The participants of the present study, 73 in total, were undergraduate students majoring in Spanish at the China Foreign Affairs University during their sixth and seventh semesters (aged between 21 and 22 years) when they received a one-year sight translation and consecutive interpretation training. The undergraduate students were Chinese, with Chinese as their native language and Spanish as their B language, mostly at Level C1 according to the Common European Framework of Reference for Languages (CEFR) levels, and English as their C language, mostly at Level B. Each student participating in this study was assigned a unique four-digit identification number. All errors in the corpus were marked with student IDs in the “ST” format, followed by four digits (e.g., ST1234), enabling the identification of error sources.

This study collected audio recordings of Chinese–Spanish sight translation sessions of the participants. The interpreting format employed in this study was sight translation. Before the translation, the students were given the interpreting theme to prepare, including to create vocabulary list, read background knowledge, etc. For the translation, students are provided with Chinese source texts on the spot and given 2 min to read through the material before rendering it into the target language. The preparation only permits making slashes for segmentation and marking important word meanings, and students cannot consult other resources or perform written translation. There were 19 original translation texts. Each text was translated by 9–20 students, and the entire corpus contained 294 audio assignments. To comprehensively analyze the difficulties of sight translation, we performed a quantitative analysis of the textual characteristics of the Chinese source texts. The HanLP and jieba were utilized for language processing, to calculate the length, tokenization (abbreviated as Token), semantic roles count (Sem. Rol), and dependency relation count (Dep. Rel). Four-character words (4CW) and terminologies (Trm) were tagged manually by the author, since online dictionaries do not incorporate the latest four-character phrases and neologisms. All statistics are shown in Table 1.

The author analyzed the collected audio recordings to evaluate errors in the students’ interpreting performance. This approach facilitates a contrastive analysis between the source and target texts, enabling an examination of students’ ability to transfer linguistic elements. In collating the error taxonomy, we reviewed existing error types, which include coherence, omission, addition (Barik, 1971; Deng, 2018; Falbo, 1998), lack of cohesion, clumsiness of expression (Falbo, 1998), logical errors, overelaboration, and self-correction (Deng, 2018). Other studies exist that aim to evaluate the interpreters’ abilities, such as proficiency in the second language and in the native language, knowledge of the topics being interpreted, ability to analyze and synthesize, mastery of interpreting strategies, communication skills (including active listening, clear expression, and empathy toward the speaker, relationships with the client and the audience, cultural adaptation, physical endurance, group work) (Deng, 2018; Novo Díaz, 2004). Based on previous studies and considering that the form of interpreting of the corpus is sight translation, and that the research objective is to evaluate the interpreter ability and language proficiency of students, this study employs the descriptive classification in the formulation of error taxonomy. The linguistic perspective serves as the primary framework for identifying and categorizing errors for that it offers a structured approach to classifying errors using concrete, measurable criteria. The errors are categorized into three main levels: lexical, grammatical and syntactic. Within each level, errors were further classified into three primary manifestations: substitution (the interpreter replaced part of the original text, causing the translation to deviate from the original content), omission (the interpreter omitted the content conveyed by the speaker), and addition (the interpreter added redundant content based on the original text). At each level, the sublevels were classified according to the pattern of interpreting errors detected in the corpus. At the lexical level, substitution errors encompass lexical use, selection, morphological inaccuracies, and terminological discrepancies. Syntactic errors relate to issues with omission and addition of syntactical elements, errors of sentence structure and word order, whereas grammatical errors include agreement in number and gender, verb tense and conjugation, subject-verb agreement, preposition, article, numerical errors and connector. The error taxonomy used in this study is presented in Table 2.

Audio recordings were analyzed and the detected errors were categorized according to the established error taxonomy and transcribed into the corpus. Erroneous sentences were manually transcribed within the corpus, with the errors clearly marked. For each error in the corpus, the following parameters were marked: student ID, article topic, sentence number, sentence containing the error, error manifestation, error level, and error sub-level if applicable. The author systematically identified and categorized interpreting errors based on the established error taxonomy. Subsequently, a corpus of students’ Chinese–Spanish interpretation errors was compiled, totaling 2,923 instances. Statistical analysis was conducted using python to examine the distribution of errors across various error levels, concrete error types, error manifestations, and sentence numbers in each original text. A qualitative analysis was also realized based on the theoretical framework to understand the error causes and characteristics. Therefore, the error analysis based on this taxonomy significantly advances the field by identifying common challenges and interpreter weaknesses, analyzing cognitive processes, and enhancing quality control criteria. It addresses a critical gap in existing taxonomies (Falbo, 1998; Kopczynski, 1981; Gile, 1987), which have overlooked the specific errors encountered in Chinese–Spanish sight translation.

