讓我受良好的教育，有幸認識身邊的老師、同學及朋友，也才有機會完成這本論文。爸爸的勤奮上進，一直是我積極向上的動力；媽媽的溫柔敦厚，則是我自信樂觀的原因。論文寫完了，帶著這麼多親朋好友的關心跟愛，還是要繼續努力不懈，跟著流走，帶著流。.

(3) Table of Contents Acknowledgement......................................................................................................... I List of Tables............................................................................................................... IV List of Tables in Appendix B ..................................................................................... IV List of Figures............................................................................................................... V Abstract....................................................................................................................... VI 摘要 .......................................................................................................................... VIII Chapter 1 Introduction ................................................................................................1 1.1 Research Background.....................................................................................1 1.2 Research Questions.........................................................................................5 Chapter 2 Literature Review .......................................................................................7 2.1 Definitions: types of interpreting and language processing ........................7 2.1.1 Simultaneous Interpreting (SI)...............................................................8 2.1.2 Sight Translation (ST)..........................................................................10 2.1.3 Simultaneous Interpreting with Text....................................................12 2.1.4 Skills of Sight Translation....................................................................14 2.2 The interpretation process and the comprehension phase........................15 2.2.1 The vertical and the horizontal perspectives........................................15 2.2.2 Reading ahead in sight translation .......................................................24 2.3 How eye movements reveal the process of reading comprehension .........25 2.4 The eye-tracking method in T&I Studies....................................................27 2.4.1 The eye-tracking method in interpretation studies...............................28 2.4.2 The eye-tracking method in translation studies ...................................32 Chapter 3 Experiment ................................................................................................37 3.1 Method ...........................................................................................................39 3.1.1 Participants...........................................................................................39 3.1.2 Design ..................................................................................................40 3.1.3 Materials ..............................................................................................41 3.1.4 Apparatus .............................................................................................43 3.1.5 Procedure .............................................................................................43 3.2 Data analysis..................................................................................................48 3.2.1 Eye movement indices and data analysis.............................................48 3.2.2 Reading ahead analysis ........................................................................51 II.

(4) 3.3 Results ............................................................................................................55 3.3.1 Eye movement indices .........................................................................55 3.3.1.1 First-pass reading time and probability.....................................55 3.3.1.2 Go-past time (GPT) and regression-out rate (ROR) .................58 3.3.1.3 Total viewing time (TVT), rereading time (RRT), and rereading probability .................................................................................61 3.3.2 Reading ahead analysis ........................................................................64 3.4 Discussion.......................................................................................................71 3.4.1 First-pass indices..................................................................................71 3.4.2 Go-past time and regression-out rate ...................................................72 3.4.3 Rereading time, rereading rate, total viewing time..............................73 3.4.4 Reading ahead ......................................................................................75 Chapter 4 General Discussion....................................................................................78 4.1 The vertical perspective and implications for interpretation training.....78 4.2 Reading ahead and implications for interpreter training .........................80 4.3 Future perspectives .......................................................................................82 4.4 Conclusion .....................................................................................................83 References....................................................................................................................85 Appendix A. Experiment Materials ..........................................................................89 Appendix B. Statistic Results of Eye movement Indices .........................................94. III.

(5) List of Tables Table 1.Correspondence of the three components of interpreting to the three conditions of this experiment........................................................................38 Table 2. Mean and standard deviation (in parenthesis) of first fixation duration, single fixation duration, and gaze duration on two-character words (ms)....56 Table 3. Mean and standard deviation (in parenthesis) of fixation probability and refixation probability on two-character words (%).......................................57 Table 4. Mean and standard deviation (in parenthesis) of go-past time on two-character words (ms) .............................................................................59 Table 5. Mean and standard deviation (in parenthesis) of regression-out rate on two-character words (%)...............................................................................60 Table 6. Mean and standard deviation (in parenthesis) of rereading time and total viewing time on two-character words (ms) ..................................................62 Table 7. Mean and standard deviation (in parenthesis) of rereading rate and regression-in rate on two-character words (%).............................................63 Table 8. Mapping between first-pass fixation and oral production ......................66 Table 9. Silent reading, reading aloud, sight translation and corresponding tasks75. List of Tables in Appendix B Table B 1. Statistic results of first-pass reading time............................................94 Table B 2. Statistic results of fixation probability and refixation probability.......95 Table B 3. Statistic results of go-past time and regression-out rate on two-character words......................................................................................95 Table B 4. Statistic results of rereading time and total viewing time on two-character words......................................................................................95 Table B 5. Statistic results of rereading rate and on two-character words............95. IV.

(6) List of Figures Figure 1. The vertical perspective/meaning-based strategy..................................17 Figure 2. The horizontal perspective/transcoding strategy ...................................19 Figure 3. The processes of the interpretation task ................................................30 Figure 4. Eye-tracking experiment procedure ......................................................47 Figure 5. Labeling first fixation onset/oral production transcription....................52 Figure 6. Reading ahead example.........................................................................54 Figure 7. Mean of first fixation duration, single fixation duration and gaze duration on two-character words (ms) ..........................................................56 Figure 8. Mean of fixation probability and refixation probability on two-character words (%)......................................................................................................58 Figure 9. Mean of go-past time on two-character words (ms)..............................60 Figure 10. Mean of regression-out rate on two-character words (%) ...................61 Figure 11. Mean of rereading time and total viewing time on two-character words (ms) ...............................................................................................................62 Figure 12. Mean of rereading rate on two-character word (%) ............................63 Figure 13. Reading ahead probability...................................................................65 Figure 14. Fixated sentence and produced sentence at the first fixation onset in each sentence ................................................................................................70. V.

(7) Abstract While the three components of interpreting have been identified as comprehension, reformulation, and production, the process of how these components occur has remained relatively unexplored. The present study employed the eye-tracking method to investigate the process of sight translation, a mode of interpreting in which the input is written rather than oral. The research focused especially on the comprehension component in sight translation, addressed the validity of the horizontal and the vertical perspectives of interpreting, and ascertained whether reading ahead exists in sight translation. Eye movements of 18 interpreting students were recorded during silent reading of a Chinese speech, reading aloud a Chinese speech, and Chinese to English sight translation. Since silent reading consists of the comprehension component while reading aloud consists of the comprehension and production components, the two tasks served as a basis of comparison for investigating comprehension in sight translation. The findings suggested that sight translation and silent reading were no different in the initial stage of reading, as reflected by similar first fixation duration, single fixation duration, gaze duration, fixation probability, and refixation probability. Sight translation only began to demonstrate differences from silent reading after first-pass reading, as shown by higher rereading time and rereading rate. Also, reading ahead occurred in 72.8% of cases in this experiment, indicating the overlap between reading and oral production in Chinese. VI.

