USE OF TRANSLATION STRATEGIES IN WRITING: ADVANCED EFL STUDENTS

This qualitative study examines how advanced EFL learners use translation in writing, its perceived effects on their written product, and their feelings concerning its use. As translation is generally associated with less competent students, this research focuses on high-level ones. Involving 29 advanced EFL students from a reputable English language faculty in Indonesia, the study collected data through questionnaires using close and open-ended questions. This research demonstrated that, although employed less, translation was still a crucial means for scaffolding among the students. It was shown that creating short chunks in L1 for subsequent translation was the most common strategy while all the participants also self-reportedly practiced mental translations from L1 to L2 in various degrees. This resort to translation often seemed inevitable, but students revealed having mixed feelings towards the use of the inter-lingual strategy. This study highlights the need to develop awareness among language learners that translation is a valid tool of language learning and the necessity to teach them the skill to translate.


Introduction
The use of translation in language learning has typically been associated with less proficient learners. As they lack mastery in L2, the use of L1 through translation often serves as a natural support for them to fall back on when performing demanding tasks in L2. However, research has suggested that more competent learners also utilize translation in their L2 production (Murtisari, 2016). As L1 and L2 coexist in learners' minds, the use of translation seems to be convenient support for any learner to produce L2 when he/she needs it. Despite this, little attention has been devoted to the examination of how more high-skilled learners employ translation. Therefore, research is necessary to further investigate the role of translation among such students. The knowledge gained from such an inquiry will be invaluable for assisting them in their subsequent language learning. Consequently, this study aims to discover post-intermediate/advanced EFL students' self-reported use of translation in their writing, the perceived effects on their written output, and their feelings regarding its use. This article will be of interest to EFL writing instructors, especially for adult learners, and those who focus on translation and its application in language teaching/learning. The translation may be generally defined as an act of mediation in which meaning in one language is reproduced in another to communicate in various contexts. It is the fifth skill language learners need to learn to communicate effectively in today's global multilingual world. Translation as a learning strategy is often termed as "pedagogical translation". Focusing on the role of translation as a means to assist learners in acquiring another language, it may be defined as "using a language for understanding, remembering, or producing another language, both at the lexical level and the syntactic level from the target language into the other language" (Liao, 2006, p.194). However, frequently interpreted only as a cross-language tool of learning, the term pedagogical translation often reduces the concept of translation to a mere transfer of forms. With this in mind, the use of translation as a means of language learning should not be divorced from its functional end, which entails the ability to communicate effectively across languages (Carreres, 2014;Murtisari, 2016).
Translation has often been valued as an old-fashioned way to acquire a language and is seen to be detrimental to language learning. However, as "an extension or alternative realization of what the learner already knows" (Widdowson, 1973(Widdowson, /1979, translation may be a useful means for "building up the knowledge of the learner while building on what he has already acquired" (Titford, 1983, p. 52). The use of L1 may serve as a bridge in the transition to acquire a higher level of L2 rather than hinder it (Kosonen, Malone, & Young, 2007). Besides, as L1 and L2 are interwoven in learners' mental processing systems (Leonardi, 2010), the connection to L1 is often inevitable when learning L2. In Kern's study (1994), for instance, instructors and learners admitted that mental translation or translating silently is natural when reading L2 texts, although both often did not see it as a positive means of support. Other studies show that students resort to mental translation when the external use of L1 is suppressed. Researchers on direct composition, for instance, noted that their participants thought in L1 and translated it into L2, although they were required to write directly in L2 (Cohen & Brooks-Carson, 2001;Kobayashi & Rinnert, 1992). With this inevitability, translation should be seen as a learning resource rather than a harmful crutch.