It is important to note that although errors at lexical, syntactic and grammatical levels show different proportional distributions in the corpus, no direct comparisons of error prominence across these levels or inferences about students’ translation difficulties can be made due to the varying total occurrences of words, sentences, and grammatical structures in the corpus. In this study, error analysis can only be conducted by comparing prominent error sublevels within each respective level. As shown in Table 3, the most frequently committed errors for each level were as follows: At the lexical level, word selection (670 errors, 37.85% of lexical errors) was the most frequent error. Substitution of terminology (19.32%) and word omission (18.64%) were other frequent problems. At the syntactic level, the predominant errors observed were substitutions related to sentence structure (237 errors, 65.29% of sentence errors) and omissions of syntactic elements (106 errors, 31.27%). The grammatical errors made by students are predominantly concentrated in agreement in gender and number (280 errors, 35.44%), verb tenses and conjugation (212 errors, 26.84%), prepositions (88 errors, 11.14%), and articles (68 errors, 8.60%). Regarding the error manifestation, the three levels show a similar error distribution pattern: high percentage of errors in substitution, followed by omission, and least amount of addition errors. Only at the syntactic level, omission errors are relatively more (30.02% of all syntactic errors), compared with the other two levels.

Regarding the correlation between interpreting topics and the error level, with a Pearson Chi-square value of 2.46% (df = 36) and a p-Value of 1.0, we find that the error rate of the lexical, syntactic and addition variables across different topics do not show a significant difference. This suggests that the topical domain is not a decisive factor in determining error patterns across various linguistic levels. A more detailed examination at the sentence level is required to understand the underlying causes of these interpretation errors. Regarding the 141 sentences that constitute our entire corpus, the median error rates of the lexical, syntactic and grammatical levels are, 1.83, 0.40, and 0.81%, respectively, while the highest error rates are 5.88, 2.22, and 3.70%, respectively. The median error rates across all three levels are below 2%, indicating a relatively low overall percentage. However, the highest error rates of sentence, ranging from 2.22 to 5.88%, indicate a significant increase in errors in individual cases. Moreover, both the median and maximum values of ER at the lexical level are notably higher than the other two levels, which, however, does not necessarily indicate that lexical errors are more significant than those at other levels, as the frequency of lexical items in a single sentence typically exceeds the number of syntactic structures and grammatical occurrences.

According to the frequency analysis, as shown in Table 4, based on different error manifestations, there were 2,363 cases of substitution errors in the corpus, accounting for 80.84% of the total errors. Omission and addition errors were relatively fewer, with 492 (16.83%) and 68 errors (2.33%), respectively. This implies that the primary issue for students in interpreting is their inability to translate the original text accurately with the correct expressions, with fewer instances of problems with redundancy or omission. The error distribution pattern presented different tendencies. In substitution, lexical errors occupied the main part (59.46%), followed by grammatical (29.88%) and syntactic errors (10.66%). In omission, lexical errors were the main part (67.07%), followed by syntactic errors (22.15%) and grammatical errors (10.77%). Regarding addition errors, the percentage of lexical and grammatical errors are relatively close, at 51.47 and 45.59%, respectively.

The Chi-square test was conducted to examine the association between interpreting topics and error manifestation. With a Pearson Chi-square value of 2.34% (df = 36) and a p-Value of 1.0. We find that the error rate of the substitution, omission and addition variables across different topics do not show a significant difference. This result is consistent with the correlation findings among error levels, suggesting that the subject matter of interpretation is not a major determinant of students’ error patterns across different error manifestations. Regarding different sentences, the median error rate of substitution, omission and addition levels are 2.46, 0.52, and 0.07%, respectively, while the highest error rates are 6.72, 2.43, and 0.79%, respectively.

The error rate of the 19 original texts was also calculated to discover the patterns of texts, as shown in Table 5. Comparing the text properties in Table 1 and the error rates in Table 5, no correlations were found, implying that the concentration of errors does not have a clear correlation with the translation topics. Error rates are influenced more by individual sentence structures and expressions rather than by specific textual characteristics. Therefore, the discussion section will provide a qualitative analysis of the sentences that concentrate the highest error rate.