(8) to English sight translation. The results supported the vertical perspective in interpreting as well as the claim of reading ahead. Implications for interpreter training are to attach more importance to paraphrasing skills and to focus more on the similarities between sight translation and simultaneous interpreting.. Keywords: sight translation, eye-tracking, vertical perspective, horizontal perspective, reading-ahead, simultaneous interpreting. VII.

(9) 摘要. 口譯的歷程包含理解（comprehension）、重組（reformulation）及產出（production）等三階段。然而目前翻譯研究中，仍缺少探討翻譯的三個階段如何及何時發生，以及翻譯過程中理解歷程之相關研究。在此前提之下，本研究使用眼動法調查視譯之閱讀理解歷程、探討序列式翻譯（vertical perspective）及平行式翻譯（horizontal perspective）兩者之合理性、並檢驗視譯過程中，閱讀與翻譯之口語產出是否發生重疊（reading ahead）之情形。研究分別記錄十八位受試者於中文閱讀、中文朗讀及中譯英視譯之眼動，比較三種不同情況下之眼動軌跡。由於閱讀包含翻譯訊息理解階段，朗讀包含理解及產出階段，皆與口譯牽涉的階段有所重疊，故可作為瞭解視譯歷程之比較基準。實驗結果顯示，視譯與閱讀在眼動的初期指標上相當一致，於晚期眼動指標上，才出現不同的認知處理歷程，故研究結果支持序列式翻譯之理論。實驗結果亦觀察到中譯英視譯過程中，理解及口語產出發生重疊的比例為 72.8%。根據本研究結果，視譯訓練應著重換句話說（paraphrasing）的技巧；另視譯與同步口譯有許多相似之處，可作為口譯教學或練習之參考。. 關鍵詞、同步口譯關鍵詞：視譯、視譯、眼動法眼動法、序列式翻譯、序列式翻譯、平行式翻譯、平行式翻譯、reading ahead、. VIII.

(10) Chapter 1. Introduction. 1.1 Research Background. As one of the world’s oldest profession, interpretation facilitates the cross-cultural communication necessary in today’s society as interpreters convert one language into another. Interpreters are language specialists who go beyond simply converting words into another language but relay concepts and ideas between languages. They must master at least two languages, have a good command of interpreting skills, and thoroughly understand the subject matter in which they work to accurately get the message across. In addition, interpreters must be sensitive to the cultures associated with their languages of expertise. Against these high demands, the 1950s and 60s saw the emergence of translation and interpretation (T&I) schools (Pöchhacker, 2004) aimed to cultivate students with the abilities required to become competent interpreters. In turn, T&I studies have also made leaps and bounds in terms of both quantity and quality. Different from translation, in which both the input and output are in the written form, interpretation is an oral form of translation which usually involves the oral input of the source language and oral production of the target language. It is a complex cognitive task which consists of three components including listening, deverbalization (also referred to as 1.

(11) reformulation or code-switching), and production (Seleskovitch, 1976). The interpreter listens to comprehend the source language, discards the form (eg. words or structures) of the input to reformulate the meaning during deverbalization, and produces the message orally in the target language. Among the three components of interpreting, comprehension plays a pivotal role because understanding the source language is a prerequisite to further processing any information. It has been estimated that an interpreter devotes 80% of his/her cognitive ability to listening and comprehension, and only 20% to speech production during interpreting (Bajo et al., 2001). The complexity of interpreting makes its study a challenging enterprise. If we are to fully understand how this task is performed, the processing of language comprehension, reformulation, and production need to be studied individually. However, while current interpretation studies have already identified the required efforts for interpreting (Gile, 1995), few have attempted to address how the three components of comprehension, reformulation, and production operate in relations to one another. Hence, the actual process of interpreting, up until today, has remained rather elusive or even debated. The lack of understanding towards the process of interpreting poses great challenges for T&I students and instructors in both learning and teaching. In skills acquisition, a novice needs to go through five different stages to become an expert (Tsui, 2003). When applied to interpretation training, the difficulty for both students and instructors lies in. 2.

(12) bridging the gap between the novice and the expert. Because both parties have very little idea of how the three components of interpreting occur inside the interpreter’s head, it is challenging for instructors to provide standardized methods to help students overcome difficulties. Generally, an instructor only sees students’ final speech production and gives feedback accordingly while students are able to witness instructors’ successful demonstrations but cannot gain knowledge on how to become more like the expert. To close the divide between the novice and the expert, light could be shed on the three components of interpreting, and especially on largely unknown and uninvestigated component of comprehension. By doing so, it will be easier to isolate and teach the demands of each individual component in interpreting and focus on its participation in the task as a whole. In turn, future T&I students and instructors benefit from learning and teaching more efficiently. One area in which we could begin to learn about interpreters’ comprehension is sight translation, a type of interpretation in which the source language is provided to the interpreter in the written rather than the oral form. Two main differences between the input in sight translation and in other modes of interpreting lie in the distinction between the oral and written language, and between the reading and listening process (Agrifoglio, 2004). Different from listening, reading comprehension can be observed from the eye movements of the reader while the speed of one’s comprehension is reflected in the speed of reading. 3.

(13) (Just & Carpenter, 1980). In this light, two types of interpreting involve the source language in its written form: sight translation (as mentioned above) and simultaneous interpreting with text. Sight translation is a mode of interpretation in which the input of the source language is written but the output of the target language is oral. It is an indispensable skill in T&I training which enhances the work efficiency of interpreters. At the same time, sight translation helps interpreter trainees react quickly and enhance their oral skills (Weber, 1990). Simultaneous interpreting with text, on the other hand, is an extension of sight translation, except that the interpreter receives both written as well as oral input to render the oral output. Since the two types of interpreting both involve written input, both could serve as the starting point in using the eye-tracking method to investigate the comprehension component in interpreting. By observing eye movements, one gains understanding of the physical reactions and cognitive activities of the person performing a cognitive task (Richardson, Dale, & Spivey, 2009). Monitoring eye movements during reading also provides valuable information regarding the moment-to-moment comprehension processes (Rayner, 1998, 2009; Rayner, Chace, Slattery, & Ashby, 2006). With its many applications, eye tracking has become one of the most important research methods in cognitive psychology ( 李筱娟 ,2007). Furthermore, the eye-tracking method has been employed in T&I research in recent years (Chang, 2009; Hyönä, Tommola, & Alaja, 1995; Pavlović & Jensen, 2009; Sharmin,. 4.

