La influencia del conocimiento previo en la interpretación simultánea de discursos especializadosun estudio empírico

Resumo

The role of prior knowledge in the simultaneous interpreting of specialized speeches This doctoral research explores the role of advance preparation in the simultaneous interpreting of specialized speeches. Specifically, this research is based on three theoretical and empirical pillars: the role of prior topic knowledge in the comprehension of specialized discourse; the cognitive process of interpreting and the development of expertise in simultaneous interpreting; and the current state of the art in the study of advance preparation in Interpreting Studies. Theoretical framework It could be said that simultaneous interpreting is the mode of linguistic mediation most often used in technical and scientific meetings and conferences. In these bilingual or multilingual events, the speeches have a series of features in which prior topic knowledge plays a crucial role, more relevant than for other genres like narrative texts (Graesser, León and Otero, 2002). Scientific discourse is a text and discourse genre whose main objective is to explain and persuade. This discourse is produced by members of a group, discipline or profession (a discourse community as defined by Swales, 1990) who share a common knowledge base and interests in similar phenomena. Specialized speeches are directed mainly at a primary audience, like the scientists, researchers, PhD students who attend the conference as delegates. But they also have a secondary, unintended, audience, that is the scientific journalists, writers and linguistic mediators involved in the event, such as translators and interpreters. The primary audience of specialized or scientific discourse is familiar with its textual structure and conceptual basis as they have prior knowledge of the terms and arguments presented by speakers, as well as of the textual clues that guide their comprehension processes (Goldman and Bisanz, 2002). However, most interpreters are not experts in all the specialized topics which they interpret throughout their professional career. This is why they are a secondary, non-expert, audience, for which the main features of scientific discourse, i.e. specialized terminology, dense and complex structures and a complex line of logic reasoning, hinder comprehension, since this secondary audience lacks the prior knowledge necessary to build a coherent and complete representation of the speech (Britton, 1994). Now, simultaneous interpreting is a highly complex task, which demands a high amount of cognitive resources, of memory and attention within a restricted temporal framework. In simultaneous interpreting, an interpreter has to understand a message delivered in one language whilst she is translating and producing a previous segment in a different language. As will be explained in detail later in this work, most theoretical models of the interpreting process recognize three macro-processes: a) comprehension of a segment of source speech; b) reformulation in target language; c) production of an equivalent segment in target language. Aside from these three macroprocesses, other executive operations of coordination and monitoring of output concur. The speed of production of the target speech is widely restricted by the speed of delivery of the source speech, so the interval of time allowed to conduct these operations is of merely a couple of seconds. (See a recent review in Christoffels and de Groot, 2005 and Padilla, Bajo and Macizo, 2007). Thus, the demands of comprehending, translating and producing a complex speech, about a topic with which they are not always familiar, in a very brief interval of time, results in interpreters having to work close to the limit of their cognitive capacity and therefore having to apply a set of strategies to avoid the interruption of the process and of interpretation. According to Gile¿s ¿Efforts¿ model¿ (1995/2009), interpreters must maintain a balance in the allocation of their limited cognitive resources in order to avoid the saturation of their processing system. Both practitioners and researchers agree on the relevance of advance preparation in the achievement of quality and satisfactory performance in simultaneous interpreting. The International Association of Conference Interpreters (AIIC), for instance, a leader in terms of professional standards and quality of performance in conference interpreters, highlights the importance of advance preparation by including it in the set of working conditions that make simultaneous interpreting possible and requires its members to demand conference documents in advance (AIIC, 2009) and to prepare them thoroughly for each assignment (AIIC, 2004). The prescriptive and descriptive literature on the professional aspects of interpretation, and simultaneous interpreting in particular reflect the same spirit. There is a consensus on the notion that interpreters are not specialists on every topic and that the topics that are interpreted vary constantly. For this reason, the literature presents methods and suggestions on how to exploit relevant documentation or, more recently, how to find relevant resources on the Internet, and extract specialized terminology and information that will support a more accurate interpretation. Most of these authors share the intuition that advance preparation might support anticipation or help accelerate the interpreting process (Seleskovitch, 1976; Moser, 1978; Gile, 2005). The literature also suggests how to prepare a glossary, which is a useful reference tool to consult while interpreting in the booth (see Seleskovitch, 1962; Gile, 1986; Moser-Mercer, 1992; Martin, 2002 y Donovan, 2001). However, on the empirical side only a few researchers have attempted to investigate the role of having background information for an interpreting task, or its effect on processing variables (like time-lag or response/reaction/production times) and performance (different measures of translation quality or accuracy). What is more, the few investigators that have shown an interest in this issue have faced huge methodological challenges (see, for instance, Anderson, 1979, 1994 and Lamberger-Felber, 2003). The results so far are not conclusive and even paradoxical, as those found by Griffin (1995) and Macizo and Bajo (2009), who observed longer reaction times in the conditions in which participants performed a translation task after studying relevant background information. For all these reasons, we believe that there is a research gap in Interpreting