More and more studies have shown that translation may assist learners in various ways. It can help learners enhance their vocabulary and grammar learning (Bahri & Mahadi, 2016;Murtisari, 2016;Murtisari, 2020;Prince, 1996), provide support to perform various tasks in L2 (Liao, 2006;Van Dyk, 2009), and raise awareness of L1-L2 differences (Cook, 2010;Kim, 2011;Scheffler, 2011;Titford, 1982). Working with advanced students, Machida (2011) andTitford (1982) believe that translation practice may help learners develop their L2 skills further to achieve near-native control. According to Titford, translating may serve as a tool to increase "the feeling for communicative appropriateness in the L2", which includes styles and naturalness (1982, p. 56). In concert with this, Machida contends that translation may develop students' L2 to a high level as students have to learn various aspects of the languagevocabulary and beyond -while translating to produce a desirable rendering.
In terms of writing, the use of translation as a strategy is generally associated with less able students, but more competent learners have also been observed to use translation. EFL students, regardless of their L2 proficiency, selfreportedly applied translation to keep going, to help them think more clearly, and to express more complex ideas (Murtisari, 2016). This resort to translation is natural as L2 composition is a demanding task even for more proficient students (Kobayashi & Rinnert, 1992). The language by which the knowledge is acquired may also lead to the use of translation in writing. Lay (1982), for instance, found that her Chinese respondents tended to think in L1 when writing on the topics studied or acquired in the language. Combined with translation, this strategy may help students retrieve information on a specific topic and improve the quality of writing without causing extra time (Friedlander, 1990). However, with the increase of L2 competence, the use of translation was found to decline (Hu, 2003). Hu concludes that there is a language use continuum where L2 learners "start by thinking of L2 in L1 (often through translation), and as the L2 develops, gradually think more in L2 and less in L1 [...]" (p. 59). As more direct L2-word meaning connections develop with practice, more proficient students will rely less on L1 (de Groot & Hoeks, 1995). However, other factors may affect learners' use of translation, such as learners' cognitive styles, their strategies for learning a language and language use, and their motivations for attaining competence in L2 (Cohen, 2001).
Research exclusively investigating the use of translation among advanced students in writing seems to be non-existent. Despite this, previous studies that examine translation in higher-skilled learners' composition process do not seem to show very positive findings. Examining 39 intermediate students, Cohen and Brooks-Carson (2001) found that two-thirds of their students wrote better French using direct compositions. Only one-third wrote better using translation. This supports Kobayashi and Rinnert's (1992) earlier findings suggesting that more proficient Japanese students (intermediate to low-advanced) did not benefit much from translation in English in their writing. While translation aided them with vocabulary and a variety of sentence structures, it did not significantly improve the quality of the content and organization compared to that of direct writing. They also had more unnatural forms and translation problems that did not successfully convey their original ideas.
Furthermore, Kobayashi and Rinnert (1992) discovered that more competent students believed their direct writing was better (76%) than by using translation (24%). They claimed it assisted them to create better organization, more natural language, and better grammar. Those who preferred translation, on the other hand, reported that it could give them more ideas and helped them express their thoughts more clearly. As many as 88% of the participants also believed that direct writing was easier than translation. One common reason for this was it was not easy to translate. Furthermore, studies have suggested, for instance, that the use of a translating strategy is not ideal when learners compose their L2 under time pressure (Lifang, 2008;Cohen & Brooks-Carson, 2001). Such a downside is very likely when students use translation substantially.
Although the above studies have shed more light on learners' use of translation, the findings need to be critically treated. As the study participants did not seem to be equipped with translation skills, they were bound to make errors in their cross-language mediation. Competence in two languages does not necessarily entail an ability to translate. Therefore, as Murtisari (2016) has pointed out, it is unfair to attribute issues such as "lexical choice errors" and "more awkward forms" (Kobayashi & Rinnert, 1992, p.197) to "translation". Secondly, as advanced students are equipped with more L2 resources, they are not very likely to develop an entire essay by using L1 through translation unless they have other reasons. More research needs to probe into their common practice of applying translation strategies to gain a more nuanced understanding of advanced students' use of translation. With this in mind, this survey study seeks to answer the following questions: 1. What specific translation strategies do post-intermediate/advanced EFL students use in their writing? 2. Why do they use translation in their writing? 3. How do they perceive the effects of using translation in their writing? 4. What do they feel about the use of translation in their writing?