As shown in the Table 6, sentences with high lexical error rate (LER) do not demonstrate notable values in their textual parameters. It is particularly noteworthy that most of these sentences do not contain a large number of four-character idioms or technical terms. Only the TD_5,¹ which lists technical terms, contains six specialized terms and new words. Our qualitative analysis shows that the sentences with high lexical error rates have the following characteristics: First, sentences with high LER may contain abstract words. Borghi and Binkofski (2014) define abstract words as those that denote intangible and imperceptible objects and concepts grounded in physical entities and concrete, singular objects. Due to their complexity (Barsalou, 2003) and semantic variability, students often struggle to precisely understand their meanings and to find appropriate Spanish equivalents. Sentences that exhibited errors of this type include TR_5, TM_2, TD_4, TS_6, and TQ_6. For example, in TR_5, students encountered difficulties in finding the Spanish equivalent of “相关单位” [entidades pertinentes]² and either omitted entire words or parts of them (ST1756, 1761, 1748). Other students made errors by choosing words with similar forms or by confusing synonyms, as in the case of “unidades relevantes” (相关单位) (ST1748, 1749), “unidades relativas” (相对单位) (ST1759), “autoridades pertinentes” (有关当局) (ST1758), “institutos competentes” (主管机构) (ST1743). Another typical example of error of this cause is the interpretation of “新愿景” [la nueva visión] in TM_4. This expression emphasizes a new perspective and expectation, rather than just a hope. For this reason, “nueva esperanza” (新希望) (ST1753, TM_4) cannot translate the meaning of guiding vision or strategic plan. Likely, “nueva previsión” (新预测) (ST1747, TM_2) refers to a forecast, prediction, or provision, which implies a more specific and concrete estimation of future events or needs. Such errors, coming from abstract words, show that students struggle to accurately understand the meanings of words in both the source and target languages within specific contexts (Xiang and Zheng, 2018). Different semantic fields between languages cause confusion in synonyms selection, increasing the coordination load for interpreters. The limited linguistic and cultural schema of the students also hindered their ability to find corresponding translations for the original text.

Second, errors tend to occur in sentences exhibiting a high density of professional terminology, neologisms, or specialized expressions within a particular domain. Prior to the translation task, participants were informed of the subject matter and could conduct pre-translation preparation, including terminology studying. However, they were not aware of the specific terms and proper nouns that would appear in the source texts. Therefore, the mastery of terminology is, to some extent, closely related to the thoroughness of interpreters’ preparation and the comprehensiveness of their established terminology database.

Table 6 shows that the terminologies of new technology and neologisms represent a challenge to the vocabulary and background knowledge of the interpreters. For instance, in TD_5, errors are detected with the relevant terminologies showcased in China International Fair for Trade in Services. “数字贸易” [comercio digital] is erroneously interpreted as “comercio de inteligente” (聪明贸易) (ST1747), “transacción digital” (数字交易) (ST1757) or “comercios internet” (互联网贸易) (ST1743). “5G通讯” [comunicaciones 5G] registers an error as “comunicación según G” (根据G通讯) (ST1747). “智慧办剬” [oficina inteligente] has various interpretations, such as “oficinas intelectuales” (智力办剬室) (ST1756), “oficio inteligente” (智慧型职业) (ST1761), “trabajo de inteligencia” (智能劳动) (ST1733), “trabajo artificial” (人工劳动) (ST1743), and “gestión inteligente” (智慧管理) (ST1757). While some of these interpretations convey the general idea, they do not accurately capture the concept of “intelligent office.” “跨境电商” [comercio electrónico transfronterizo] is interpreted erroneously as “comercios electrónicos internacionales” (国际电子商务) (ST1756), “comercio online de transporte” (运输服务在线交易) (ST1749), etc. In summary, the main issues with the interpretations of these technical terms and neologisms are inconsistency, inaccuracy, and the use of nonstandard or improvised interpretations that do not effectively convey the original Chinese concepts.

In sentences of TD_6 and TM_4, the Chinese neologism “新业态” [nuevos formatos] and “新模式” [nuevos modelos] create difficulty for interpreters to understand precisely the meaning and find the suitable Spanish words. Errors are detected such as “la nueva normalidad y modalidad” (新常态与新模式) (ST1755), “nuevos formatos y modos” (新形式与新方式) (ST1754), “una nueva situación y un nuevo modelo” (新形势与新模型) (ST1750), “nuevo tipo de negocio” (新型业务) (ST1733), “nuevos aspectos” (新方面) (ST1745), and “nueva moda y nueva industria” (新时尚与新产业) (ST1747).

Moreover, the substitution of professional terminology, including institutional names and economic terminology detected in our corpus, significantly contributes to translation errors due to students’ unfamiliarity with Spanish expressions and their lack of corresponding linguistic-cultural knowledge. For example:

In the sentence TC_6, “澳门特别行政区” [el Gobierno de la Región Administrativa Especial de Macao] is interpreted as “distrito Aomen” (澳门区) (ST1743), “región especial Aomen” (澳门特区) (ST1746) or “la comunidad especial de Aomen” (澳门特别社区) (ST1753).

