Findings of Interview Results - Shadowing Results

Chapter 4 Findings

4.1 Shadowing Results

4.2.2 Findings of Interview Results

Guided by the literature reviewed, this study categorized the participants' comments into the following categories:

(1) Intelligibility: The participants' comments on the relationship between

intelligibility and comprehensibility, and their subjective report on the phonological variations that interfered with intelligibility.

(2) Survival skills: Coping and comprehension strategies that the participants resorted to when shadowing Speaker S.

(3) Effort allocation/management: The participants' response to the question whether the allocation of effort in different processes differed when shadowing the accented speaker.

(4) Attitudes toward the task: Aspects of participants' use of language that reveals their attitudes toward the task of shadowing for the accented speaker.

(5) Pedagogical implications: The participants' response to the question on what they think about including foreign-accented speakers in interpreting classrooms.

The following paragraphs present professional and trainee interpreters' comments under each of the above categories in an attempt to answer Research Question 2:

What are the differences in professional and trainee interpreters' coping and processing strategies when listening to foreign-accented speech?

4.2.2.1 Intelligibility

4.2.2.1.1 The relationship between intelligibility and comprehensibility

An equal number of professional interpreters (P2, P3, P4, P6) and trainee

interpreters (T1, T5, T6, T10) commented that although the foreign accent has made it difficult to recognize every word in the source text, the comprehension nevertheless was not substantially impaired. Just like what T10 has said, "I probably cannot understand every single word he [Speaker S] said, but I am confident that I know the main message of his speech.", these participants felt that the fact that not every word in the foreign-accented input is intelligible does not lead to serious comprehension loss. Moreover, four professional interpreters (P2, P3, P6, P7) and one trainee

interpreter (T6) also commented that the challenge posed by Speaker S' accent can be conquered in interpreting, because "if I am interpreting for him, I would still know what he is talking about, only that I may miss some words" (P6).

In short, although all the participants suffered intelligibility degradation when listening to Speaker S, they regard their comprehension of the speaker not quite as much affected. Unfortunately, this claim cannot be verified with objective evidence due to the lack of a comprehension assessment in the data collection procedures.

Resorting to shadowing results to verify comprehensibility may also produce

misleading results. This is because, some participants, when referring to the omissions that they made in the shadowing task, reported that they eventually understood what the utterance was talking about, only that they did not have time to paraphrase it, or that the instruction to shadow without changing the words the speaker used made them refrain from paraphrasing. There were indeed instances where participants substituted a word they could not make out with another one that they deemed reasonable based on contextual information, or paraphrased sentences that they understood but could not make out or remember the exact words used. Nevertheless, as can be seen in Table 4.4 and 4.5, these instances were a few. In other words, the shadowing transcripts do not show that any participant attempted to do systematic paraphrasing. This indicates that these participants somehow did not or could not reflect their comprehension in the shadowing output. Finally, even if one tabulates all the subjective reports saying 'Yes, I understood this, only that I did not say it', or 'Yes, I came to understand this later' to correct some of the shadowing output that seem to show wrong comprehension, the fact the claim that 'Yes, I understood this' was made only after the participants were shown the script and had some time to read through the scripts may make these comments no less subjective than the current approach, that is, taking the general claim of

participants as what it is and assume that overall comprehension of the text was not seriously affected by the loss of intelligibility.

Table 4.4 Word-level substitutions

Participant Source text Shadowing output (number of

instances)

T2, T6, T14, P6

Conducive to Related to (3), reason of (1)

T2, T10 The job of... The goal of... (2)

T4, T12, P7 Germany Europe (3)

T4, T12, T13, P5

pleasant Decent (3), peace (1)

P4 psychologists Professionals in the field

P4 Paintings of garbage Paintings that are not worthwhile T10 a great impulse behind... the thinking behind...

Table 4.5 Paraphrasing at sentential level

Participant Source text Shadowing output

T2 What is going on inside anybody other than your own self

What (*) is going on (*) ahh emm within another person (*)¹¹

11 (*) represents a pause over 200 mms in the participants' shadowing output.

T14 The situation is essentially the same for Britain and also for West Germany.

The same (+)¹² situation can be found in Britain and Germany.