Studies, specifically in the research of the role of advance preparation. We argue that exploring the role of an instance of topic knowledge acquisition is of interest to Interpreting Studies, considering the role of prior knowledge in the comprehension of scientific discourse and the cognitive complexity of the simultaneous interpreting task. We also believe that it is relevant that research on this issue takes into account possible differences between experienced and inexperienced interpreters as this may shed light on the role of preparation at different stages of interpreting competence development. Empirical and experimental research on the acquisition and development of interpreting skills by deliberate practice also shows that professional interpreters' performance is qualitatively and quantitatively different from that of inexperienced interpreters or bilinguals and that these skills are susceptible of training and develop specifically to carry out the task of simultaneous interpreting (Christoffels, de Groot and Kroll, 2006; Yudes, 2010; Liu, 2008). Objectives, research questions and hypotheses The main objective of this study was to contribute to the empirical study of simultaneous interpreting by investigating the role of advance preparation in the simultaneous interpreting of scientific speeches. This subject has been addressed by previous researchers, but the results have been rather elusive as they have encountered important methodological obstacles. Thus, the main research question was: Is it possible to observe the effect of advance preparation in simultaneous interpreting? This question aims to test certain intuitive opinions that emerge from experience in professional and interpreter training practices and that are shared by several authors in Interpreting Studies literature. These intuitions, as Gile (2002:26) calls them, point towards a positive effect of preparation that might be reflected both in accelerated processing and in enhanced performance. However, testing these intuitions has not been an easy task. Simultaneous interpreting is a complex linguistic and human activity during which several variables come into play. In some previous studies, the lack of control over these variables has impeded the observation of clear effects (Anderson, 1979, 1994). In other cases, the high variability of the data did not allow for the extrapolation of the results to larger populations and statistical significance was not tested (Lamberger-Felber, 2003). Furthermore, in terms of knowledge acquisition, the study of speech manuscripts is not necessarily the same cognitive activity as the study of related conference materials and the elaboration of a glossary. In order to overcome these deficiencies, an experimental approach was taken and special attention was paid to controlling extraneous variables both in the design of materials, task, procedure and in the data analyses. We expected that this experimental setup would allow for the observation of measurable data and of a significant effect of advance preparation in processing and performance. However, the study of the role of prior topic knowledge in comprehension and simultaneous interpreting is quite a broad subject. In this research we have narrowed it down and restricted it to three specific research questions, objectives and hypotheses. As mentioned above, simultaneous interpreting is considered to be a complex task, not only in terms of cognitive processing, but also as a communicative activity. Hence the first specific research question is a two-fold research question: Does advance preparation have an effect on i) simultaneous interpreting processing; and ii) the accuracy of target speeches. Thus, the first specific objective of this research is to obtain a wide overview, a preliminary but comprehensive idea of the effect of advance preparation by measuring variables that have been traditionally studied in interpreting research, such as ear-voice span and target speech accuracy. Some researchers and scholars in Interpreting Studies have typically suggested that prior knowledge supports faster processing and improves performance and accuracy. Our first specific hypothesis is, then, that the acquisition of prior topic knowledge, by means of advance preparation, will be reflected in better processing and performance, as measured by ear-voice span, target speech accuracy scores and lexical-semantic content. Our second specific objective is to explore the role of advance preparation in overcoming difficult features of specialized source speeches. We considered that since coping with difficulties involves strategic problem-solving processing, it would be interesting to analyze the choice of reformulation strategies applied. So we set out to this task by measuring ear-voice span, target speech accuracy and the choice of reformulation strategies at segments of source speech classified as ¿neutral¿ and ¿difficult¿. Also, we further classified difficult segments according to the type of difficulty that presented: specialized terminology, complex syntactic structure and non-redundant information (proper names, figures, acronyms, etc.). Therefore, our second specific hypothesis is that advance preparation will support the interpreting process and performance in segments that contained difficult features of scientific speeches, such as terminology, complex syntactic structure and non-redundant information. We also expected that difficulty could have a detrimental effect on processing and performance, but hypothesized that preparation could probably contribute to mitigating this effect. Finally, due to the results obtained by the group of interpreting students, a second group of experienced interpreters was added to the design to learn more about the effect of preparation as a function of expertise in simultaneous interpreting. The specific research question in this regard was: Do experienced and inexperienced interpreters differ in the benefit they gain from advance preparation? Hence, the third objective of this research deals with investigating the role of expertise and its possible interaction with the acquisition of prior topic knowledge. Our hypotheses in this regard are less specific since, as reviewed in Chapter 3, there are aspects of simultaneous interpreting in which superior performance has been observed in experienced interpreters, when compared to novice interpreters (target speech accuracy, for instance). However, expertise research in interpreting is still in early stages of development, so any hypothesis in this regard needs to be taken with caution. Our