In TG_4, “国家统计局” should be translated as “Instituto Nacional de Estadísticas” or “Oficina Nacional de Estadísticas,” but students did not comprehend the expression and invented words:

Errors were detected such as “Ministerio de Estadísticas” (统计部) (ST1762), “Oficina Nacional de Estrategia” (国家战略办剬室) (ST1754) and “Centro Nacional de Estadísticas” (国家统计中心) (ST1758). In TO_5, there are relevant cases of errors of the economic terminology “增加值” [valor añadido], which is translated as “el valor acelerado” (加速值) (ST1763), “el volume total de producción” (总产量) (ST1761), or just omitted (ST1751, 1756). The main causes of terminology errors include a lack of terminological and cultural schemata both in Chinese and Spanish. This unfamiliarity and insufficiency also increase the cognitive load for the interpreters, leading to omissions or inaccuracies in their translations (Hardy et al., 2013; Horváth, 2019). These findings agree with the equivalent recall discussed by Putranti (2017).

The third type of sentence involves common sayings, chengyu fixed expressions, and poetic phrases. In TF_2, errors arise from the inability to transfer semantic relations, such as interpreting “金秋九月” [septiembre dorado de otoño] as “En el otoño dorado y el septiembre,” (在金色的秋天和九月) (ST1756), and “en este otoño de oro” (在这个金色的秋天) (ST1755). Other students fail to select the correct Spanish word to translate “丹桂” [osmanthus fragante], interpreting it as “flores” (花) and “árboles” (树) (ST1755, 1743).

Furthermore, certain fixed formulations or collocations exhibited a four-character pattern, predominantly within the political domain. The compact structure, its incisive meaning, and the concentrated embodiment of linguistic rhetoric and cultural accumulation (Kang and Yang, 2022) are characteristics of these idioms, and cause interpreting difficulties for students. For example, in TG_3, “六稳” “六保” [las “seis estabilidades” y las “seis garantías”] is a numeric abbreviation which is often used in Chinese political texts. Its concise expression and cultural and political context create a high level of difficulty for understanding and translating. Error causes can include unfamiliarity with the background knowledge and the significance of the term, and committing errors, such as “seis estabilidad y seis seguridad” (六稳六保) (ST1733), “seis estabilizaciones y seis garantías” (六稳六保) (ST1753), and “seis estabilidad y las garantías” (六稳和保障) (ST1762). In TN_2, the interpretation of “开放包容” [apertura e inclusión] led to errors like wrong selection and formation of the words “la apertura e inclugencia” (开放与包容) (ST1478), “la tolerancia abierta” (开放的包容) (ST1464), “la abirdad y tolerancia” (开放和宽容) (ST1479), and “la actitud tolerante y abierta” (宽容开放的态度) (ST1476), as well as omission of information – “la abierta” (开放) (ST1469).

Apart from the typical sentences that contain higher LER, word selection, substitution of terminology and omission of words are the three lexical sublevels with prominent errors (as shown in Table 4). The reason for the concentration of errors in word selection may be insufficient vocabulary comprehension (Jiménez Ivars and Hurtado Albir, 2008), and multiple similar vocabulary choices may exist, as discussed in the part of abstract words. Interference by the translator’s mother tongue (intralingual interference) and lack of linguistic schemata of the target language can also cause translation errors (Phạm, 2018; Utami, 2017). Moreover, words with abstract meanings and common sayings are more difficult to process cognitively than concrete words. Research has shown that concrete words are more easily remembered and processed than abstract words, which may contribute to higher error rates in interpreting abstract concepts (Fliessbach et al., 2006). The terminology error has been analyzed in the second type of high LER, while the third sublevel, omission, is spread across different sentences, mainly due to lack of understanding of the correct use of the Spanish expressions, or failing to remember the expression and omitting information, as is shown in TC_5:

In this sentence words like “创新” [innovación] (ST1761), “合作” [cooperación] (ST1733), “包容” [inclusión] (ST1749, 1747, 1753), “复苏, 保障” [recuperar, garantizar] (ST1753) and “人民” [el pueblo] (ST1762) are omitted. Word usage errors are also detected, including translating “愿” [estar dispuesto a] as “tiene muchas ganas para” (非常渴望) (ST1743), and “为…保障人民身体健康作出贡献” [contribuir a la salud del pueblo] as “contribuir (a^*3) la protección de la salud de los ciudadanos” (为保护剬民健康做出贡献) (ST1743).

