Context-embedded phonological memory in interpreters
-
Hye-Yeon Chung
Abstract
This pilot study examines the hypothesis that the phonological memory of interpreters is superior to that of non-interpreters when embedded in a context. Context-embedded phonological memory (CEPM) is a product of the interaction between semantic and phonological memory, and becomes a component of interpreters’ professional expertise because of their attentional control. Four groups with different years of interpretation experience were asked to recall texts with distinct CEPM features. CEPM and its retrieval process were measured using WordSmith 6.0 and My Screen Recorder. Professional interpreters retrieved a high amount of CEPM (39.6 %), but were outperformed by interpreter trainees (40.52 %), roughly 15 years younger on average. Group difference was statistically significant only between the groups with interpretation experience and the teacher group. Only professionals’ performance in retrieving highly abstract information and the proper names and numbers in a context suggests that professionals are skilled at exploiting phonological information to create an elaborate semantic network.
1 Introduction
In the 1960s-1970s, Seleskovitch (1978) explained interpretation characteristics by using the concept of ‘sense’. According to this theory, interpretation is not a mere transposition of words but an act of delivering the sense, that is, the meaning of a text. Arguing that sense is both “conscious” and “non-verbal” in nature, she characterized it as “dissociated from any language form in cognitive memory” (Seleskovitch 1978:336).
When interpreting, retaining the language form can be important and even essential in some special situations, such as technical contexts. Seleskovitch may also have intended to stress the importance of the message and did not claim that the retention of form was entirely inappropriate or did not occur during interpretation (metaphor of raisin bread in Seleskovitch 1975).
Studies that compared source text (ST) and target text (TT) in interpretation revealed that experienced interpreters display greater accuracy than novices or non-interpreters (Barik 1975; Dillinger 1990; Liu 2001). The accuracy of TT is, to a certain extent, related to the retention (but not verbatim translation) of the language form of the ST. Although accuracy is not measured by the degree of literal translation, some of the interpreters’ strategies, such as paraphrasing (e. g., ‘it was a tradition’ → ‘traditionally’), are based on the presumption that the language form is retained as a phonological remnant in the memory. In this regard, well-developed phonological memory (PM)[1] may not be a sufficient condition for accuracy, but it is at least a necessary condition.
This indicates that interpreters remember the phonological information of the ST relatively well, but there is little evidence to support this idea in research on interpreters’ memory. Studies using digit and word span tests have often failed to provide solid evidence of the superiority of interpreters’ memory (not only simultaneous interpreters’ memory) (Section 2). As both digit and word span tests concern the memory of random words or numbers, they require a form-dependent strategy based on surface-level processing of language, whereas the memory of a text is created while constructing meaning. The results of these experiments suggest that professional interpreters’ PM alone cannot be considered superior to those of control groups.
This study is based on the hypothesis that professional interpreters’ PM is better than that of novices or non-interpreters when placed within a context. For convenience, this type of memory is hereafter referred to as ‘context-embedded phonological memory (CEPM).’ This study aims to build a theoretical construct for this hypothesis and test it empirically. Section 2 reviews the literature on interpreters’ memory from interpreting studies, psychology, and cognitive neuroscience to identify which aspects of memory will become part of the interpreters’ expertise. Section 3 draws on the information provided in Section 2 to establish the concept of CEPM. Finally, Sections 4, 5, and 6 will show and discuss the results of an empirical study on CEPM.
2 Interpreters’ memory as a component of expertise
2.1 Psychological aspects of interpreters’ memory
A memory created during interpretation can be categorized as working memory (WM), given its nature as temporary storage for immediate use and its capacity and time limitations. For this reason, WM has been regarded as contributing to interpreters’ expertise (Pöchhacker 2016:110–112). In the early research phase, studies focused on memory storage, particularly on measuring WM capacity. They mainly used a digit span test (Padilla 1995; Chincotta/Underwood 1998; Nordet/Voegtlin 1998) or a listening span test (Liu et al. 2004; Köpke/Nespoulous 2006) to compare professionals with control groups. These are reliable tests to measure WM capacity, but they might not be the most suitable means to reveal the expertise of professional interpreters, who are adept at delivering meaning. During the span tests, subjects have to exploit a rehearsal strategy rather than constructed meaning since the numbers or words in the tests are semantically unrelated. The auditory inputs the subjects receive circulate in the phonological loop of the WM system (Baddeley/Hitch 1974; Baddeley 2000), helping them to retain the phonological information. Alternatively, subjects could chunk items into meaningful units, in which case the semantic information of the word is forced into a relationship with that of other words. Neither rehearsal nor chunking strategy is based on the same mental process that interpreters experience while working with coherent texts. In the process of interpretation, salient phonological information is put together and turned into semantic memory (SM) without much effort (Darò/Fabbro 1994).
This might be the reason why experiments with these span tests provided inconsistent results in the early phase of research. Padilla (1995), Chincotta/Underwood (1998) and Liu et al. (2004) reported better memory in interpreters than in the control groups but the group difference was not always statistically significant. Nordet/Voegtlin (1998) and Köpke/Nespoulous (2006) found that professional interpreters exhibited less memory capacity than student interpreters.
Since the findings from digit and listening span studies in the early phase have proven contradictory, researchers turned to tasks involving semantic analysis. Bajo et al. (2000) applied semantic categorization to demonstrate that interpreters’ SM was greater than that of non-interpreters in the task. Liu et al. (2004) showed that the superiority of interpreters’ memory lies in their semantic competence by verifying that interpreters are better at identifying and interpreting critical sentences within a text than are control groups.
This inconsistency in the results in the early phase could be attributed to the difference in experimental designs (e. g., tasks and subjects). Evidence found in more recent studies that exploit span tests and semantic tasks (Stavrakaki et al. 2012; Hiltunen et al. 2016) supports stronger WM for, at least, simultaneous interpreters (Mellinger/Hanson 2019; Bae/Jeong 2021).