(14) Špakov, Räihä, & Jokobsen, 2008). However, most of the studies conducted with the eye-tracking method have focused mainly on translation, in which the output is written, as opposed to interpretation, in which the output is oral. From the field of interpreting, Dillinger (1994) emphasized that cooperative research is the only rational solution in future interpretation studies since interpreters and experimenters need to work with each others’ expertise. From the field of cognitive psychology, Rayner (1998) echoed this by highlighting that the future success of eye movement research depends on the ingenuity of researchers to design interesting and informative studies. For those very reasons, the present study sailed into unchartered waters and applied the eye-tracking method to investigate the three components of sight translation in general and the comprehension component in particular. This study may be critically important in laying the groundwork which will pave the way to revealing the process of sight translation, a fundamental mode of interpreting.. 1.2 Research Questions. The purpose of the present study was to explore the process of sight translation through tracking the eye movements of interpreting students during three tasks: silent reading of a Chinese speech, reading aloud a Chinese speech, and sight translation of a Chinese speech into English. Since silent reading and reading aloud include at least one component of 5.

(15) interpreting, they could be analyzed and compared to sight translation so that certain components of interpreting could be further understood. In detail, the primary research questions that we addressed are as follows: (1) To investigate how and when the comprehension component in interpreting occurs in sight translation; (2) To examine whether the comprehension and reformulation components overlap in sight translation, that is, to explore the validity of the vertical and horizontal perspectives in interpreting; (3) To ascertain whether the conventional wisdom of “reading ahead” is sound in sight translation, that is, to determine whether the comprehension and production components overlap during sight translation.. The remainder of this thesis is organized into four sections. Chapter 2 provides basic definitions of important concepts and reviews previous studies on sight translation and eye tracking. Chapter 3 describes the methodology, procedures, and results of an eye-tracking experiment on silent reading, reading aloud, and sight translation. Finally, Chapter 4 offers a general discussion on how the findings of the experiment can be applied to T&I while suggestions are made for future research.. 6.

(16) Chapter 2. Literature Review. 2.1 Definitions: types of interpreting and language processing. It is vital to understand the types of interpreting as they are the basics to why sight translation is of enormous value to interpretation training. Generally speaking, interpreting can be categorized into two major modes: simultaneous interpreting (SI) and consecutive interpreting (CI). Simultaneous interpreting (SI) means new input is continuously presented and the interpreter comprehends the continuously incoming input while simultaneously reformulating the message and producing it orally in the target language. In contrast to the immediacy of SI, output in consecutive interpreting (CI) begins only after the speaker has verbalized a group of words or sentences. The interpreter alternates between listening and speaking, and only starts to translate after the speaker has finished speaking. In SI, the interpreter multitasks and coordinates various efforts while in CI, the interpreter utilizes note-taking skills and capitalizes on short-term memory skills. The two modes of interpreting require distinct language processing skills while different training programs are designed in T&I schools to equip interpreter with relevant capabilities.. 7.

(17) 2.1.1 Simultaneous Interpreting (SI). In real-life simultaneous interpreting (SI) settings, the speaker speaks continuously and does not pause for the interpreter to render his/her oral translation. In conference settings, the interpreter usually sits inside a sound-proof booth and wears headphones to listen to the speaker’s delivery. While the interpreter listens to the speaker, he/she talks into a microphone to the audience, who are also wearing headphones. The interpreter listens to the source language and orally produces the target language while at the same time still listening to the speaker’s continuously incoming segments. In other words, the interpreter continuously hears new input while simultaneously comprehending the input and stores segments of it in memory. While this is happening, an earlier segment has to be reformulated mentally into the target language, and an even earlier segment has to be orally produced (Christoffels & De Groot, 2005; Liu, Schallert, & Carroll, 2004). Furthermore, the interpreter, while already listening to the source language and producing it in the target language, has to listen and monitor his/her own speech production to ensure no mistakes are made. The simultaneity of comprehension and production imposes a severe strain on cognitive processing capacity. This is one of the reasons that SI is such a cognitively demanding task, which also explains why professional interpreters normally work in pair or groups of three for 20-minute periods each (Christoffels & De Groot, 2004; Lambert, 2004). The simultaneity of SI also means that the speaker and the interpreter both speak at the 8.

(18) same time, yet the interpreter’s orally produced content falls behind the speaker as the he/she needs to listen and understand the message before producing the oral output.. Gile (1995) proposed an effort model for simultaneous interpreting: Simultaneous Interpreting= listening and analysis effort + short term memory effort + speech production effort + coordination effort. The linguistic input is oral in SI, and therefore listening and analysis efforts both play critical roles. The interpreter needs to store information that is heard in his/her short-term memory for the time interval between the moment the speech is heard and the completion of its target language production (Agrifoglio, 2004) while all of the aforementioned efforts need to be coordinated. Gile’s (1995) effort model proved that SI is a cognitively demanding task since the coordination of many types of efforts is required. Gerver (1976) also pointed that SI is a complex task for it involves perception, storage, retrieval, transformation, and transmission of verbal information. As the most widely used form of interpreting in international. 9.

(19) conference settings, SI has been a core subject in T&I training programs while students have had to undergo rigorous training and extensive practice to master its skills.. 2.1.2 Sight Translation (ST). Sight translation (ST) is a form of interpreting in which the interpreter’s linguistic input is in the written rather than the oral form. During the process of ST, the interpreter reads the source text while rendering the oral interpretation in the target language (Weber, 1990). Unlike simultaneous interpreting, in which the interpreter has no control over the speed of the input, the interpreter can control the speed in which the written input is perceived. However, the task of ST is still challenging because the demand on the quality of oral production is very high. ST is perceived as an oral translation of a written text that should sound as smooth as if the interpreter were merely reading a document written in the target language (Angellini, 1999). Any pause of over 2 seconds would be considered an error (楊承淑, 2005). The difficulty of ST lies in that fact that the interpreter needs to read the source text, comprehend its content, translate and produce the speech in another language while monitoring his/her oral production (Syysnummi, 2003). In this regard, ST resembles simultaneous interpreting because it also involves multitasking.. 10.

(20) Gile (1995) proposed the effort model of sight translation: Sight Translation = reading and analysis effort + speech production effort. According to Gile’s effort model, ST consumes reading and analysis efforts as well as speech production efforts. Hence, ST is regarded as a combination of interpretation and translation, which echoes Lambert’s views (1989). It has been argued that ST is difficult not because of the written form of the source text but because of the interpreters needs to smoothly coordinate the reading, memory and production efforts while working to avoid the interference of the source language. Gile (1995) further explained that the listening and analysis effort becomes a reading effort in ST while the production effort remains. Since information is always available on paper, there does not seem to be a memory effort similar to the one in simultaneous or consecutive interpreting. In contrast, Agrifoglio (2004) claimed that there seems to be a memory effort involved in ST, which is similar to the short-term memory demands of simultaneous interpreting, because the syntactic differences between languages may force the interpreter to store some information in memory until it could be appropriately produced in the target language. The two opposing claims still need to be tested by further evidence and the results may vary between different language combinations. However, both. 11.