Method
This survey-based qualitative research was carried out in an English language faculty of a reputable university in Central Java, Indonesia. English is used as the medium of instruction in over 90% of the program's courses, but students usually use Indonesian or local languages outside the classroom within the EFL setting. The faculty normally enrolls new students of elementary to upper-intermediate English proficiency, around 20% of which are expected to have reached a post-intermediate level in the third year. While high-achieving students tend to be able to learn English more effectively, it is crucial to assist them to use all the available resources they have to enhance their language skills, including writing. One potential resource such learners may benefit from for writing is the use of L1 through translation. As L2 writing is a demanding task, more research on how translation may assist high-level students to improve their language skills in English is paramount. Therefore, it is crucial to examine how students use translation in writing. However, such research is scarce.
This study involved 29 ends of third-year advanced EFL students with a GPA (grade point average) of a minimum of 3.5 (out of 4), who made up 18% of the top students in the faculty's undergraduate English programs (English Language Education and English Literature). The faculty's high-achieving EFL students were expected to reach an advanced level of language proficiency in their third year based on the ACTFL proficiency scale. The participants were recruited using a convenience sampling method based on their self-reported eligibility in all the translation classes of the third semester of the 2019/2020 academic year. Well-informed of the student's abilities, the class lecturers were involved in facilitating recruitment. This was also to make sure that all the eligible students could participate in the research.
As this study aimed to develop a general map of the advanced EFL students' tendencies regarding their attitudes on the use of translation in writing, a questionnaire was used to collect the data. The questionnaire consisted of sixteen Likert-scale statements, eleven complementary open-ended questions, and two multi-response sets. It consisted of four parts and the first section was aimed to discover whether the participants used translation strategies. The students who claimed to use written translation were required to answer the next two parts (Section 2 and 3), while those who did not were asked to answer the last two sections (Section 3 and 4). Section 2 was designed to elicit data on students' uses of specific translation strategies, their reasons for using translation, their views on its perceived effects on writing, and students' feelings. Section 3 required all the participants to answer whether they had employed L1-L2 mental translation when writing in L2 and explain their reasons. Finally, through the last section (Section 4), those who reported that they did not apply written translation were asked to write their reasons for not using the strategy. The questionnaire was piloted and underwent some changes before it was administered for the present research's data collection. In addition to the questionnaire, semi-structured written interviews were also employed to obtain additional information from several students. Although the researcher wished to involve more students in the interviews, this could not be done due to her limited access to the participants.
The main data collection was carried out at the end of the second term. The participants anonymously filled out the questionnaire in small groups or individually after class. A short briefing was given to ensure that they gave their responses truthfully and encouraged them to supply all the required information. Their answers were subsequently logged into a Microsoft Excel program and converted into percentages for analysis. Before being calculated, their responses to the open-ended questions were classified into common themes. To ensure accuracy, the researcher rechecked the logged data to avoid incorrect inputs. The responses of the students who claimed not to use written translation strategies were converted into "never" or "not applicable", as relevant to the questionnaire statements in the first part, which were included in the percentages (except for Statement 4, to which students who self-reportedly did not use a written translation method also gave their responses). This was done so that the quantities could be based on the total number of participants to enable a more holistic perspective of the results. Several participants were contacted after the survey for short written interviews through Whatsapp.

General tendencies in the use of written and mental translations
The participants' responses showed that the vast majority employed written translation in their English writing (79%), which suggested that translation was a common strategy among the advanced EFL students. Only 21% of the respondents reported that they did not implement written translation when composing their essays. The latter students' main reasons for not using it were to train themselves to think directly in English, to save time, and that it was easier to write directly in the target language. One student said it was easier to produce more natural English using a direct composition because his sources of information were mainly in English, which supports Lay's finding (1982). Furthermore, three of these students had the highest GPAs among the overall participants, but several other learners with the top GPAs reported using a written translation. This corroborates previous findings that language competence is not the only factor leading to the employment of written translation.