Apart from the above-mentioned aspects of memory, executive or attentional control has also been examined as a critical aspect of interpreters’ memory processes. Cowan (1988) and Conway/Engle (1994) argued early that the superiority of interpreters’ WM lies more in attentional control than in memory capacity. Stating that WM is an activated element of long-term memory (focus of attention (FOA)), Cowan 2000) suggested that FOA capacity may be associated with an individual’s potential to become an interpreter. Meanwhile, Moser-Mercer et al. (2000) and Timarová et al. (2014) proposed that interpreters specialise in the competence to suppress irrelevant information among diverse aspects of attention. Moser-Mercer et al. (2000) found that professional interpreters demonstrated significant proficiency in reading tests with delayed auditory feedback. The authors concluded that this result indicates that professional interpreters’ attentional control was not deterred by distracting information. This finding was corroborated by Stavrakaki et al. (2012), who showed that this kind of competence was found among those with high verbal fluency. Timarová et al. (2014) reported similar results. Assuming that the superiority of WM among interpreters stems from a discrepancy in controlled attention, they demonstrated that professional interpreters were better able to suppress interfering distractors than the control groups. Swift shifting between tasks is another feature of professional interpreters. In Liu et al. (2004), professional interpreters were better able to quickly switch their attention to continuation sentences directly following critical sentences than interpretation students. As suggested by Köpke/Nespoulous (2006), interpreters equipped with this attention control competence are better able to perform complex tasks than non-interpreters. They used free recall with articulatory suppression and category probe tasks and concluded that the more complex the task, the better interpreters performed compared to the control groups, ascribing this result to their ability to shift their attention to tasks requiring greater attention. The advantage of professional interpreters in executive control has been reconfirmed in recent studies, such as Injoque-Ricle et al. (2015) and Hiltunen et al. (2016).
2.2 Neurocognitive aspects of interpreters’ memory
These findings were further supported by neurocognitive research. To remember phonological information, people commonly resort to articulatory rehearsal. In this case, memory is maintained while sound remains in the phonological loop of WM (regions within the left front-orbital area and the left temporoparietal junction, Aboitiz et al. 2010). This type of memory, that is, verbatim memory of verbal information or PM, seems to be more affected by biological conditions[2] such as age (Nittrouer et al. 2016) than SM, whereas the competence to retain SM is more likely to be improved by extrinsic factors, such as training (Miotto et al. 2020).
This distinguishing factor may also be valid for interpreters. Younger and less-trained interpreters score higher on simple non-word repetition than older interpreters (Signorelli et al. 2012). Experienced interpreters have demonstrated their superiority over younger students, mostly in semantic tasks such as reading and speaking span tests (Christoffels et al. 2006). In this regard, studies on interpreters’ WM that required subjects to memorize random items may have been disadvantageous for professional interpreters, who are usually older than the control groups.
However, the distinction between PM and SM cannot be so clear regarding the process of memory creation. A large number of brain regions subserving the PM overlap with those involved in creating the SM, which includes the junction between the parietal and temporal lobes (Binder et al. 2009). This suggests that SM is created by actively exploiting phonological input. However, not all phonological inputs reach the level of semantic abstraction, but only those with salient features and attract listeners’ attention. Binder/Desai (2011:21) proposed that the extent of abstraction during semantic tasks varies with “concept familiarity, demand for perceptual information, and degree of contextual support.” This means that the regions subserving the PM also contribute to the formation of SM in interaction with long-term memory, as well as with the attentional system of the brain. This interplay between PM-SM and the attentional system could occur most actively during interpretation and lead to its automatization.[3]García (2019:180), who extensively reviewed neurocognitive research on translation and interpreting, also suggested that the network of simultaneous interpreters’ cognitive control comprises “several hubs of frontostriatal” (attention, automatization) and “perisylvian circuits” (phonological and semantic WM).
One region engaged in creating SM is the inferior parietal lobule (IPL), in which visual and auditory inputs are integrated into a semantic representation. Especially the right IPL plays an important role also in “maintaining attention on the current task and responding [to] salient new information” (Singh-Curry/Husain 2009:1434–1435). Changes in the functions and structures of the IPL in interpreters’ brains could indicate that interpreting experience significantly influences interpreters’ SM. In their fMRI study, Elmer et al. (2011) found that professional interpreters showed more responsiveness in the left IPL than in the bilingual control group. Elmer et al. (2014) and Hervais-Adelman et al. (2011) used fMRI and discovered that the gray matter volume[4] of the left supramarginal gyrus (SMG), which is the lower part of the IPL, changes according to the hours of interpreting training and experience. Similarly, Hervais-Adelman et al. (2017), who conducted a longitudinal fMRI study on 34 interpreter trainees, discovered that the cortical thickness of the left SMG and right IPL increased after 15 months of interpreting training. More recently, Hervais-Adelman/Babcock (2019) mentioned in their review article that IPL is one of the regions that have been repeatedly implicated in studies on interpreters’ brains.
Hervais-Adelman et al. (2011) and Elmer et al. (2014) also provided results that support the superiority of attentional control among professional interpreters. One of the most relevant brain regions for attentional control is the prefrontal cortex (PFC), which is responsible for conflict resolution and response selection (Hervais-Adelman et al. 2015). The other region is the anterior cingulate cortex (ACC), which plays a key role in monitoring language use, allocating attention, conflict monitoring, error detection (Abutalebi/Green 2007; Hernandez 2009), and suppressing automated reactions (Banich/Compton 2011). Elmer et al. (2014) and Hervais-Adelman et al. (2011) found changes in gray matter volume in the left rostral ACC of interpretative students after interpreting training and experience.
The findings discussed above lead us to conclude that the superiority of memory among professional interpreters seems to lie in creating SM that exploits the relevant phonological information and their attentional competence, rather than in the capacity of PM.
3 Interpreters’ CEPM as a component of expertise
Section 2 shows that interpreters’ memory expertise lies in semantic and attentional competence rather than phonological competence. One could ask how this is related to the hypothesis of this study that interpreters have better phonological memory.
Professional interpreters learn to shift or divide attention strategically during interpreting training. As interpreting is a multitasking process that consists of several subtasks, interpreters must shift or divide their attention between tasks to perform successfully. They do this in several ways. They allocate their attention to subtasks such as comprehension, production, and memory (Gile 2009). Their attention devoted to comprehension and memory can be further divided into more- and less-important information. As seen by Liu et al. (2004), professional interpreters can identify essential information as a critical factor in the meaning-making process and devote strategic attention to it. Since essential information attracts interpreters’ attention, its semantic and even phonological features could profit from this advantage. Phonological information during interpretation is often strategically suppressed (Moser-Mercer et al. 2000; Timarová et al. 2014) to avoid language interference. However, assuming that they have developed a WM with a broader FOA[5], professional interpreters' attention might extend even to phonological information and retain it in memory. In particular, these phonological inputs, which contribute significantly to creating SM, are likely candidates for PM since interpreters strategically allocate their attention to key information. As interpreters’ attentional load for SM gradually lessens, surplus attention can, albeit unintentionally, be directed to the phonological information surrounding semantically meaningful information.