At the syntactic level, similarly, we detected the 10% sentences with the highest syntactic error rate (SYER), as shown in Table 7. Firstly, the sentences including poems or ancient sayings present a higher level of difficulty for interprets’ understanding and skills, therefore leading to more errors. The sentences TB_4, TF_2, and TE_4 are relatively low in sentence length, semantic roles, and dependency relationship, but their condensed and context-dependent expressions, culture-specific allusions and references, as well as the parallel structure signify difficulties for interpreters. In TB_4, the main detected error is omission, meaning that interpreters tend to omit part of the sentence for evasion due to the difficulty of understanding the Spanish expression in Chinese. Incomplete sentences were found like “cincuenta años de viento y lluvia inusuales, cincuenta años de arduo trabajo.” (五十年风雨不寻常, 五十年艰苦奋斗) (ST1745), “durante el trabajo de más de cincuenta años” (在五十多年的工作中) (ST1733), “han transcurrido cincuenta años de singularidad y de cultivo duro” (经历五十年的非凡与艰辛耕耘) (ST1761), “tras cincuenta años de trabajo duro” (五十年的艰苦奋斗后) (ST1762), and “cincuenta años de aventura y esfuerzos son inusuales” (五十年的冒险与努力是非凡的) (ST1749). In TF_2, the errors include “En el septiembre con la aroma de la flora” (在充满花香的九月) (ST1761), “En el otoño de septiembre” (在九月的秋天) (ST1749), and omission of the whole sentence (ST1754). In TE_4, we detected omission of the whole sentence (ST1856), as well as the following substitution errors, indicating sentence structure distortion and ambiguous expressions in Spanish, like “Amigos con ambiciones no se llevan las montañas y los mares tan lejos.” (志同道合的朋友不嫌山海远) (ST1862), and “Incluso los voluntario no para montañas y mares lejos.” (连志愿者也不因山海遥远而却步) (ST1845).

Second, the complexity of sentence structures is another source of errors, due to factors including sentence length, numerous semantic roles, and dependency relationships, like in sentences TN_6, TP_5, TD_4, TL_9, TP_3 and TQ_6. In TN_6, the syntactic errors occur in the adverbial clause “在最后这不到一个月的时间里” [En este último tiempo de menos de un mes], for its high dependency relationship and the need to adjust word order due to differences between the two languages: “En este tiempo que menos de un mes” (在这不到一个月的时间里) (ST1467), “En los últimos tiempos, no alcanza un mes” (在最近不到一个月的时间里) (ST1472), “Durante los últimos tiempos que de menos de un mes” (在最近不到一个月的时间里) (ST1479), “En el/este último mes” (在过去这一个月里) (ST1476, 1,469, 1,473), “En la última etapa en que no llega un mes” (在不到一个月的最后阶段) (ST1477). Moreover, in TN_6, “确保论坛顺利圆满举行, 推动论坛取得更多丰硕成果, 助力“一带一路”国际合作开辟新局面, 迈上新台阶!” [Asegurar que el foro se lleve a cabo de manera fluida y exitosa, promover que el foro logre resultados más fructíferos y contribuir a abrir una nueva etapa y alcanzar un nuevo nivel en la cooperación internacional de la Franja y la Ruta.] is a serial verb construction in Chinese, which creates an additional challenge for the translator.

The syntactic complexity and unfamiliarity with Spanish grammar lead the ST1470 and the ST1469 to produce syntactic calques, failing to capture the syntactic logic and accurately reconstruct the sentence. Syntactic calques of the Chinese structure especially occur in sentences with relatively high dependency rate, or in run-on sentence. In TP_3, the interpretation “una de las relaciones bilaterles más que se ha armoniza profundamente en el ámbito de cooperación más grande y tiene el beneficio mutuo también más grande.” (其中一个最为和谐的双边关系, 在最广泛的合作领域中深入发展, 并具有最大的互利效益。) (ST1733) commits errors of awkward phrasing and incorrect word order.

In TP_5, errors are detected in the use of specific syntactic structures, particularly in comparative constructions:

These errors clearly stem from using the comparative structure, especially in a sentence with a high dependency value (27), which reflects students’ inability to transfer complex Chinese sentences to a target language of the Spanish norms. This also supports the findings of several studies suggesting a greater cognitive load posed by sentences that need to be syntactically restructured in the target language (Gernsbacher and Shlesinger, 1997; Ma et al., 2021; Viezzi, 1989). Due to the differences in word order between Spanish and Chinese, the interpreters must rearrange the sentence elements to conform to the Spanish syntactic norms while preserving the original meaning, thus increasing the coordination effort (Chernovaty et al., 2023; Lee, 2012).

Apart from long and complex sentences, problems also occur in the distortion of syntactic structures. In the corpus, “在中国抗击疫情最困难的时刻” [en el momento más difícil de la lucha de China contra la pandemia] (TQ_9, SenT. Len 59, Dep. Rel 37, SYER 0.47%) is interpreted as “Cuando en el momento más difícil en China” (在中国最困难的时刻里) (ST1733) with the redundancy of “cuando” and “en.”

Regarding the sublevel of syntactic errors, apart from the already discussed substitution, omission is another serious syntactic error manifestation, that may occur when students encounter time constraints, complex sentence structures, or a substantial volume of information. Sentences TF_4, TR_6 in Table 7 have a high ER primarily due to partial or complete omissions. Their positions within the speech texts led to inappropriate attention allocation by the students. Specifically, TF_4 follows a lengthy sentence, while TR_6 is situated in the concluding section of the speech. The cognitive pressure experienced by interpreters appears to have contributed to them overlooking these sentences.