(21) of these assumptions shed some light on the necessary efforts of ST, which serve as foundation for further research. T&I scholars such as Weber (1990), Moser-Mercer (1994), Lambert (2004), and Sampaio (2007) have highlighted the benefits of ST, which has been considered an ideal pedagogy in interpreter training programs for several reasons: (1) the interpreter becomes familiarized with the technical terms in context and develops immediate reflexes of these terms; (2) the interpreter can rehearse speech texts thoroughly in advance before the actual interpreting assignment; (3) the interpreter develops skills of speed reading and gives more fluent production after reading the source text (Weber, 1990). In terms of the applications, ST is used usually, though not exclusively, in judicial and medical interpreting. It is also an essential skill applied when the speaker reads from a prepared speech. In the US, the National Association of Judiciary Interpreters and Translators (NAJIT) offers a rigorous examination including two sight translation tests from the first language into the second language and vice versa. In Brazil, professionals have to qualify for exams administered by the Board of Trade which include sight translation to become qualified sworn-in interpreters (Sampaio, 2007).. 2.1.3 Simultaneous Interpreting with Text. Another valuable use of sight translation is in simultaneous interpreting with text (SI 12.

(22) with text). In SI with text, the interpreter receives two sources of input: listening to the speaker’s oral presentation and also reading a written text. SI with text could be defined as simultaneous interpreting with the extra task of sight translation As opposed to sight translation, SI with text is one step closer to simultaneous interpretation as the source language is presented both orally and visually (Lambert, 2004). Usually in authentic interpreting settings, the interpreter obtains the speaker’s text beforehand and the speaker reads aloud the text during the actual speech. Although SI with text is not performed by participants in this study, it is still worthwhile to mention since it is an extension of sight translation and even one step closer to authentic interpreting settings. In SI with text, the interpreter devotes efforts to both listening and reading. Gile (1995) did not propose an effort model for sight interpretation, but judging from the efforts needed in both sight translation and simultaneous interpreting, the efforts needed in SI with text include the listening and analysis effort, reading and analysis effort, production effort, and coordination effort. To sum up, the benefits and importance of sight translation have proved to be self-evident. ST encompasses all the essential abilities of a conference interpreter and enhances the cognitive processing speed of the interpreter (Weber, 1990). At the same time, the rapid and efficient visual-brain-vocal coordination required by ST standards serves as the foothold which helps an interpreter master consecutive and simultaneous interpreting. 13.

(23) skills (Sampaio, 2007).. 2.1.4 Skills of Sight Translation. Weber (1990) pointed out that the guidelines to sight translation include the following: (1) analyzing a text rapidly; (2) producing the meaning rather than a word-for-word interpretation; (3) rapid conversion of information from one cultural setting (language) to another; (4) public speaking techniques. Before actually proceeding with sight translation, student interpreters should skim through the speech quickly while conducting segmentation and making marks to indicate the order in which the speech will be interpreted (何慧玲, 1997). From the author’s experience as an interpreting student and practitioner, strategies of sight translation often taught by instructors of interpreting include the following: (1) Scanning a document rapidly for content and style; (2) Analyzing units of meaning which form each sentence; (3) Anticipating syntactic rearrangement necessary in the target language; (4) Rendering sight translation in the target language while reading ahead to prepare to produce next units of meaning; (5) Rendering sight translation with accuracy and fidelity to the text; (6) Employing effective speech skills: the individual needs to use clear diction, 14.

(24) appropriate pauses and intensity, delivers the message with fluidity, and in a well modulated voice. Although certain guidelines have been proposed for training of ST, Sampaio (2007) noted that literature which document the sight translation pedagogy has been scant. This is very likely due to the fact that no research findings are yet available concerning the cognitive process of ST, which will be discussed in the following section.. 2.2 The interpretation process and the comprehension phase. Theories of interpretation have noted the importance of comprehension process in the interpretation task (Dillinger, 1994). However, beyond the comprehension process, interpreters perform a reformulation or code-switching process between the two languages and produce the output in the target language. Generally speaking, interpreting can be categorized into three components which include comprehension, reformulation (also referred to as code-switching), and target language production (Gerver, 1976; Seleskovitch, 1976).. 2.2.1 The vertical and the horizontal perspectives. Despite the fact that theorists agree about the components of interpreting. 15.

(25) (comprehension, reformulation, and production), there exist two different views on the way these operations occur, namely, the vertical perspective and the horizontal perspective (Macizo & Bajo, 2004, 2006). The vertical perspective is also referred to as the meaning-based strategy. The interpreter is thought to retain the meaning of information chunks during comprehension to reformulate the meaning, and to produce it in the target language (Fabbro & Gran, 1994). Meaning-based interpreting is conceptually mediated and the input is fully comprehended in a way similar to ordinary comprehension. The interpreter’s job is to give lexical expression to the meaning extracted from the full comprehension of the input. The vertical perspective is also in line with the deverbalization theory proposed by Seleskovitch (1976). The theory claimed that interpreting involves first the processing of information in the source language to obtain its meaning. Second, after the comprehension process is complete, the message is restructured according to target language grammar while specific linguistic form of the source language is discarded. This is the so-called deverbalization process, which occurs only after the comprehension process has been completed. The message is then reformulated to be produced in the target language. According to this strategy, interpreting involves full comprehension of the source language in a way similar to common comprehension of speech (Christoffels & De Groot, 2005; Macizo & Bajo, 2006). Therefore, from the vertical perspective, comprehension and reformulation are. 16.

(26) performed sequentially rather than concurrently without any direct links between the source language and target language at the lexical/syntactic levels of analysis (Macizo & Bajo, 2004, 2006). Figure 1 (Macizo & Bajo, 2004, 2006) shows the sequence of processes involved in interpreting under the vertical perspective/meaning-based strategy. The left hand side refers to the interpreter’s understanding in the source language (SL) while the “abstract” part indicates the extraction of the meaning of the SL. The right hand side shows the production in the target language (TL) after obtaining the meaning of the original message.. Figure 1. The vertical perspective/meaning-based strategy (Macizo & Bajo, 2004). In contrast to the vertical perspective, there is a varying view called the horizontal perspective, or the transcoding strategy. The horizontal perspective sees interpreting as the direct processes of recoding from one linguistic code to another. The interpreter may. 17.