third specific hypothesis is that experienced interpreters will show superior performance to students, as measured by target speech accuracy scores. We also expect that preparation has a positive effect on both groups, although we would not know to what extent. We further hypothesized that preparation could interact with expertise and perhaps bring students¿ performance closer to that of an experienced interpreter. Furthermore, it could also be expected that we find aspects of processing and performance in which both experienced and inexperienced interpreters behave similarly. Methods Participants. Twenty-three participants, 16 students of English-Spanish conference interpreting and 7 professional interpreters participated in the study. Student interpreters form the advanced students¿ group and professional interpreters, the professional interpreters¿ group. Both groups were homogenous in terms of linguistic proficiency in their L1 language skills and in their oral comprehension skills in L2. Differences in L2 reading, speaking and writing have been explained by the professional interpreters¿ age and the period of time they have been actively using L2, both of which are, as would be expected, significantly longer in the group of professional interpreters. There was a significant difference in the working memory span of both groups, - interpreters in the professional group had a significantly higher memory span than student interpreters. The students who participated in this study had a high record of academic performance, as rated by their own interpreting trainers at the Universidad de Granada and the Universidad de Las Américas, and are thus considered advanced interpreting students. Interpreters had, on average, 11 years of experience in professional interpreting. Participants also reported having little to no knowledge about the topics, ideas and terminology of the experimental source speeches, especially the students and no significant differences were found between the groups in their prior knowledge about the speeches' topic, ideas and terminology and the personal interest participants had about them. Materials. The main materials used in this research were the experimental speeches and the preparation materials. Six speeches were pre-selected from real life scientific seminars available on the internet. Transcriptions of these six pre-selected speeches were handed to a panel of independent judges (four Conference interpreting lecturers and professors at the Department of Translation and Interpreting, University of Granada) who were asked i) to assess, on a scale of 1¿7, the overall difficulty of each speech; and ii) to choose from the pool of six speeches the two that were more similar in degree of difficulty. Two pairs of speeches emerged with a very high correlation (r = .92), of which the pair with the highest degree of concordance among judges (Kendall W = 1) was selected as the two experimental source speeches. The judges were also asked to identify in these speeches the segments that they considered to be difficult to interpret without proper prior knowledge. Forty-six segments were identified as the "Difficult". According to the judges' explanations, these segments were further classified as having a lexical, syntactic or semantic difficulty. Each speech contained then 46 difficult segments and 34 neutral segments, or segments that should not entail a difficulty for participants considering their linguistic and interpreting competence. The topics of the experimental speeches were ¿The genetics of psychiatric disorders¿ and ¿Neuroscience of visual perception¿. The source speeches were later recorded by a native speaker of English at a mean delivery rate of 99.9 words per minute. The preparation materials included a 250-word summary of the speech; a CV or brief biography of the speakers (80-word); a slide presentation based on the speech (9 slides), a mock programme of the conference where the speech was supposedly presented (for communicative context) and a brief English-Spanish glossary that contained 30 specialised terms extracted from the corresponding source speech. The glossary contained the specialized term in English, its equivalent term in Spanish and, in some cases, additional information about the term or concept. These preparation materials were designed and written by the investigator, taking as a reference the set of documents and types of information that professional interpreters would realistically use when preparing for a conference in advance, based on her personal experience as a conference interpreter and as several authors in interpreting training literature suggest (see Chapter 4 and Moser-Mercer, 1992; Gile, 1995; Abril and Ortiz, 1998; Donovan, 2001; among others). Task and procedure. In the experimental task, participants had to simultaneously interpret two speeches from English into Spanish. One speech was to be interpreted with advance preparation, while the second speech had to be interpreted without advance preparation. The order of the experimental conditions (preparation) and of the presentation of the speeches was counterbalanced in four conditions (see Table 5.6). In the preparation condition, participants were first given 30 minutes to study the preparation materials. They were provided with the preparation materials and supplied with pens, markers and additional blank sheets, so they could take their own notes. The speech was interpreted immediately afterwards. In the non-preparation condition, participants were instructed to interpret without any preparation or knowledge about the topic of the speech. Both tasks were separated by a 15-minute break. Results and conclusions The results of the study show that advance preparation supported a more efficient processing and performance. Both the students and experienced interpreters who participated in our study obtained significantly better results when they had the opportunity to prepare relevant documentation for the interpretation task. The interpretation was more efficient since they produced significantly richer and more accurate target speeches, in terms of use of specialized terminology, an idiomatic rendition in the target language and a correct and complete rendition of the source speech meaning, while the time-lag allocated for this processing was significantly shorter. Advance preparation also had an effect on the choice of reformulation strategies. When participants prepared the task, their pattern of strategies was very homogenous. They consistently used the terminology provided in the glossary and applied the