For this reason, information immediately following essential information can often be remembered, as Liu et al. (2004) showed in the task of continuing sentences following critical sentences. As a result of the overall process, the PM of the entire text (i. e., general CEPM) can increase for professional interpreters.[6] Once the general amount of CEPM rises, text can be more elaborately structured and accurately retrieved based on the construction of CEPM so that one can provide more stable memory and more accurate interpretation.
4 Experiment
This experiment was designed to examine the hypothesis that professional interpreters’ CEPM is better than that of non-interpreters. This hypothesis is based on the concept of CEPM explained above and is subdivided into the following hypotheses:
Hypothesis 1: CEPM is influenced by interpreting experience.
Hypothesis 2: The superiority of interpreters’ CEPM is more distinct for essential information than secondary information.
Hypothesis 3: Interpreters’ CEPM results from the attentional shift/division between the PM and SM.
Participants
A total of 40 subjects, ten from each group, were recruited based on their years of interpretation training, experience, and age. Interpreter trainees were second-year students from the Graduate School of Interpretation and Translation (GSIT) and professional interpreters, graduates from GSIT had, on average, 11.5 years of professional experience in consecutive and simultaneous interpreting. Undergraduates were selected from the same university and faculty and had relatively homogeneous general cognitive abilities. Teachers who all worked in the same school were included as a comparative group for professional interpreters since both are regarded as similar in terms of age and linguistic ability. All the subjects agreed to participate in the experiment and received a small monetary reward for participation.
Subjects
| Group | Undergraduates | Interpreter trainees | School teachers | Professional interpreters |
| Mean age | 20.7 | 25.8 | 40.8 | 39.6 |
| Gender | M:F=1:9 | M:F=4:6 | F=10 | M:F=1:9 |
| Background | Humanities majors |
1 year of interpreter training | Teachers of humanities | 11.5 years of experience |
Materials
Four different types of text were selected based on their features in terms of CEPM. Instead of word or listening span tests using random words, we chose texts of adequate length (85–90 words per text) to allow the creation of a semantic structure, but not verbatim memory. Text 1, for example, was excerpted from a novel that describes scenery, and Text 3, a blog excerpt, consists of sentences with abstract content. The descriptive character of Text 1 should facilitate visual association of the scene, whereas the abstractness of Text 3 should make it challenging to create a semantic structure. Both texts were selected to ascertain the extent to which the PM contributed to SM creation. Texts 2 and 4 contain proper nouns, technical terms, and word lists. These texts should help reveal whether professionals are actually able to shift or divide their attention between PM and SM. All texts were presented in Korean, the mother language for all subjects, to minimize the effect of language proficiency.
Apart from the proportion of PM and SM in general, that of kernel words (PM) and kernel propositions (SM) are of special interest, because they are crucial for estimating whether essential information is remembered in its original form, and whether this result can be ascribed to the effective shift/division of attention. Seven professional interpreters with five or more years of experience participated in the selection of kernel words and propositions (not the subjects of the experiment). They were asked to mark the words and propositions they considered to contain essential information within each text. Those selected by four or more interpreters (measured using WordList of WordSmith 6.0) were deemed kernel words and propositions.
Materials
| Text 1 | Text 2 | Text 3 | Text 4 | |
| Text typeProposition / word (total) Proposition / word (kernel) |
Novel19 / 85 11 / 31 |
News article20 / 90 11 / 38 |
Blog post20 / 87 10 / 41 |
Speech15 / 90 10 / 44 |
Procedure
Subjects were asked to complete a form with their gender, age, and interpretation experience. During the experiment, they listened to each of the four texts (presented in Korean) to the end and then typed what they could recall afterward into a computer in the original language, Korean. The texts were recorded and read at an average speed of 100 wpm. The subjects’ computer activity (typing, hesitation, corrections, etc.) was recorded using My Screen Recorder (MSR).
Measurement
The unit for measuring PM is a word[7], and that for SM is a proposition, consisting of a subject and predicate. The results, that is, the percentage of words in the TT (memory output) matching those in the ST, were assessed using the Concordance of WordSmith 6.0. This tool allows us to measure how many exact matchs (in this case, words) ST and TT have, in which part of the text, and how often. Figure 1 below shows us that professional 6 recalled 36 out of 85 words (total words of Text 1) from Text 1 with the exact wording.
ST-TT Concordance of Text 1, Professional 6 (WordSmith 6.0)
Not only the product but also the process of retrieval can give us insight into the subjects’ memory. An MSR video was examined to observe the CEPM retrieval process.
Statistical Analysis
What intrigues us in this experiment is the difference between professional interpreters and three control groups. Because we have more than two groups in our experiment, the group difference was mostly assessed by ANOVA (analysis of variance). If the results did not turn out to be normally distributed for one or more groups (Shapiro-Wilk’s test), ANOVA was substituted by the Kruskal-Wallis test and the p-value in the post-hoc test was corrected by Bonferroni correction. In the comparison of two groups (e. g., between essential and secondary information), a t-test (difference in means between two populations) was used.
We are interested in analysing the relationships between variables as well, e. g., between CEPM and interpretation experience or between CEPM and age. In this case, we exploited the Pearson correlation coefficient, which was used in the comparison between SM and PM as well.
5 Results
5.1 Hypothesis 1: influence of interpreting experience
It was not the professional interpreters but the interpreter trainees who remembered the greatest number of words, and the professionals were the second best. The only text in which professionals outperformed trainees was Text 3. As for the group difference, F-value (ANOVA) (F(3,36) = 3.499*, p = .025) was substituted by Kruskal-Wallis H, as the overall CEPM ratios were normally distributed for undergraduates (p = .614) and professionals (p = .554) and nearly for interpreter trainees (p = .043) but not school teachers (p = .002) (Shapiro-Wilk's test). A Kruskal-Wallis test indicated that the overall CEPM ratios significantly differed over groups (H(3) = 11.393*, p = .010). Group difference was significant between professionals (p = .042) and school teachers and between trainees and school teachers (p = .035) (Bonferroni correction).