(27) engage in partial reformulation and seek the equivalent of the smallest meaningful unit in the TL while still reading and comprehending the source text. The lexical units in the TL are supposed to be activated continuously in a parallel manner, before the source language (SL) meaning units are fully comprehended. The horizontal approach has also been referred to as a word-based or word-for-word strategy (Fabbro, Gran, Basso, & Bava, 1990). However, it does not mean literally that words per se serve as the transcoding unit in interpreting. Rather, it indicates that each meaning unit is reformulated before the comprehension process of that meaning unit has been completed. In other words, comprehension and reformulation occur concurrently rather than serially, which is opposed to the claim of the vertical perspective. Figure 2 (Macizo & Bajo, 2004) shows the sequence of processes involved in interpreting from the horizontal perspective/transcoding strategy. The left hand side refers to the interpreter’s understanding in the source language (SL) while the arrows pointing from the TL to the SL at the lexical, syntactic, and discourse levels indicate the ongoing transcoding or reformulation processes at different levels during the course of the comprehension process. The right hand side shows the production in the TL after reformulation of the original message.. 18.

(28) Figure 2. The horizontal perspective/transcoding strategy (Macizo & Bajo, 2004). It should be noted that the horizontal and the vertical perspectives are the possible approaches for interpretation strategies rather than proven theories. Also, what the unit is for these two perspectives has not yet been specified in literature. However, what is certain is that for the vertical perspective, SL comprehension plays a pivotal role in interpreting. There is no parallel access to the TL as the interpreter receives SL input. Because reformulation only occurs after interpreters have extracted the SL meaning, normal reading and reading for the purpose of interpreting should impose similar demands on the interpreter’s working memory. In contrast, for the horizontal perspective, partial reformulation already takes place while the interpreter reads the SL. The partial reformulation process consumes working memory and adds a greater loading to the cognitive resources than that for normal comprehension. As a result, reading processes. 19.

(29) would be more demanding when reading for the purpose of interpreting because of the extra demands on working memory. Also, the increased cognitive load would be especially high when comprehension of the SL is difficult. Macizo and Bajo (2006) conducted two types of self-paced reading experiments to determine whether the horizontal and vertical perspectives was valid. Their prediction was that if reading for the purpose of interpreting took longer time than normal reading, this would be evidence for the horizontal perspective. In one type of experiment, the task (reading for repetition or reading for interpreting from Spanish into English) and the lexical ambiguity of the target word (ambiguous: homograph or unambiguous) were manipulated within participants. Memory load (low or high) was a between-groups variable, which was manipulated by varying the number of words between the target word and the disambiguating context (5 words versus 7 words). Sixteen professional translators were divided into two groups composing the two memory load conditions. The stimuli were sentences which appeared word-by-word in the middle of a computer screen. Participants were told to repeat the sentence or to interpret the sentence. They could read at their own pace by pressing the space bar every time they wanted to see new words. The time between consecutive key presses was taken as an index of the processing time for the displayed words. The same experiment was repeated on 16 Spanish-English bilinguals. The findings suggested that when participants read and interpreted sentences, global. 20.

(30) comprehension and the speed of the reading processes were affected by the presence of lexical ambiguity and memory load. Reading for interpreting became slower and understanding became less accurate when the sentences contained ambiguous words and the distance between the ambiguous word and the disambiguating context was large (high memory load condition). In contrast, when participants were instructed to only read, understand, and repeat the sentences, the presentation of an ambiguous word did not affect reading times in either of the two memory load conditions. Macizo and Bajo claimed that whereas reading for interpreting requires working memory resources for parallel activation of the TL lexical entries and switching the two languages involved, reading for repetition does not need these additional resources. The results were in agreement with the predictions of the horizontal perspective. In the other type of experiment, Macizo and Bajo (2006) tried to prove that there was parallel activation of TL lexical entries when reading for the purpose of interpreting. Sixteen professional translators were asked to read sentences which contained cognate words (words that resemble its target language equivalent, eg. “cebra” in Spanish vs. “zebra” in English) at the beginning and at the end of sentences for the purpose of repetition and interpreting. The same experiment was repeated on 16 Spanish-English bilinguals. The prediction was that the presence of cognates would facilitate comprehension when reading for interpreting. Results showed that when the cognates were at the end of the. 21.

(31) sentence in the source language, reading times were shorter when reading for interpretation than reading for repetition. In other words, the study concluded that interpreters’ comprehension was facilitated by the presence of cognates at the end of the sentence. The cognate effect was not observed when reading for repetition. Hence, the results indicated the activation of the TL during reading for the purpose of interpreting and thus supported the horizontal perspective. Results of studies by Macizo and Bajo (2004, 2006) supported the horizontal perspective. As they claimed to prove, reading for the purpose of interpreting demanded more working memory resources than within-language reading and the additional demand was due to parallel activation of TL lexical entries. However, even though these results yielded support for the horizontal perspective, questions still remain when it comes to the comprehension process in sight translation. For one, stimuli used in their experiments were sentences rather than passages, which deviated from the normal practice in sight translation. Secondly, the display of each stimulus was controlled by each participant voluntarily as they pressed the spacebar key when they wanted to see a new word in the sentence on the computer screen. The time between consecutive key presses was taken as an index of the processing time for the displayed words. This could be a confounding factor since this is not a natural setting for normal reading. Thirdly, participants were told either to repeat or to interpret the stimulus. Short-term memory efforts may have prolonged the time between. 22.

(32) consecutive key presses, since the interpreter needed to remember the content that they were supposed to produce orally. Under this light, a method which reveals the moment-to-moment information of reading comprehension would be more suitable for investigating such a process. McDonald and Carpenter (1981) employed the eye-tracking method to explore interpreters’ chunking strategies and how error detection occurs during sight translation. Though their purpose was not to validate either the horizontal or vertical perspective, the results showed that the average first-pass (when the eyes fixate on a region for the first time) reading rate for interpretation is very similar to the rate of normal silent reading, and that interpreters began to produce interpretation after the first-pass. It was proposed that the additional time spent after the first-pass is the process needed to compose an interpretation, to take into account the syntactic constraint of the target language, and to output at least some of the interpretation. This study showed that interpretation builds on normal comprehension processes and that the task of interpretation partially reflected normal reading processes. Since the experiment showed that reading for interpretation is similar to normal reading and that comprehension and reformulation occur serially, it yielded support for the vertical perspective. Furthermore, the eye-tracking method which was used to conduct the study provided more detailed information regarding moment-to-moment comprehension process.. 23.