minimax strategy, or a minimum manipulation of the input. On the contrary, when participants did not prepare for the task, a host of remedial strategies appeared and both groups resorted more often to omission, paraphrasis, lexical and syntactic transcodification and generalization, especially when dealing with difficult segments. We were also able to observe how an increased level of difficulty affected processing and performance by slowing down ear-voice span and deteriorating the accuracy of the target segments. However, we could also observe that advance preparation played a significant role in mitigating these effects in the two experimental groups. Finally, our expertise approach allowed us to establish a set of aspects in which inexperienced and experienced interpreters behaved differently but also some in which they behaved similarly. The effect of expertise was significant in the superior performance that professional interpreters showed in the accuracy of the target speeches and in the frequency with which they chose the most efficient reformulation strategies, such as the use of direct target language equivalent and the minimax strategy. However, there were no significant differences in the ear-voice span of both groups, nor in some of the remedial strategies that they applied in the non-preparation condition or at neutral segments. Moreover, in the case of omissions, students omitted more segments than professional interpreters, but only in the non-preparation condition. Our third hypothesis dealt with the possible differences that we might find in the processing and performance of interpreting students and professional interpreters. We expected that the performance of professional interpreters would be superior to that of students, as measured by target speech accuracy scores. We also expected to observe differences in the other variables, although we would not know to what extent. We also hypothesized that there could be aspects in which experienced and inexperienced interpreters would behave similarly. Differences between experienced and inexperienced interpreters Our results showed that, as we expected, experienced interpreters had a superior performance in terms of the accuracy of their target speeches, which were significantly better. The target speeches produced by professional interpreters also had a significantly higher lexical-semantic similarity index, which indicates that they were more homogeneous and closer to the semantic representation of the scientific domain. These results are consistent with almost every previous study that has measured performance in students and professional interpreters (Liu, 2011; Liu et al., 2004; Hild, 2011, among others). Besides the superior performance in terms of accuracy and content of target speeches, we also found differences in the way dealt with difficult segments of speech. While when reformulating difficult segments that contained specialized terminology, complex syntactic structure and non-redundant items students resorted more often to remedial strategies such as paraphrasis, generalization and omission, professional interpreters maintained the same strategies of producing direct target language equivalents and the minimax strategies. These results indicate also a superior processing by professional interpreters, since these two strategies entail a minimum manipulation of the content of the segments and both correlated significantly with higher accuracy scores and shorter ear-voice span. This finding might explain why previous studies that did not used highly technical materials, like Anderson¿s (1994), failed to establish significant differences in the performance of experienced and inexperienced interpreters. Similarities between experienced and inexperienced interpreters As we also expected, there were aspects of the performance in which both students and professional interpreters behaved alike. Surprisingly, both groups of participants maintained a similar ear-voice span in the two conditions. As previous studies would suggest (Barik, 1973; Bajo et al., 2000; Christoffels et al, 2006) and considering their highest memory span and linguistic competence (see Chapter 5, Section 5.3.2-5.3.4), we would have expected professional interpreters to have a shorter ear-voice span than students, but both groups maintained an almost identical ear-voice span in the two conditions. Although memory span and linguistic competence was controlled for, a factorial analysis (which did not take those variables into account) did not find significant differences either. One possible explanation for this result is that the speech was delivered at a presentation rate comfortable for both groups, not too fast for students and not too slow for professional interpreters, so both groups maintained the time lag necessary to perform the cognitive operations according to their abilities. The comparative analysis of students and professional interpreters also allowed for the confirmation of a trend observed separately in both groups. Here, the ear-voice span of segments that contained proper names, figures and acronyms was significantly shorter than the ear-voice span that preceded specialized terms or embedded clauses. We have explained this observation, as Setton (1999) did, by arguing that since these are low-redundancy items, they are both difficult to incorporate into a mental representation and to retain in working memory. Since they can only be retained in working memory for a short time before they decay, both professionals and students adjusted their ear-voice span to quickly produce these items. We know that participants actually produced these items in the target language because this analysis does not take into account omitted segments. A correlation analysis together with the variation of ear-voice span according to an increased level of difficulty, indicates that ear-voice span also varies according to the type of input that interpreters are processing. Finally, we did not find any differences in the patterns of strategies applied to reformulate neutral segments. We defined these neutral segments as segments that could easily be translated considering the linguistic and interpreting competence of both groups of participants. Both students and professional interpreters used in equal proportion strategies such as direct target language equivalent, paraphrasis, generalization and minimax. This result again shows that superior performance of experienced interpreters may not be evident at general, non-technical materials, but instead at more difficult or complex input.