Overall CEPM (%)
| Undergraduates | Interpreter trainees | School teachers | Professional interpreters | |
| T1 | 27.76 | 37.06 | 27.88 | 36.82 |
| T2 | 32.33 | 41.33 | 27.00 | 36.11 |
| T3 | 29.08 | 34.25 | 28.05 | 37.82 |
| T4 | 33.89 | 49.44 | 33.22 | 47.67 |
| M | 30.77 | 40.52 | 29.04 | 39.60 |
| SD | 10.25 | 12.30 | 16.05 | 9.77 |
| MR | 16.60 | 26.70 | 12.30 | 26.40 |
| MR: mean rank | ||||
Overall CEPM (Kruskal-Wallis test)
| Kruskal-Wallis H | df | Asymp.Sig. | ||
| total | 11.393* | 3 | .010 | |
| * p < .05 ** p < .01 ***p < .001 | ||||
Professionals and trainees were additionally compared to examine the influence of interpretation experience and age on CEPM and the interactions among them. CEPM (overall) correlates positively, albeit with no statistical significance, with interpretation experience (r = .150, p = .528) and negatively with age (r = -.138, p = .560). The interaction effect between age and interpretation experience showed no statistically significant differences (age × experience × CEPM: F(1,18) = .338, p = .719).
During the analysis of the MSR video, professional interpreters showed salient behaviors concerning CEPM. They made seemingly unnecessary corrections, even when they made little difference in meaning.[8] They also reproduced the greatest number of words, presented them in their original order (Figure 1), and were least influenced by the recency effect (Chung 2018). Although less frequent than professional interpreters, similar cases were found within the interpreter trainee group. The number of corrected words exactly matching those in the ST was seven for professional interpreters, five for interpreter trainees, one for undergraduate students, and zero for teachers.
5.2 Hypothesis 2: essential information and secondary information
Here again, interpreter trainees had the highest scores. However, compared to the overall CEPM, the difference between interpreter trainees and professionals shrank, and professionals obtained scores as high as those of trainees in Text 1 and higher scores in Text 3. Professionals showed stable performance across the texts, but group differences were significant for none (Dunnett test).
CEPM of essential information (%)
| Undergraduates | Interpreter trainees | School teachers | Professional interpreters | |
| T1 | 27.42 | 39.03 | 31.94 | 39.03 |
| T2 | 33.68 | 36.05 | 22.63 | 31.58 |
| T3 | 33.17 | 36.10 | 27.56 | 39.27 |
| T4 | 32.95 | 47.73 | 37.27 | 46.82 |
| M | 31.81 | 39.73 | 29.85 | 39.17 |
| SD | 10.69 | 13.09 | 14.94 | 12.21 |
CEPM of essential information (ANOVA)
| M | SD | F | P | Ƞp2 | |
| total | 35.14 | 12.73 | 2.803 | .054 | .189 |
| * p < .05 ** p < .01 ***p < .001 | |||||
For secondary information, the result was similar to that of the overall CEPM. Professional interpreters performed the best only in Text 3, and interpreter trainees topped the scores in others. A Kruskal-Wallis H indicated that the CEPM of secondary information significantly differed over groups (H(3) = 12.016**, p = .007).[9] Group difference was significant between trainees and school teachers (p = .010) (Bonferroni correction).
CEPM of secondary information (%)
| Undergraduates | Interpreter trainees | School teachers | Professional interpreters | |
| T1 | 26.98 | 34.72 | 25.28 | 34.53 |
| T2 | 28.57 | 32.04 | 21.84 | 28.57 |
| T3 | 16.98 | 23.49 | 17.91 | 29.53 |
| T4 | 35.68 | 48.92 | 24.86 | 42.43 |
| M | 27.05 | 34.79 | 22.47 | 33.77 |
| SD | 11.68 | 13.99 | 16.64 | 9.65 |
| MR | 17.80 | 27.90 | 11.40 | 24.90 |
| MR: mean rank | ||||
CEPM of secondary information (Kruskal-Wallis test)
| Kruskal-Wallis H | df | Asymp.Sig. | ||
| total | 12.016** | 3 | .007 | |
| * p < .05 ** p < .01 ***p < .001 | ||||
The difference between the essential and secondary information was statistically significant for professionals (t = 2.000*, p = .049) and teachers (t = 2.047*, p = .044), but not for trainees (t = 1.453, p = 0.150) and undergraduates (t = 1.704, p = .092).
5.3 Hypothesis 3: attentional shift/division between PM and SM
To determine whether Tables 3, 5, and 7 are indeed the results of the interaction between PM and SM and not simply PM, the rate of SM should be estimated as well. The following shows how well each group remembered the meaning of the kernel propositions (SM).
SM of essential information (%)
| Undergraduates | Interpreter trainees | School teachers | Professional interpreters | |
| T1 | 40.91 | 63.64 | 46.36 | 63.64 |
| T2 | 74.55 | 73.64 | 46.36 | 66.36 |
| T3 | 42.00 | 53.00 | 32.00 | 64.00 |
| T4 | 58.00 | 75.00 | 52.00 | 76.00 |
| M | 53.86 | 66.32 | 44.18 | 67.50 |
| SD | 18.77 | 19.72 | 18.50 | 17.51 |
| MR | 16.25 | 27.2 | 11.8 | 26.75 |
| MR: mean rank | ||||
SM of essential information (Kruskal-Wallis test)
| Kruskal-Wallis H | df | Asymp.Sig. | ||
| total | 13.006** | 3 | .005 | |
| * p < .05 ** p < .01 ***p < .001 | ||||
Professional interpreters demonstrated the best performance in retrieving SM. Group differences were greater and more evidently significant compared to CEPM (H(3) = 13.006**, p =.005)[10]. Both professionals (p = .025) and interpreter trainees (p = .019) exhibited the significant difference in comparison with school teachers (Bonferroni correction). For the comparison of texts, the group difference was the highest in Text 3 (H(3) = 12.151**, p = .007), the most difficult text to be semantically structured.