In summary, the sentences that exhibit higher SYER include poems, ancient saying, and sentences of long and complex structures. The predominant errors observed were substitutions related to sentence structure and omissions of syntactic elements. Additionally, lower proficiency in mastering various sentence structures and the usage of connecting words in Spanish hindered the translators’ ability to adjust sentence structures effectively. The causes of syntactic errors in interpretation are mainly the cognitive load caused by complicated and lengthy sentences, the differences in syntactic structure between Spanish and Chinese and the difficulty of adjusting the sentence structure, and the difficulty in using appropriate connectors between clauses. This result is also in accordance with previous studies that pointed out that difficult words and complex sentences cause increased cognitive effort (Shreve et al., 2010) and dysfluency (Shreve et al., 2011) in sight translation.

At the grammatical level, we detected the 10% sentences with the highest grammatical error rate (GER), as shown in Table 8. The results indicate that the sentences with the most grammatical errors are not necessarily those with relevant length or dependency relationship, but sentences that have different grammatical uses between Chinese and Spanish, like TC_3, TQ_2, TQ_1, TQ_7. Spanish verbal inflections, particularly in terms of tense and mood, represent grammatical features that are absent in Chinese, thus frequently leading to translation errors. For example, TC_3 has a high GER due to the verb tense. Students used the present tense for “强调” [enfatizar] (ST1761, 1758, 1755) and “应该” [deber] (ST1751, 1749, 1733, 1762, 1748, 1743, 1754), which is a common error of verb tense for Chinese students. In TQ_2, we detected errors of using the wrong tense when translating “期盼” [esperar], where the present perfect tense should have been used.

Moreover, for TQ_2, errors in subject-verb agreement were also detected.

TQ_1 includes errors regarding expressing dates, like “en (*el) 14 de septiembre” (9月14日) (ST1758), as well as subject-verb agreement errors, as in “Los dirigentes de China, Alemania y la Unión Europea se entrevistará” (中国, 德国和欧盟的领导人将进行会晤。) (ST1761), “Los líderes de China, Alemania y la Unión Europea se está reuniendo” (中国, 德国和欧盟的领导人正在举行会晤。) (ST1748), “los líderes de China, Alemania y la Unión Europea va a organizar una conferencia” (中国, 德国和欧盟的领导人将组织一次会议。) (ST1743), and using the simple past tense “se reunieron” (会晤) (ST1746) in the place where the future tense should have been used. TQ_7 shows the errors in determining the right verb tense and translating the predicate “密切合作, 互帮互助” [han trabajado estrechamente para ayudarse mutuamente] resulting in errors, such as:

TD_6 is an example of errors from using verb tense “estarán atrayendo a la atención” (将吸引关注) (ST1754) or subject-verb agreement in interpreting “吸引着外媒的关注” [han atraído la atención de los medios extranjeros], like “ha atraído la atención” (吸引着外媒的关注) (ST1756, 1745, 1747). Although the structure of the Chinese sentences is not inherently complex, the fundamental differences between Chinese and Spanish grammar are the main cause of these errors. Moreover, insufficiency in linguistic and formal schemata of the target language also contribute to such errors.

Furthermore, the sentences with significant length and dependency relationship have a tendency of exhibiting more errors, like TG_2, TR_2, TR_3, TI_7, TC_7, due to the bigger memory and coordination pressure. Take the sentence TG_2 for example:

ST1751 makes errors in subject-predicate agreement due to the length of the sentence, therefore increasing the coordination pressure. In the TD_3 of ST1743, we detected the errors of using the past participle form “incluido” (包括) of the verb “incluir” (包括) instead of using its gerund form “incluyendo” (包括). There are also errors of agreement in number and gender in the number “doscientos” (二百), and in the verb tense of “participar” (参加).

In TI_7 of ST1683, we detected errors of the verb tense of “señala” (指出), where “señalado” (指出的) should be used, as well as erroneous use of “la política monetaria” (货币政策), where the plural form should be employed:

These errors are caused due to the complexity of tense conjugation, the voice and mood of Spanish verbs, and the difficulty of identify the subject in a long sentence. The insufficiency of grasping the grammatical rules of Spanish is also a cause of errors. Chinese lacks grammatical gender and a robust system of noun-adjective agreement based on number. This fundamental difference in language structure makes it challenging for Chinese learners to internalize the Spanish system of grammatical gender and number agreement. Another possible error cause is that the students prioritize conveying meaning, which, however, could deviate their attention from the grammar in long and complex sentences.