T2 Go on and live a creative life without worrying too much about becoming crazy

Having (*) a creative life (*) doesn't have to (*) worry about becoming crazy

P1 If people are happy according to their income relative to other people

if people (*) are (*) rated (*) by their income

P1 What are the chances of ...and therefore develop the insight into where in the brain creativity comes from

So (*) finding hmm the creativity in the brain (*) may be (*) difficult.

P4 What are the chances of ...and therefore develop the insight into where in the brain creativity comes from

and to understand their brains(*)and how they work in the realms of creativity(*)that would be very unlikely(*)

12 (+) represents an elongated vowel, which is a sign of hesitation or disfluency in speech production.

Nevertheless, the participants' claim that their comprehension was intact under the foreign-accented situation do seem plausible. First of all, existing research substantiates that comprehensibility of a passage is achievable without one hundred percent

intelligibility. In Field (2003), it was suggested that listeners' decoding failures can be compensated by resorting to contextual knowledge at the higher level. In the field of speech disorders, researchers have also found that listeners' comprehension of dysarthric patients of varying degree of severity are consistently higher than

intelligibility judgment (Hustad, 2008; Hustad & Beukelman, 2002). Hustad reasoned that this may be because when engaging in the comprehension task, the listeners' goals were to form a coherent picture of the utterance by actively drawing upon their world knowledge to process the unfolding passage. In short, it is possible to

comprehend the sense of a message that is limited in intelligibility by drawing the resources from the higher level of processing. Secondly, the participants did make these claims out of their own initiatives, not as a response to any specific questions that inquired on whether they found comprehensibility intact despite the lack of

intelligibility. In particular, most participants reported this at the very beginning of the retrospective interview, when they were asked to talk about what they felt in general during the shadowing task based on the notes they have jotted down in between the two shadowing sessions. Moreover, T1 actually summarized the content of the two

speeches at the beginning of the interview. Thirdly, as have been said earlier, there are several instances where participants substituted an unintelligible word with a word they generated based on the context. The occurrences of such kind of substitutions listed in Table 4.4 were all identified by the participants themselves in the

retrospective interview. Paraphrasing at sentential levels also occurred sporadically, with participants being able to explain correctly their comprehension of the source text of the paraphrased segment. For instance, for the last three examples in Table 4.5, P1 and P4 summarized their understanding of the source text in the retrospective interview, and explained that they only did paraphrasing because they came to understand the sentence at the latter part of its delivery and thus could no longer remember the exact wording.

4.2.2.1.2 Phonological variations that interfered with intelligibility

When asked if they have noticed any phonological deviations that caused

intelligibility difficulties when listening to Speaker S, P3, P5, P7, T1, T2, T6, and T13 mentioned both suprasegmental (i.e. pause patterns, intonation, lexical stress, prosody, etc.) and segmental problems (i.e. pronunciation, phonemic variation, etc.), P2, P4, P6, T4, and T15 only referred to segmental realizations, and T5 and T7 only referred to prosodic deviations that have troubled them. In other words, the majority of the

participants agreed that both suprasegmental and segmental deviations contributed to intelligibility difficulties. Just at P5 has nicely summarized it,

His pronunciations, actual phonological realizations. I am not very sure what words he was saying, and then there was no prosodic information to help me. I felt I recognized the words based mainly on his pronunciations, and used the prosody to help me. But his pronunciation was not accurate, and when you want to use prosody to help recognizing the words, the prosody was no use either. (P5)

In particular, suprasegmental deviations that were brought up include unnatural pause patters (e.g. "The pace of his speech is different from that of English speakers'.

He paused at odd places.", T6), unnatural intonation (P3), lack of prosodic stress (e.g.

"Every word was spoken with the same speed and 'weight'...I could not tell which words are more important.", T2), and incorrect primary stress (T13). Segmental deviations that were said to be troublesome include many references to Speaker S' realization of rhotic consonants, or R-like sounds (P2, P6, T4, T6, T13), and a few references to word-endings (P3, P4) and vowels (P4, T4).