(33) 2.2.2 Reading ahead in sight translation. Besides discussion on the sequence or relationship between comprehension and reformulation in interpreting, there have also been discussions on whether production overlaps with comprehension in interpreting. Training in T&I programs has focused on “reading ahead” in sight translation (ST). Weber (1990) remarked that ST mainly consists of interpreters’ reading ahead what he/she is orally producing in the target language. This is done to avoid any hesitation or prolonged pauses during the delivery of interpretation. Sampaio (2007) commented that in ST, the oral production is practically concurrent with the comprehension of the text. Weber (1990) emphasized that a central ability to develop is to reformulate ideas within a short period of time while devoting the entire concentration to the analysis of the following meaning units. Agrifoglio (2004) suggested that although in ST, the interpreter can control his/her rhythm of perception, smooth delivery is only possible when he/she starts reformulating while still reading. From this view, many would agree that the reception and analysis phase overlaps with the production phase in ST. In other words, the interpreter is perceived to be producing a target-language version of the previous sentence while reading/listening to the current sentence. The aforementioned assumptions suggest that during ST, the interpreter is orally interpreting while reading the upcoming text in order to render the interpretation. 24.

(34) Conventional wisdom accepted in the field of interpreting are in line with this view. But essentially this proposal is an intuitive hypothesis, lacking empirical investigation. To date, little literature has been published on the overlap of comprehension and oral production in Chinese to English sight translation. Therefore, another focus of this study is to investigate the phenomenon of reading ahead in detail. It is believed that the eye-tracking method will yield clues in terms of the extent of the feasibility of reading ahead.. 2.3 How eye movements reveal the process of reading comprehension. Just & Carpenter (1980) proposed two assumptions for the relationship between eye movements and reading comprehension. The first was the immediacy assumption, meaning that readers immediately process the fixated word so that information processing is not deferred but happens immediately upon fixation. The second assumption was the eye-mind assumption, which means the eye remains fixated on a word as long as the word is being processed. Also, the time needed to process a newly fixated word was directly indicated by fixation duration. Based on those two assumptions, oculomotor movements have, to some extent, indicated the mental processes that take place in reading (Inhoff & Radach, 1998; Just & Carpenter, 1980). Eye movements are necessary because due to the anatomical limitation of the eye, the central portion of the retina, known as the fovea, subtends only about two degrees visual 25.

(35) angle. Humans are able to perceive stimulus with the highest visual acuity at the fovea. Vision outside the fovea deteriorates and hence readers need to move their eyes to the location of the stimulus where the fovea could perceive it clearly, constituting eye movements (Rayner, 1998; Richardson, et al., 2009). Even though there are instances in which attention is allocated independently of eye positions in simple tasks, Rayner (2009) contended that in complex tasks, eye location and attention are at the same location while attention precedes the eye to the next target. In other words, the eyes move to where attention is allocated, which provides the explanation to why eye movements could be used as indices to human cognitive processes during reading comprehension. Eye movements in reading are composed of saccades and fixations, which serve as indicators to information processing. Saccades are the series of rapid movements which move the eyes from one fixated location to another during reading. The pauses separating saccades are called fixations, which generally last between 200 – 250 ms. Fixations yield valuable information about the process of reading because new information is encoded and processed only during fixations (Radach & Kennedy, 2004; Rayner, 1998). The advantage of the eye-tracking method is that its data not only provides the behavioral end products of our cognitive process, but also offers us clues to the process through which they are achieved. Eye movements can be captured instantaneously to provide on-line indicators which effectively manifest the details of the mental process and. 26.

(36) are tightly linked to moment-to-moment goals and subject-tasks (Duchowski, 2003). More importantly, this measure of cognitive processing does not interrupt task processing (Richardson, et al., 2009). With the eye-movement recording technique, readers can read sentences under normal conditions. Hence, the eye-movement method proves to be an ideal tool for T&I studies. Comprehension of the source text in sight translation belongs to the realm of reading. Therefore, reading comprehension studies is helpful in understanding how an interpreter comprehends written text, and may possibly help interpreters overcome difficulties in comprehension (楊承淑, 2000). In light of this, the eye-movement recording technique was adopted in the present study. The aim was to capture the process of how the interpreter comprehends the source text and how the demand of sight translation affects reading comprehension.. 2.4 The eye-tracking method in T&I Studies. To date, there have only been a limited number of studies which applied the eye-tracking technique to T&I research. Though very scarce, previous studies have lent strong support to the applicability of the eye-tracking method to studying T&I. The review below is by no means exhaustive, yet provides a large picture of past research relevant to the present study. 27.

(37) 2.4.1 The eye-tracking method in interpretation studies. A pioneering study which applied the eye-tracking method to interpretation studies was conducted by McDonald and Carpenter (1981). The study examined the processes in sight translating ambiguous phrases and how these phrases were parsed during comprehension and interpretation. Also, how errors were detected during sight translation was also investigated. Two expert interpreters and two amateur German-English bilinguals sight translated 44 texts from English into German while their eye fixations and oral translations were recorded. The texts contained idiomatic phrases such as “hit the nail on the head” or “break the ice”, which could be comprehended and interpreted literally or idiomatically under two different contexts which either primed a literal interpretation or an idiomatic interpretation. This was done by analyzing the chunking of ambiguous phrases manifested through eye movement indices. A chunked unit was defined by the farthest word to the right that was fixated in a first-time forward fixation scan before a regression, also called the first-pass reading. Comprehension of an idiomatic phrase was classified as idiomatic if the entire phrase was read as one chunk without any regressions in first-pass reading. For example, the entire phrase of Mike hit the nail right on the head, if read in one pass, would be regarded as idiomatic comprehension. In contrast, comprehension was classified as literal if the phrase was chunked at any of the syntactic boundaries in the sentence. For example the chunking Mike hit / the nail right on the head or Mike hit the nail 28.

(38) / right on the head, both not read in one entire chunk but parsed by regressions at syntactic boundaries (as indicated by the slash), were both classified as literal comprehension. Interpreters’ eye fixation patterns not only demonstrated how idiomatic phrases were interpreted but also reflected the subprocesses of interpretation. Each phrase received at least two scans: the initial comprehension of the phrase was marked by the initial sequence of forward gazes, which is the first-pass reading in eye movements. The subsequent process was marked by regressions, which constituted the second scan or the rereading of the phrases. This second scan (also referred to as second-pass reading) was also when oral interpretation took place. Further, if the participant detected an error in his/her comprehension/interpretation after the second pass, the eye regressed back again to the preceding part which contained the ambiguous idiomatic phrase while the previous oral interpretation was corrected. The results of this study showed that interpreters’ eye fixation patterns depended on the preceding context as well as semantic and syntactic cues. Although this study was not aimed at exploring the process of sight translation, it revealed that the initial process of comprehension in sight translation was similar to the comprehension processes that occur in normal English reading. This was evidenced by the fact that the reading speed for the initial reading of a phrase in sight translation was similar to that for normal reading (about 200 to 300 words/minute). These results supported the vertical perspective, because readers need. 29.