Regarding the manifestation of grammatical errors, as shown in Table 4, the most prominent sublevel of error is the agreement between gender and number, manifested in various forms. Moreover, this sublevel of errors is not concentrated on specific expressions or grammatical structures, but rather spread across a variety of linguistic contexts. For example, we detected the following errors in sentence TR_2:

From the errors, we can see that the forms vary from gender agreement of articles and adjectives “el esfuerzos comunes” (共同努力) and “las visita” (访问), to errors in number agreement within noun phrases, such as “ambos partes” (双方) and “los esfuerzos conjunto” (共同努力). Agreement errors may also be generated because of the source text influence and short-memory effort failure in long sentences.

Errors in verb tense and conjugation, inconsistency in verb tense, and in singular and plural verbs and in subject-predicate are also prevalent, as discussed in the previous examples. Prepositions and articles are two grammatical phenomenon that do not occupy a high percentage of errors, but worth noticing that within these two sublevels, the most frequent errors are substitution of preposition, and omission and addition of articles. The cause of the prepositional errors could be interference from the mother tongue, as Chinese and Spanish have very different prepositional systems. Spanish has a wide range of prepositions, each with specific uses and meanings. Some prepositions can be used interchangeably in certain contexts, whereas others have very distinct uses. This complexity can be challenging for Chinese interpreters to grasp fully. Some prepositional errors may arise from a lack of in-depth understanding of the Spanish language’s grammatical structures and idiomatic expressions.

Regarding the errors in articles, omission and addition were the main error manifestations. An obvious error is the omission of definite articles, such as “invitar a (*el) Sr. Zhang” (有请张先生) (TR_7, ST1754), “Los líderes de China, Alemania, (*la) Unión Europea” (中国, 德国, 欧盟的领导人) (TQ_1, ST1746) “reunirse en (*el) 14 de septiembre” (9月14日会晤) (TQ_1, ST1758), and “el intercambio de (*el) personal” (人员交流) (TP_6, ST1733). The addition occurs when using vocative case, such as “Estimados todos, los líderes, los invitados, los maestros y los alumnos” (各位领导, 各位来宾, 老师们, 同学们) (TB_1, ST1756); or adding articles before a proper noun, for example, “la Europa y China” (欧洲和中国) (TQ_10, ST1743).

We can see that grammatical errors made by students are predominantly concentrated in areas where Spanish grammar differs from or lacks direct equivalence to Chinese grammar, such as agreement in gender and number, verb tense and conjugation, prepositions, and articles. These errors are also typical difficulties for Chinese learners of Spanish, and they are highlighted in the Chinese–Spanish interpretations. This finding is in line with Agrifoglio (2004), who found that interpreters often lost the referent and forgot the gender, number, and person. This problem may have been exacerbated by the fact that inflection is more common in Spanish and that adjectives, pronouns, and verbs need to be put into proper agreement.

The error distribution and frequency analysis show that students have major problems correctly using Spanish expressions, whereas errors of redundancy and omissions are not as serious (see Table 4). This result is in accordance with the discovery by Agrifoglio (2004), who argues that sight translation has an advantage over simultaneous interpretation and consecutive interpretation in the content, thus committing fewer errors related to the content, such as omissions, changes in meaning, or incomplete sentences, but more errors in expression, such as lexical, syntactic, or grammatical problems, calques, and language style, mostly caused by short-term memory failures of retrieving information from the beginning of sentences, or the formulation the interpreter has already embarked on, especially where grammatical structures differ markedly between the two languages.

Table 9 shows that sentences that frequently contain substitution errors also appear to have higher error rate at the lexical, syntactic or grammatical levels (refer to Tables 6–8). These sentences have complex structures or high dependency relationships (see Table 10, TP_5, TD_3, TC_7, TL_9), include terminology or neologisms that students have not yet mastered (TD_6, TG_3, TD_5, TD_4), use four-character or abstract words with difficult-to-find Spanish equivalents, or exhibit more than one of these characteristics (TL_9). This phenomenon is also explained by the high percentage of substitution at each of the three error levels and in the total count of the corpus (Table 4). Therefore, the high percentage of substitution errors at all levels suggests a possible correlation between substitution error rate (Sub. ER) and error concentration in sentences (Table 9), though further investigation may be needed to confirm this relationship.

The second common form of error is omission (Table 11), as students tend to omit words in certain professional areas, terminology, poetic and expressions (TF_2, TE_4, TB_4), among others, which they do not know how to express in Spanish. For example, in TO_5, professional terms, like “增加值” [valor añadido] (ST1751, 1746), are omitted, as well as numerical expressions “15%左右” [cerca de 15%] (ST1751, 1758), and the adverbial clause “近年来” [en los últimos años] (ST1756, 1746). In TD_4, “全貌” [el panorama] is omitted because the students do not know how to express it in Spanish (ST1756, 1749, 1733, 1770, 1748, 1747, 1743, 1754). “相关单位” [entidades concernientes] in TR_5 (ST1756, 1733, 1760, 1748) and the neologism of “新业态” [el nuevo formato] in TM_4 (ST1733, 1762, 1760, 1748, 1757, 1746, 1753) are omitted for the same reason.