For the correlation between PM and SM, teachers (r = .884**, p = .001) showed the highest score, probably because they retrieved the smallest number of propositions and words, closely followed by professional interpreters (r = .712*, p = .021). Interpreter trainees (r = .683*, p = .030) and undergraduates (r = .549, p = .100) were third and last, respectively.
The high PM-SM correlation above might be a reliable indicator for effective shift/division of attention, but it does not show an immediate relationship between an individual PM and the SM containing it. More definite evidence is needed to confirm that professional interpreters shift or divide their attention between PM and SM. Numbers, appropriate nouns/technical terms, and word lists can serve this purpose. PM for the lexicons in Texts 2 and 4 were measured in the following categories: numbers, proper nouns/technical terms, and word lists.
| Exp. 1) A study by Yale University in the United States surveyed 38,000 doctors and 80,000 nurses.– SM: The whole sentence– PM: proper nouns and numbers[11] |
CEPM of numbers, proper nouns, and word lists (%)
| Undergraduates | Interpreter trainees | School teachers | Professional interpreters | |
| T2 | 50.32 | 63.33 | 43.93 | 62.26 |
| T4 | 70.60 | 88.69 | 76.55 | 92.62 |
| M | 60.46 | 76.01 | 60.24 | 77.44 |
| SD | 24.04 | 14.53 | 21.54 | 13.30 |
CEPM of numbers, proper nouns, and word lists. (ANOVA)
| M | SD | F | p | Ƞp2 | |
| total | 68.1 | 16.36 | 4.863** | .006 | .288 |
| * p < .05 ** p < .01 ***p < .001 | |||||
Professional interpreters had the highest score in almost every category (except for proper nouns/technical terms in Text 2). The group differences in total CEPM were statistically significant (Table 12).
Correctly retrieved PM and SM in an individual unit (ANOVA)
| M | SD | F | p | Ƞp2 | |
| PM and SM (○) | 67.87 | 24.78 | 3.837* | .018 | .242 |
| PM or SM (∆) | 27.18 | 20.41 | 2.503 | .075 | .173 |
| None (×) | 5.58 | 13.48 | 1.772 | .17 | .129 |
| * p < .05 ** p < .01 ***p < .001 | |||||
Meanwhile, the measurement of the propositions (SM) including numbers (PM), yielded the following results: First, interpreter trainees best remember both information types (SM, PM) simultaneously and correctly (marked with ○ in Table 13), closely followed by professional interpreters. Second, the failure rate was the lowest among the professional interpreters. None of the professionals omitted both types of information concurrently (∆ Table 13). The interpreter trainees came second in terms of failure rate. The group with the highest rate of missing phonological or semantic information (× Table 13) was undergraduate students.
| Exp. 2) A study by Yale University in the United States surveyed 38,000 doctors and 80,000 nurses.– Both SM and PM correctly remembered: ○– SM or PM correctly remembered: ∆– None of SM and PM correctly remembered: × |
6 Discussion
The results of the overall CEPM do not support Hypothesis 1. Professionals who have the longest interpreting experience are outperformed by interpreter trainees, who are approximately 15 years younger than the professionals (Tables 3 and 4). Group differences of both groups were only significant in comparison with school teachers. Neither interpretation experience nor age showed a significant correlation with CEPM, supposedly because of a small sample size.
Another way to test Hypothesis 1 is to use CEPM-related features of individual texts. As mentioned above, Text 3 consisted of highly abstract information and is, therefore, the most difficult for semantic structuring and difficult to memorize without using a phonological strategy. The fact that the professionals yielded the highest CEPM in Text 3 (Tables 3, 5, 7, and 9) might indicate that they are indeed skilled at exploiting PM to create SM.
This result is consistent with the results of the MSR video analysis. Professionals’ successful attempts to correct TT to maintain the exact wording of the original text provide evidence that professionals are more attentive to phonological information when it is embedded in the context. This and the previous outcome could be interpreted in favor of Hypothesis 1; however, this result should be corroborated by statistical evidence in further studies with a larger sample.
The results for Hypothesis 2 diverge from the assumption. It is expected that professionals would perform better in CEPM of essential information than in that of secondary information even in comparison to three control groups because CEPM contributes to creating meaning, which is shown to be the professionals’ expertise. However, interpreter trainees performed better than professionals in CEPM for both types of information and the group differences among four groups were not significant (Tables 6 and 8). As for the difference between essential and secondary information, the only group with a significant result was that of professionals. This means that the professionals, albeit with a less good performance, paid their attention strategically to essential information much more than secondary one.
Finally, one could ask how to tell if the CEPM shown above is actually PM contributing to the formation of SM and not just PM. To answer this question, we assessed and analysed the PM and SM. Since this study assumed that CEPM is created through effective attentional shift/division (Hypothesis 3), both PM and SM should be retrieved successfully. According to previous studies, creating and retrieving SM is one of the strengths of professionals. This result was confirmed in this pilot study (Tables 9 and 10). In retrieving the SM of essential information, in which the professionals performed the best, the group difference was the highest in the most difficult text (Text 3). Professionals may have exploited their stable CEPM to create a more elaborate semantic network. Non-interpreters (undergraduates and teachers) who produced relatively low CEPM could have experienced difficulties in creating SM based on the text as abstract as in Text 3.
The question is then whether professionals’ competence in creating SM can be extended to CEPM since we assumed that their focus of attention might have expanded during their interpreting training and that they are more able to divide their attention. There are two ways to determine whether the subjects actually shifted or divided their attention with success. First, we can calculate the correlation between PM-SM, and second, we can analyse propositions containing proper nouns and numbers. The results showed that professionals performed the best in remembering numbers, etc., in context (Tables 11 and 12) and that they were the second-best in the correlation between PM and SM and the concurrent retrieval of both (Table 13). This result indicates that professional interpreters are inclined to pay attention to SM and PM concurrently, and can divide their attention to both information types. The results for Hypothesis 3 were the clearest among the three hypotheses because they reached a respectable level of statistical significance and effect size (Tables 10 and 12), which suggests that professional interpreters’ expertise lies in SM and, although it is less clear, also in CEPM, especially in that of the words they are trained for (such as numbers).