Nevertheless, observed effects and subjective perceptions are different. Due to the lack of systematic phonological analysis of Speaker S that can enable the

connection between participants' shadowing errors to specific phonological deviations, it is impossible to verify if these phonological features that have been 'noticed' by the participants did cause decoding failures. While it may be possible to argue that these phonological features that have been identified by the participants were particularly problematic to decode, one cannot exclude the possibility that the phonological variations that have been pointed out are hurdles to intelligibility that they have conquered. All in all, due to the lack of observable evidence, it is beyond the scope of the current investigation to conclude whether the phonological features identified are indeed those that result in intelligibility difficulties.

4.2.2.2 Survival Skills 4.2.2.2.1 Coping strategies

Many of the participants across both groups seemed to agree that when

shadowing for Speaker S, it is better to adopt a longer EVS (P1, P2, P5, P6, P7; T1, T5, T6, T8). They also reported that a longer EVS allows them to comprehend the meaning of the utterance better before shadowing the words. T6, for instance, feel that this is a general strategy he would adopt when shadowing/interpreting for

foreign-accented speakers: "I feel that, for speakers with an accent, my strategy

should be not to follow the utterances too closely, and listen for the structure and main ideas in the sentence".

P4 and T10, on the other hand, adjusted their EVS in the opposite direction. T10 reported that she followed Speaker S more closely because "I was afraid that I would not be able to understand him". P4 also stated the same reason for closing in on EVS, stating that "This is more like a natural reaction. I followed the speaker more closely when I could not understand the passage."

As for what they could do when they failed to recognize the actual words that have been said by Speaker S, the participants' comments indicate that paraphrasing and omissions are equally eligible strategies disregarding whether they have

comprehended the message delivered. Unlike P1, P4, P7, and T12, who paraphrased the message when they could not recognize the actual words but could understand the meaning, P2 and P3 said they sometimes deliberately chose to omit those segments.

P3 specifically commented that she did it to "take some rest", but for some words that she did not recognize at first but came to realize what the meaning was later in the text,

"I would add something to explain the earlier part I have missed". Other occasions of omissions include not being able to recognize the words (P6, T13), wishing to focus on comprehension (T2), and preparing to start from a new sentence (T9).

From the participants' references to coping strategies, it is found that these participants seem to be using interpreting skills in the shadowing task. In a shadowing task, paraphrasing should not be an eligible choice, for the instruction was to repeat every word as exactly it is. Moreover, since the requirement is to repeat word by word, the flexibility in EVS should be rather limited by the number of words our working memory can contain, too.

One problem that may rise from the references to paraphrasing was: if

participants paraphrased extensively, it may affect the validity of the shadowing task.

However, the truth is, the occurrences of 'paraphrasing' at the sentential level in the target text of P1, P4, P7, and T12, who reported that they paraphrased sometimes, were very few (i.e. two instances for P1 and P4 respectively, and one instance for P7 and T12 respectively). The retrospective interview also found that most of the paraphrasing at the word level were done because the decoding of the substituted word failed, i.e., they could not make out what the word was based on the auditory input, so they substituted it with a word they deemed reasonable based on contextual information (e.g. Europe for Germany, not worthwhile for garbage, peace for pleasant).

Paraphrasing, omission, and EVS adjustment are all production related strategies.

The preceding two paragraphs have made clear that, due to the constraint of the

shadowing exercise, the participants' use of these production-related strategies were limited. Moreover, as have been said in Chapter 1, the scope of investigation in this study is the perception of foreign-accented English input in simultaneous mode.

Therefore, coping strategies related to the production stage of simultaneous interpreting cannot be dealt with in the current investigation.

4.2.2.2.2 Comprehension skills: Anticipation and contextual clues

These comments include references to specific comprehension skills employed both in the shadowing task in this study and interpreting for foreign-accented speakers in general. Judging from these comments, professional and trainee interpreters all acknowledged the importance of anticipation and contextual clues to assist the recognition of unintelligible input. Nevertheless, there is an inter-group difference in how successfully these comprehension skills were used in the listening process.

Most of the participants expressed that anticipation is a part of their working process in simultaneous mode. It can be shown in comments such as, "I am used to doing anticipation" (P6), "For the speech on government...it was easier to anticipate what he was going to say about this theme" (T9), or "Anticipation is very important"

(T13).

The picture of the participants' use of contextual knowledge is also very similar.