Moreover, students also omitted words in complex structure sentences, long sentences, or high dependency relationship, due to linguistic differences and memory and coordination pressure. For example, in TR_5, students (ST1750, 1751) omitted the object “感谢” [agradecimientos] probably due to the long adverbial clause and modifiers. In TD_3, which is another long sentence, students omitted “线上线下” [en línea y de forma presencial] (ST1756, 1733, 1748, 1746) or “参展参会” [participar en la exposición] (ST1756, 1746) because of the sentence length and the unfamiliarity with the equivalent Spanish expressions. There are also cases of omission of whole phrases. We detected that ST1754 and ST1746 omitted parts of the sentence TO_6 for evasion due to their lack of ability to translating the words and sentence structure.

In TR_7, a high omission error rate (Om. ER) is also detected, although it’s a sentence with simple structure and words. Two cases of omission of whole sentence are detected (ST1750, 1746), as well as one case of omission of the second part of the sentence “有请省政府外事办主任张先生主持。谢谢大家!” [Invitamos al señor Zhang, director de la Oficina de Asuntos Exteriores del Gobierno Provincial, a presidir. ¡Gracias a todos!] (ST1748). This phenomenon is unusual in the corpus, so we assume that the omission is caused by the concentration of attention in the core information of the text and the ignorance of the last sentence. Other omission errors of this sentence are related to the terminology, like interpreting “省政府” [la Oficina de Asuntos Exteriores del Gobierno Provincial] as “Oficina de los Asuntos Exteriores” (ST1762) or omitting “张先生” [el señor Zhang]. We assume that this omission is partly due to unfamiliarity of expression of Chinese institutions in Spanish, but the main cause of error should be the pressure interpreter feels and the attention allocation issue.

In conclusion, the textual characteristics of sentences that concentrate errors of omission can be terminological or poetic expressions, or other words that are difficult to translate, leading to students omitting information for evasion. Omission can also be caused by cognitive load or attention allocation issue.

Within addition, which is the third type of error manifestation, we firstly examined grammatical additions. In TB_1 and TR_7 (Table 12), students used articles in unnecessary positions, concretely speaking, before vocatives. The following are errors detected in TB_1:

In TR_7, when interpreting “省政府外事办主任张先生” [el señor Zhang, director de la Oficina de Asuntos Exteriores del Gobierno Provincial], students add articles like “el director de la Oficina de Relaciones Exteriores del Gobierno Provincial” (省政府外事办主任) (ST1745), “el jefe de los Asuntos Extranjeros del Gobierno” (政府外事办主任) (ST1733) or “el director de los Asuntos Exteriores del Gobierno” (政府外事办主任) (ST1747). Addition of articles also appeared when listing, as in TQ_8:

In TF_7, we detected two cases of additional article use in four-character words, like “una historia espléndida de la cooperación y las ganancias compartidas” (合作与共赢的辉煌历史) (ST1762), and “capítulos más espléndidos en el camino de la cooperación y la ganancia mutua” (合作与共赢道路上更辉煌的篇章) (1761). Additions are also detected in expressions of dates, such as “en 14 de septiembre” (9月14日) (TQ_1, ST1733), and regarding countries, like “la Europa” (欧洲) (TQ_10, ST1743). Moreover, the addition of words occurred when the students failed to grasp the meaning or the correct use of certain words, adding unnecessary expressions. For example, when interpreting “大会” [congreso] in TC_1, students committed errors, like “la primera conferencia del foro” (论坛第一次会议) (ST1749), “la primera asamblea del foro” (论坛第一次大会) (ST1751), and “el primer congreso de foro” (论坛第一次代表大会) (ST1762). In TR_3, ST1762 added the information “los líderes” (领导), which is not mentioned in the source text:

The sentence length of TR_3 probably increases the cognitive load, as well as memory and coordination effort of the interpreter. Student may therefore add words without being able to memorize all the information.

To sum up, the addition occurs mainly at the grammatical and lexical levels. The diverse textual features of sentences with high addition error rate (Add. ER) indicate that the students are prone to making such errors across various text properties. Some sentences employ a more formal and elegant language style, for which the students often lack the sufficient Spanish expressions repertoire and therefore fail to express correctly. Addition of articles in Spanish is frequent, discussed in depth in the error level section, due to the grammatical differences between Chinese and Spanish. Word additions, on the other hand, stem from insufficient Spanish linguistic schemata at both the lexical and syntactic levels, resulting in heightened cognitive load and psychological pressure. We can see that the addition is committed mainly for lack of linguistic schemata. The addition of information is rare in the corpus, probably due to, to a large extent, the form of sight translation, which includes the continuous presence of the source-language text and easier monitoring of the output.