7 Conclusion
This paper examines tentatively the hypothesis that interpreters’ phonological memory (PM) is superior to that of non-interpreters when embedded in a context (CEPM). CEPM is a product of the interaction between PM-SM and the attentional system and can account for interpreters’ accurate rendition of ST.
PM competence per se could not be regarded as part of the interpreters’ memory expertise, as demonstrated by several digits and word span tests with (in many cases trainee) interpreters. Their expertise lies in semantic and attentional competence. However, with increasing semantic and attentional ability, their PM can also be improved as they gain sufficient cognitive resources to strategically shift or divide their attention between PM and SM. This is even more so for PM types that contribute to creating SM. This strategic shift/division of attention might be more distinct for essential information than for secondary information since interpreters pay more attention to the former. Based on this theoretical framework, three hypotheses are proposed and tested, with the following results.
(1) The professional interpreter group performed well in the overall CEPM but was outperformed by the interpreter trainees, who were 15 years younger on average. However, analysing the memory retrieval process via MSR videos revealed that the professionals were attentive to phonological information in the context.
(2) For the CEPM of essential information, the interpreter trainees again topped the scores, followed by professional interpreters, who came in second place. But the difference between essential and secondary information was statistically significant only for professionals.
(3) For the concurrent retrieval of PM and SM (memory of sentences with numbers and proper nouns), professionals yielded the second-best result in that case as well. The rate of correctly retrieved PM-SM was the highest among interpreter trainees, and that of failure was the lowest among professionals.
In sum, professional interpreters[12] demonstrated not the best but a reasonably high CEPM throughout all the categories. Their superiority in retrieving highly abstract information (Text 3) and in creating CEPM of numbers, proper names, and PM-SM correlation might indicate that they are indeed skilled at exploiting PM to create an elaborate network of SM.
However, in most cases, the interpreter trainees were the group with the best performance,[13]perhaps because this group fulfills two advantageous conditions for high CEPM: youth and interpretation experience.
These results should be corroborated by more definitive statistical evidence since the hypotheses of this study were tested on a small sample. In particular, the precise mechanism of the interplay between SM and PM in creating CEPM might be important in improving the accuracy of interpretation and should therefore be investigated in future research.
Funding
This work was supported by the Hankuk University of Foreign Studies Research Fund (of 2022).
References
Aboitiz, Francisco/Aboitiz, Sebastian/García, Ricardo R. (2010): The phonological loop: A key innovation in human evolution. Current Anthropology 51(1), 55–65. doi:https://doi.org/10.1086/650525.Search in Google Scholar
Abutalebi, Jubin/Green, David (2007): Bilingual language production: The neurocognition of language representation and control. Journal of Neurolinguistics 20, 242–275. doi:10.1016/j.jneuroling.2006.10.003.10.1016/j.jneuroling.2006.10.003Search in Google Scholar
Baddeley, Alan D./Hitch, Graham (1974): Working memory. In: Bower, G.A. (Ed.): The psychology of learning and motivation. New York: Academic Press. 48–79. doi.org/10.1016/S0079-7421(08)60452-1.10.1016/S0079-7421(08)60452-1Search in Google Scholar
Baddeley, Alan (2000): The episodic buffer: a new component of working memory? Trends in cognitive sciences 4, 417–423. doi.org/10.1016/S1364-6613(00)01538-2.10.1016/S1364-6613(00)01538-2Search in Google Scholar
Bae, Mun Jung/Jeong, Cheol Ja (2021): The role of working memory capacity in interpreting performance. Translation, Cognition & Behavior 4(1), 26–46. doi.org/10.1075/tcb.00050.bae.10.1075/tcb.00050.baeSearch in Google Scholar
Bajo, Maria T./Padilla, Francisca/Padilla, Presentación (2000): Comprehension processes in simultaneous interpreting. In: Chesterman, Andrew/Gallardo San Salvador, Natividad /Gambier, Yves (Eds.): Translation in context. Amsterdam: John Benjamins, 127–142. doi.org/10.1075/btl.39.15baj.10.1075/btl.39.15bajSearch in Google Scholar
Banich, Marie T./Compton, Rebecca J. (2011): Cognitive neuroscience. 3rd ed. Boston: Wadsworth Cengage Learning.Search in Google Scholar
Barik, Henri C. (1975): Simultaneous interpretation: qualitative and linguistic data. Language and Speech 18(3), 272–297. doi:10.1177/002383097501800310.10.1177/002383097501800310Search in Google Scholar
Binder, Jeffrey R./Desai, Rutvik H. (2011): The neurobiology of semantic memory. Trends of Cognitive Science 15(11), 527–536. doi:10.1016/j.tics.2011.10.001. 10.1016/j.tics.2011.10.001Search in Google Scholar
Binder, Jeffrey R./Desai, Rutvik H./Graves, William W./Conant, Lisa L. (2009): Where is semantic system? A critical review and meta-analysis of 120 functional neuroimaging studies. Cerebral Cortex 19, 2767–2796. doi: 10.1093/cercor/bhp055.10.1093/cercor/bhp055Search in Google Scholar
Chincotta, Dino/Underwood, Geoffrey (1998): Non temporal determinants of bilingual memory capacity: The role of long-term representations and fluency. Bilingualism: Language and Cognition 1, 117–130. doi:10.1017/S1366728998000248.10.1017/S1366728998000248Search in Google Scholar