Professional and trainee interpreters alike think contextual information, including background knowledge of the theme and the message so far, helps comprehension when perfect intelligibility is impossible. For instance,

"Based on the preceding and succeeding text, I could comprehend what the speaker meant." (P3)

"At the beginning of a sentence, I only heard some voices. As I proceeded to the later part of the sentence, I may get more information, and then was able to recognize what words was used." (T15)

The participants agreed that context is an important factor that can compensate the loss of intelligibility. For example,

"For me, anticipation and background knowledge is more important [when interpreting for accented speakers]" (P6)

"Sometimes, comprehension relies on not only listening proficiencies, but also your own background knowledge, and whether or not you understand the subject matter" (T14)

However, contextual knowledge and anticipation are not fail-proof strategies.

First of all, when words that are impossible to anticipate or difficult to fit into context appear, interpreters from both groups failed to decode. For example, in the speech on creativity, the speaker made the following example to illustrate why it is difficult to understand the psychology of creativity.

Source text: However, what are the chances of ever being able to get a Mozart, stick him in the brain scanner at the moment of inspiration and scan the brain at that moment, and therefore develop the insight into where in the brain creativity comes from?

Of all the eleven participants (four professional and seven trainees) who shadowed this passage as read by Speaker S, only T4, T5, and T9 correctly recognized the word Mozart. Almost all of those who could not decode it were surprised when they read this word in script and said that they could not anticipate that Mozart the musician would appear here in this speech. This also had an impact on the decoding of the later part of this sentence. On average, out of this 44-word sentence, the participants failed to decode 24.63 words. T4, T5, T9, however, decoded on average 14.3 words wrongly while the others decoded on average 28.5 words wrongly.

Secondly, over-reliance on contextual knowledge may cause the participant to overlook the auditory input. For example,

"I have once heard a speech. It argued that the GDP is not the right way to measure happiness. So although I heard 'was the GDP', I said 'was not the GDP'. I listened some more and realized that he did say 'was'." (P2)

However, inter-group differences emerged when one looks at the overall

discourse in which the word 'anticipation' and 'contextual knowledge' was embedded.

When referring to anticipation, professional interpreters mostly used it to point out specific performance or comprehension collapse, such as the example about Mozart.

Also, there was no comments stating that the level of anticipation when shadowing Speaker S was higher than that for Speaker A. Trainee interpreters, nevertheless, refer to anticipation not only to account for specific intelligibility/comprehensibility

collapses, but also to describe the need to boost up their anticipation or that they have failed in general to anticipate the message better (T1, T2, T3, T5). For instance,

"A lot of my problems concerned anticipation....I tried my best to think about what he wanted to say, but sometimes I could not do it." (T1)

"I was unprepared the first time [I heard Speaker S], so I panicked. During the second time, I reminded myself to do more anticipation or analyzing." (T2)

The second inter-group difference concern trainee interpreters' sentiment toward the use of contextual knowledge. It seems that trainees felt that resorting to context when intelligibility falters is a kind of guesswork (T4, T7, T8, T12, and T13). For example,

"If you know a subject matter well enough, and have enough vocabularies, even if the speaker has a heavy accent, you can still guess what he means."

(T4)

"When the speaker is foreign-accented, I feel that I was not recognizing the words based on the phonological input, but based on the context, and there was a lot of guesswork." (T13).

Trainees' comments of failures in applying contextual knowledge also abounded.

"I told myself it was alright, understand the message first and then speak. I think this was a method I could try, but it did not mean I succeeded in doing it." (T8)

"During the second half of the speech, I wish I could use the information I have heard during the first half [delivered by Speaker A] to help me understand the message. But it did not work very well, because it was too demanding." (T14)

To summarize, professional and trainee interpreters all acknowledged the importance of anticipation and contextual knowledge in assisting successful comprehension/word recognition, despite the fact that sometimes, words and ideas that do not fit the interpreters' anticipation or contextual knowledge disrupt their performances. Nevertheless, trainee interpreters seem to need to deliberately increase the level of anticipation and use of contextual knowledge in assisting comprehension when shadowing Speaker S and sometimes could not succeed in their attempts. The language of 'guess work' in the use of contextual knowledge is also unique to the trainee group.

在文檔中外國口音英語辨識度與對應策略之認知差異：專業口譯員與學生口譯員對比研究 (頁 76-101)