(39) to comprehend first before proceeding to interpret a meaning unit. Figure 3 depicts a model proposed by McDonald and Carpenter (1981). The first phase “read, integrate and parse” was manifested in the first-pass reading of eye movements. The second phase “reread for lexical retrieval in target language” was manifested in rereading, while error recovery was manifested by regressions to the previous fixated regions, as shown by the two arrows pointing back to the previous regions on the right hand side.. Figure 3. The processes of the interpretation task (McDonald & Carpenter, 1981). In sum, this research was groundbreaking for it applied the eye-tracking method to interpretation study, yielding insightful information to the parsing of ambiguous phrases, error detection, as well as the process of comprehension in sight translation. Despite the findings, this study was not intended to determine the validity of the vertical and horizontal perspectives. Even though it was mentioned in this study that initial reading time of a 30.

(40) phrase in sight translation was similar to that of normal reading, normal reading was not included as one of the conditions in the experiment. Until the present study, there has been a lack of follow-up research which continued the investigation into sight translation with the eye-tracking method. Moreover, there have not been studies targeted at comparing the process of sight translation with normal reading in Chinese nor the respective components of sight translation (comprehension, reformulation, and production). Other studies used the pupil diameter as a measurement in interpretation research (Hyönä, et al., 1995; Tommola & Hyönä, 1990; Tommola & Niemi, 1986), since pupil diameter was positively correlated with cognitive loading (Rayner, 1998). Chang (2009) commented that Tommola and Niemi (1986) were the first to use pupil diameter as a measurement of cognitive loading during simultaneous interpreting. In this study, one participant conducted simultaneous interpreting (SI) of five Finnish texts into English. The results showed that the participant’s pupil diameter was the highest during interpreting when restructuring of the output English sentence was required because of the syntactic differences between the source language (Finnish) and the target language (English). Another example of the aforementioned studies was undertaken by Hyönä et al. (1995) to understand the cognitive processing load in simultaneous interpreting (SI), speech shadowing, and passive listening. Two experiments were conducted in this study: in Experiment 1, nine Finnish-English T&I students conducted SI, shadowing, and passive. 31.

(41) listening to a text of 500-600 words from English into Finnish. The findings showed that SI produced a higher degree of pupil dilation than did speech shadowing, and speech shadowing yielded a higher level than did listening (average pupil diameter: listening=4.20mm, shadowing=4.72mm, SI=5.22mm). In Experiment 2, 18 T&I students participated in a 2x3x2 within-subject design experiment. The factors were input language (Finnish and English), task type (listening, shadowing, oral translation), and word translatability (easy, difficult). Effect of task type showed that overall average increase in pupil size was largest for lexical translation (0.33mm), followed by shadowing (0.24mm), and then listening (0.16mm). Effect of language direction indicated that the pupil dilated more for English words (0.27mm) than for Finnish words (0.22mm), demonstrating that repeating back words in a non-native language (English) was accompanied by increased pupil dilations, in comparison to repetition in the subject’s native language. Effect of word translatability was also manifested: words that were determined to be more difficult to translate induced a higher increase in pupil size (0.29mm) than did easily translatable words (0.20mm). This study lent strong support to the use of the pupillary response as an indicator of cognitive processing load.. 2.4.2 The eye-tracking method in translation studies. Aside from the studies mentioned so far, previous work which employed the 32.

(42) eye-tracking method have mostly focused on translation rather than interpreting, while most of them dealt with the issue of cognitive processing load (Chang, 2009; Dragsted & Hansen, 2008; Jakcobsen & Jensen, 2008; O'Brien, 2007; Pavlović & Jensen, 2009; Sharmin, et al., 2008; Sjørup, 2008). This has to do with the fact that whereas the mode of input for interpretation is mostly oral, the mode of input for translation is written, and hence the eye-tracking method can be more widely applied. One example of such studies was conducted by O’Brien (2007). Four professional translators’ pupil dilation measures and processing speed were recorded when using SDL Translator’s Workbench, a translation memory tool. Eye-movement data was measured under four different conditions representing the varying degrees of high to low match between the source text and the translation memory tool database. The results showed that the source text that had the least matches to the translation memory tool database (No Match) required the most cognitive efforts while source text that had the most matches (Exact Match) required the least efforts, as reflected in the average percentage change of pupil dilation (No Match: 11.84%, Exact Match: 5.83%) and the average number of source-text words processed (No Match: 0.32 words/minute, Exact Match: 4.11 words/minute). Pavlović and Jensen (2009) applied the eye-tracking method to investigate cognitive loading in the two translation directions between Danish and English and also between the. 33.

(43) source text (ST) and target text (TT). Indices including gaze time, average fixation duration, total task length and pupil size were analyzed from the T&I performance of eight translation students and eight professional translators. The source text and target text were displayed on the same computer screen in split windows. Five hypothesis were tested: (1) processing the target text (TT) required more cognitive efforts than the source text (ST) in both directions of translation; (2) translating into the second language (L2 translation task) on the whole required more efforts than translating into the native language (L1 translation task); (3) cognitive efforts in the processing of source text (ST) was higher in L2 to L1 translation (in which the ST is in L2) than L1 to L2 translation (in which the ST is in L1); (4) cognitive efforts in the processing of the target text (TT) was higher in L1 to L2 translation than in L2 to L1 translation; (5) students needed to invest more cognitive efforts to translation tasks than professionals in both directions of translation. The results of this study wholly confirmed hypothesis (4): in both directions of translation, processing the target text (TT) demanded more cognitive loading than the source text (ST), as reflected in significantly higher gaze time, average fixation duration, and pupil dilation. In L1 and L2 translation tasks alike, participants fixated longer on the TT than on the ST (81.2% more when translating into L1 and 118% more when translating into L2). Their average fixation duration values were higher on the TT than ST in the L1 translation task by 53.1% and the L2 translation task by 55.1%. The pupil dilation values. 34.