Christoffels, Ingrid K./de Groot, Annette M. B./Kroll, Judith F. (2006): Memory and language skills in simultaneous interpreters: The role of expertise and language proficiency. Journal of Memory and Language 54, 324–345. doi: 10.1016/j.jml.2005.12.004.10.1016/j.jml.2005.12.004Search in Google Scholar
Chung, Hye-Yeon (2014): Measuring interpreters' memory (in Korean). Korean Journal of Translation Studies 15(5), 221–244. doi:10.15749/jts.2014.15.5.009.10.15749/jts.2014.15.5.009Search in Google Scholar
Chung, Hye-Yeon (2018): Sequentiality in interpreters’ memory (in Korean). Korean Journal of Translation Studies 19(1), 227–246. doi:10.15749/jts.2018.19.1.009.10.15749/jts.2018.19.1.009Search in Google Scholar
Conway, Andrew/Engle, Randall W. (1994): Working memory and retrieval: A resource-dependent inhibition model. Journal of Experimental Psychology: General 123(4), 354–373. doi:10.1037//0096-3445.123.4.354.10.1037/0096-3445.123.4.354Search in Google Scholar
Cowan, Nelson (1988): Evolving conceptions of memory storage, selective attention and their mutual constraints within human information-processing system. Psychological Bulletin 104(2), 163–191. doi: 10.1037//0033-2909.104.2.163.10.1037/0033-2909.104.2.163Search in Google Scholar
Cowan, Nelson (2000): Processing limits of selective attention and working memory. Interpreting 5(2), 117–146. doi:10.1075/intp.5.2.05cow.10.1075/intp.5.2.05cowSearch in Google Scholar
Darò, Valeria/Fabbro, Franco (1994): Verbal memory during simultaneous interpretation: Effects of phonological interference. Applied Linguistics 15(4), 365–381. doi.org/10.1093/applin/15.4.365.10.1093/applin/15.4.365Search in Google Scholar
Dillinger, Mike (1990): Comprehension during interpreting: What do interpreters know that bilinguals don’t? Interpreters’ Newsletter 3, 41–58. doi.org/10.1075/btl.3.14dil.Search in Google Scholar
Elmer, Stefan/Hanggi, Jürgen/Jäncke, Lutz (2014): Processing demands upon cognitive, linguistic, and articulatory functions promote grey matter plasticity in the adult multilingual brain: insights from simultaneous interpreters. Cortex 54, 179–189. doi:10.1016/j.cortex.2014.02.014.10.1016/j.cortex.2014.02.014Search in Google Scholar
Elmer, Stefan/Meyer, Martin/Marrama, Lucas/Jäncke, Lutz (2011): Intensive language training and attention modulate the involvement of fronto-parietal regions during a non-verbal auditory discrimination task. European Journal of Neuroscience 34(1), 165–175. doi:10.1111/j.1460-9568.2011.07728.x.10.1111/j.1460-9568.2011.07728.xSearch in Google Scholar
García, Adolfo (2019): The neurocognition of translation and interpreting. Amsterdam/Philadelphia: John Benjamins.10.1075/btl.147Search in Google Scholar
Gile, Daniel (2009): Basic concepts and models for interpreter and translator training. Amsterdam/Philadelphia: John Benjamins.10.1075/btl.8Search in Google Scholar
Hernandez, Arturo E. (2009): Language switching in the bilingual brain: what’s next? Brain and Language 109, 133–140. doi:10.1016/j.bandl.2008.12.005.10.1016/j.bandl.2008.12.005Search in Google Scholar
Hervais-Adelman, Alexis/Moser-Mercer, Barbara/Michel, Christoph M./Golestani, Narly (2015): fMRI of Simultaneous interpretation reveals the neural basis of extreme language control. Cerebral Cortex 25(12), 4727–4739. doi:10.1093/cercor/bhu158.10.1093/cercor/bhu158Search in Google Scholar
Hervais-Adelman, Alexis/Moser-Mercer, Barbara/Golestani, Narly (2011): Executive control of language in the bilingual brain. Frontiers in Psychology 2. Article 234, 1–8. doi:10.3389/fpsyg.2011.00234.10.3389/fpsyg.2011.00234Search in Google Scholar
Hervais-Adelman, Alexis/Moser-Mercer, Barbara/Murray, Micah M./Golestani, Narly (2017): Cortical thickness increases after simultaneous interpretation training. Neuropsychologia 98, 212–219. doi:10.1016/j.neuropsychologia.2017.01.008.10.1016/j.neuropsychologia.2017.01.008Search in Google Scholar
Hervais-Adelman, Alexis/Babcock, Laura (2019): The neurobiology of simultaneous interpreting: Where extreme language control and cognitive control intersect. In: Abutalebi, Jubin/Clahsen, Harald (Eds.): Bilingualism: Language and cognition. Cambridge: Cambridge University Press, 1–12. doi:10.1017/S1366728919000324.10.1017/S1366728919000324Search in Google Scholar
Hiltunen, Sinikka/Pääkkönen, Rauni/Vik, Gun-Viol/Krause, Christina M. (2016): On interpreters’ working memory and executive control. International Journal of Bilingualism 20(3), 297–314. doi.org/10.1177/1367006914554406.10.1177/1367006914554406Search in Google Scholar
Injoque-Ricle, Irene/Barreyro, Juan Pablo/Formoso, Jesica/Jaichenco, Virginia (2015): Expertise, working memory and articulatory suppression effect: Their relation with simultaneous interpreting performance. Advances in Cognitive Psychology 11(2), 56–63. doi.org/10.5709/acp-0171-1.10.5709/acp-0171-1Search in Google Scholar
Jeong, Hyeonjeong/Tatsumi, Keiko/Sugiura, Motoaki/Kawashima, Ryuta (2014): Neural correlates of simultaneous interpretation: The role of expertise and language proficiency. Proceedings of the 16th annual international conference of the Japanese Society for Language Sciences, 171–172. Search in Google Scholar
Kandel, Eric (2007): In search of memory: The emergence of a new science of mind. New York: W. W. Norton & Company.Search in Google Scholar
Köpke, Barbara/Nespoulous, Jean-Luc (2006): Working memory performance in expert and novice interpreters. Interpreting 8(1), 1–23. doi:10.1075/intp.8.1.02kop.10.1075/intp.8.1.02kopSearch in Google Scholar
Lilienthal, Lindsey/Tamez, Elaine/Shelton, Jill-Talley/Myerson, Joel/Hale, Sandra (2013): Dual n-back training increases the capacity of the focus of attention. Psychonomic Bulletin & Review 20, 135–141. doi: 10.3758/s13423-012-0335-6.10.3758/s13423-012-0335-6Search in Google Scholar