(44) were also higher for the TT than ST in both L1 and L2 translation tasks (2.4% and 2.6% higher respectively). Chang (2009) conducted two experiments which used pupil size, number of fixations, task time, and blink frequency to explore cognitive efforts in translational directionality. In Experiment 1, 15 native Chinese participants performed four tasks: translating from English into Mandarin, typing English, translating from Mandarin into English, and typing Mandarin. In Experiment 2, eight native Spanish participants performed six tasks including reading Spanish, reading English, typing Spanish, typing English, translating from English into Spanish, and translating from Spanish into English. Findings showed that (1) translating the first language into the second language was more cognitively demanding than translating in the opposite direction at a textual level; (2) reading the second language was more cognitively demanding then the first language.. To conclude, previous T&I studies using the eye-tracking method have yielded results on cognitive loading and translation directionality. What has not yet been addressed, however, is the use of the eye-tracking method to explore the process of comprehension in interpreting to address specific issues of the horizontal and vertical perspectives and to explore the details of reading ahead in sight translation. Furthermore, eye-movement measures employed in past studies mainly centered upon pupil size, number of fixations,. 35.

(45) and fixation duration. Number of fixations and fixation duration provide preliminary information to the general efforts demanded in reading comprehension. However, fixation duration and number of fixations can be further categorized into more sophisticated eye-movement measures including first fixation duration, single fixation duration, gaze duration, fixation probability, refixation probability, go-past time, regression-out rate, rereading time, rereading rate, and total viewing time. By doing so, we will be able to gain more insight into the efforts needed during the various stages of reading comprehension as opposed to merely a general view. On the other hand, while pupil size reflects cognitive loading of different tasks, the present study did not measure pupil dilation since it is more of a physiological response which does not reveal the various phases of reading comprehension which can be shown by the eye-movement indices mentioned above.. 36.

(46) Chapter 3. Experiment: eye tracking silent reading, reading aloud, and sight translation. In this experiment, eye movements of interpreters were recorded during three tasks: silent reading, reading aloud, and sight translation. Sight translation contains the three components of interpreting (comprehension, reformulation, and production). Reading aloud consists of the comprehension and production components while silent reading consists of only the comprehension component. Table 1 is a summary of the three components of interpreting involved in this experiment. The intent of this experiment was to compare the eye movement indices of sight translation with two tasks containing one or two of the three components of interpreting. If interpreting and the two other tasks that contain all but one of their components was systematically compared, the demands of each individual component and its participation in the task as a whole can then be isolated and learned (Bajo, et al., 2001). Both the eye movement data as well as the oral production data of participants were recorded so that the two modes of information could be matched up to understand the relationship between comprehension and production.. 37.

(47) Table 1.Correspondence of the three components of interpreting to the three conditions of this experiment Silent Reading. Reading Aloud. Sight Translation. Comprehension. V. V. V. Reformulation. --. --. V. Production. --. V. V. The experiment was designed to answer the following research questions: (1) When and how each component of interpreting occurs in sight translation and how the three components of interpreting operate with one another, with special focus on comprehension in sight translation; (2) Whether eye movements during sight translation reflect a vertical or horizontal perspective in interpreting. Experiment results yielded by McDonald & Carpenter (1981) were opposed to the results by Macizo & Bajo (2004, 2006), while there were also some who claim that the two perspectives may not be mutually exclusive and can both be applied by the experienced interpreter (Christoffels & De Groot, 2005) . According to the vertical perspective, the interpreter’s eye movement patterns in sight translation would be similar to that of silent reading in the initial stage of reading (when the eyes fixate for the first time on a region) because the process of comprehension would be similar for the two tasks. However, in the later stage (after the first time the eyes fixate on a region and move back to that region) of sight 38.

(48) translation, eye movement patterns between sight translation and silent reading would be distinct as the interpreter needs to perform reformulation and production. On the other hand, according to the horizontal perspective, sight translation would consume more efforts than silent reading in the initial stage of reading because reformulation would already be taking place. (3) Whether “reading-ahead” (Agrifoglio, 2004; Sampaio, 2007; Weber, 1990) exists in sight translation. Should “reading ahead” be valid, the oral production of a current segment would overlap with the reading of the following segment during sight translation. In other words, when the first fixation on a region occurs, the interpreter is either orally producing in the target language the content of a prior region, orally producing the content that he/she is currently fixating, or not orally producing any output in the target language (eg. pausing). The first situation would be regarded as reading ahead while the second and third situation would not.. 3.1 Method. 3.1.1 Participants. The participants for this research were selected from the population of interpretation students enrolled at graduate programs of translation and interpretation in Taiwan. Eighteen. 39.

(49) students (ages 23 – 40) were paid to participate in this experiment. All participants’ native language was Mandarin Chinese and second language was English. They had passed T&I graduate program entrance exams and completed sight translation courses in the first year of their graduate program. Participants were fluent in both Chinese and English and capable of basic sight translation skills. All had normal or corrected-to-normal vision. All participants signed an informed consent form before participating in the experiment.. 3.1.2 Design. A one-factor, three-level (silent reading, reading aloud, and sight translation), within-subject design was employed. Six passages were divided into three blocks, each consisting of two passages. Participants performed tasks in different orders of rotating blocks. Each block was silently read, read aloud, and sight translated by the 18 participants an equal number of times. This design was adopted to assess the eye movement patterns of a same text in three modalities by different interpreters. Participants were asked to answer two comprehension questions after performing different tasks to each passage to ensure that they comprehended the content despite different demands of the tasks. Before each block, participants practiced on a passage in advance to familiarize themselves with the upcoming procedures.. 40.

(50) 3.1.3 Materials. Materials used in this experiment consisted of six Mandarin Chinese speech texts of approximately 150 words which were all excerpts from authentic speech texts. Part of the texts were excerpted from real speech texts used in the Chinese to English sight translation training program at the Graduate Institute of Translation and Interpretation, National Taiwan Normal University, while the other part of the texts were excerpted from model speeches written by members of the Chinese branch of Toastmasters International. It should be noted that the texts were not artificially composed to include certain linguistic characteristics but were originally written to be communicative and read to an audience. The experiment sought to eliminate major discrepancies in the difficulty level of test materials. Also, passages of general rather than specific topics were preferred to avoid the bias resulting from participants’ familiarity of a topic. In light of these concerns, the most important criterion for material selection was that all texts were comprehensible to the general public while manageable (eg. interpreted smoothly with very little difficulties) for sight translation to novice interpreters. Materials selected were also of varying topics to prevent participants from becoming overly familiar with the content through repeated practice (eg. sight translating or reading speech texts which were all about a similar topic). Nine pieces of speech texts, which were determined by the experimenter to be of similar levels of difficulty, were first selected. The nine passages were evaluated for their 41.