Liu, Minhua (2001): Expertise in simultaneous interpreting: A working memory analysis. Doctoral thesis, The University of Texas at Austin. Search in Google Scholar
Liu, Minhua/Schallert, Diane L./Carroll, Patrick J. (2004): Working memory and expertise in simultaneous interpreting. Interpreting 6(1), 19–42. doi:10.1075/intp.6.1.04liu.10.1075/intp.6.1.04liuSearch in Google Scholar
Mellinger, Christopher D./Hanson, Thomas A. (2019): Meta-analyses of simultaneous interpreting and working memory. Interpreting 21(2), 165–195. doi.org/10.1075/intp.00026.mel.10.1075/intp.00026.melSearch in Google Scholar
Miotto, Eliane C./Balardin, Joana B./Martin, Maria da Graça M./Polanczyk, Guilherme V./Savage, Cary R./Miguel, Euripedes C./Batistuzzo, Marcelo C. (2020): Effects of semantic categorization strategy training on episodic memory in children and adolescents. PLOS ONE 15(2), 1–17. doi.org/10.1371/journal.pone.0228866.10.1371/journal.pone.0228866Search in Google Scholar
Mizuno, Akira (2005): Process model for simultaneous interpreting and working memory. Meta 50(2), 739–752. doi.org/10.7202/011015ar.10.7202/011015arSearch in Google Scholar
Moser-Mercer, Barbara/Frauenfelder, Ulrich Hans/Casado, Beatriz/Künzli, Alexander (2000): Searching to define expertise in interpreting. In: Dimitrova, Brigitta Englund/Hyltenstam, Kenneth (Eds.): Language processing and simultaneous interpretation: Interdisciplinary perspectives. Amsterdam: John Benjamins, 107–131. doi.org/10.1075/btl.40.09mos.10.1075/btl.40.09mosSearch in Google Scholar
Nordet, Lilan/Voegtlin, Lara (1998): Les tests d'aptitude pour l'interprétation de conférence. La mémoire. Diploma thesis, Genève: Université de Genève. Search in Google Scholar
Nittrouer, Susan/Lowenstein, Joanna H./Wucinich, Taylor/Moberly, Aaron C. (2016): Verbal working memory in older adults: The roles of phonological capacities and processing speed. Journal of Speech, Language, and Hearing Research 59, 1520–1532. doi: 10.1044/2016_JSLHR-H-15-0404.10.1044/2016_JSLHR-H-15-0404Search in Google Scholar
Padilla, Benítez Presentación (1995): Procesos de memoria y atención en la interpretación de lenguas. Doctoral thesis, Universidad de Granada. Search in Google Scholar
Pöchhacker, Franz (2016): Introducing interpreting studies. 2nd ed. New York: Routledge.10.4324/9781315649573Search in Google Scholar
Riby, Deborah M./Porter, Melanie A. (2010): Williams syndrome. Advances in Child Development and Behavior 39, 163–209. doi:10.1016/b978-0-12-374748-8.00005-6.10.1016/B978-0-12-374748-8.00005-6Search in Google Scholar
Seleskovitch, Danica (1975): Langage, langues et mémoire. Paris: Minard.Search in Google Scholar
Seleskovitch, Danica (1978): Language and recognition. In: Gerver, David/Sinaiko, Wallace (Eds.): Language interpretation and communication. New York: Plenum Press, 333–341. doi.org/10.1007/978-1-4615-9077-4_29.10.1007/978-1-4615-9077-4_29Search in Google Scholar
Signorelli, Teresa M./Haarmann, Henk J./Obler, Loraine K. (2012): Working memory in simultaneous interpreters: Effects of task and age. International Journal of Bilingualism 16(2), 198–212. doi:10.1177/1367006911403200.10.1177/1367006911403200Search in Google Scholar
Singh-Curry, Victoria/Husain, Masud (2009): The functional role of the inferior parietal lobe in the dorsal and ventral stream dichotomy. Neuropsychologia 47, 1434–1448. doi: 10.1016/j.neuropsychologia.2008.11.033.10.1016/j.neuropsychologia.2008.11.033Search in Google Scholar
Stavrakaki, Stavroula/Megari, Kalliopi/Kosmidis, Mary H./Apostolidou, Maria/Takou, Eleni (2012): Working memory and verbal fluency in simultaneous interpreters. Journal of Clinical and Experimental Neuropsychology 34(6), 624–633. doi:10.1080/13803395.2012.667068.10.1080/13803395.2012.667068Search in Google Scholar
Timarová, Šárka/Čeňková, Ivana/Meylaerts, Reine/Hertog, Erik/Szmalec, Arnaud/Duyck, Wouter (2014): Simultaneous interpreting and working memory executive control. Interpreting 16(2), 139–168. doi:10.1075/intp.16.2.01tim.10.1075/intp.16.2.01timSearch in Google Scholar
Wagner, Richard K./Torgesen, Joseph K. (1987): The nature of phonological processing and its causal role in the acquisition of reading skills. Psychological Bulletin 101, 192–212. doi:10.1037/0033-2909.101.2.192.10.1037/0033-2909.101.2.192Search in Google Scholar
© 2023 Walter de Gruyter GmbH, Berlin/Boston
Dieses Werk ist lizensiert unter einer Creative Commons Namensnennung 4.0 International Lizenz.
Articles in the same Issue
- Frontmatter
- Frontmatter
- Artikel
- The reception and critique of literature in translation: Ahmad Shāmlū, the Poet-Translator and the case of Margot Bickel
- « Aujourd’hui, la cloche doit naître. Compagnons, vite au travail ! »
- Zur (Un-)Verständlichkeit der Rechtssprache am Beispiel der deutschen Rechtssprache
- La disqualification de l’expert : Le cas de l’expert traducteur
- Context-embedded phonological memory in interpreters
- Localising INGOs websites into Arabic: A case study of IFRC
- Rezension
- Holger Siever (2021): Komplexe Translationstheorie. Übersetzen im 21. Jahrhundert. München: AVM
Articles in the same Issue
- Frontmatter
- Frontmatter
- Artikel
- The reception and critique of literature in translation: Ahmad Shāmlū, the Poet-Translator and the case of Margot Bickel
- « Aujourd’hui, la cloche doit naître. Compagnons, vite au travail ! »
- Zur (Un-)Verständlichkeit der Rechtssprache am Beispiel der deutschen Rechtssprache
- La disqualification de l’expert : Le cas de l’expert traducteur
- Context-embedded phonological memory in interpreters
- Localising INGOs websites into Arabic: A case study of IFRC
- Rezension
- Holger Siever (2021): Komplexe Translationstheorie. Übersetzen im 21. Jahrhundert. München: AVM