Despite this, 21% of the students who claimed not to use written translation indicated that they retreated to an L1-L2 mental translation at the phrase and sentence levels mainly at high frequencies. In other words, translation remains a relatively essential crutch among these students, although employed below the discourse level and not performed explicitly. All of the total participants reported that they used L1-L2 mental translation. The majority (65%) claimed to use the strategy at high frequencies (always 17%, often 48%), mostly at the phrase and sentence levels (word 28%, phrase 38%, sentence 62%, paragraph 24%). The main reasons for this were that it saved time and helped them express their ideas and improve their English. The view that translating silently may enhance one's language learning is interesting. Perhaps requiring one to work at an abstract level with more direct associations with L2, the technique was regarded as the next step from written translation.

Students' use of more specific translation strategies
It is worth noting that this section and the next three were based on the responses of the participants who reported using a written translation except for Statement 4 (S4). However, as previously mentioned, the answers of those who claimed not to use written translation were incorporated as "never" or "not applicable", as relevant to the questionnaire statements to allow for a holistic perspective of the findings. *Two students did not respond to this statement.
In terms of the use of more specific translation strategies, the advanced learners typically did not rely much on translation, but it remained an essential tool to assist them with minor problems and as a post-writing checking tool. Table  1 shows one common strategy with the most significant use, which is writing part or parts of their drafts to be translated at a later stage (52% often, 21% sometimes). Those who reported implementing this strategy said they utilized it mainly below the discourse level (word 43%, phrase 52%), but the use could extend to more than one sentence (35%). However, none of the advanced students claimed to use it at the paragraph level. This study's result is in contrast with the finding of Murtisari, Widiningrum, Branata, and Susanto (2019) of students in general of the same program, where 36% to over 40% of the respondents claimed to have used translation (Google Translate) to help them write one paragraph or more. Based on these comparable findings, advanced students appear to use much less translation at further discourse levels to assist them in writing than general students. This appears to corroborate Hu's (2003) finding that the use of translation decreases with the L2 proficiency improves.
Other strategies were significantly less utilized, but the use of L2-L1 mental translation to check what has been written is relatively common (often 17% or sometimes 45%). Unlike what the participants practiced with the previous strategy, most of them claimed to implement this silent translation at significant extents (paragraph 48% or whole draft 30%). This indicates that L1 remains a dominant cognitive tool among the advanced EFL learners in their L2 composition process. As they are still learners and their L2 is still developing, this is not surprising.
Finally, writing the whole draft in L1 and then translating it into English was the least popular, with 83% claiming never to implement the strategy. Students tended to avoid this strategy mainly because it was considered time-consuming. Having enough L2 resources to write a substantial portion of their composition directly in the language, advanced learners will have significant extra work if they translate their whole essay from L1. Although unexpectedly several students reported applying it very rarely, they also did not write favorably of the strategy because it took their time. Table 2 shows that most of the participants self-reportedly used translation for the stated reasons at significant frequencies. Students' responses revealed two of the most common causes. The first of these was to help them write complex ideas (52% always or often, 14% sometimes), which suggests the dominance of the EFL participants' L1 in processing ideas. The second most frequent reason was to keep going (41% always or often, 31% sometimes). However, as the participants indicated, using translation to continue did not necessarily mean that they could save time. A significant number of students pointed out that translating what they had written in L1 could take time (41%) and could be difficult (14%). Several students who were available for further contact believed that they had applied the strategy despite the downsides because they were compelled to do so in certain contexts, such as because of an excessive cognitive load or limited L2 linguistic resources. In other instances, they used translation because they felt more comfortable, found it easier to write their ideas in their mother tongue, and then translate it, or just were used to applying the technique. Regarding the latter, translation may not only serve as a means of scaffolding but also may become a student's preference or style in approaching his/her writing work (see Cohen, 2001). Furthermore, although fewer learners reported using translation when they feel there is a lot in their brains, a relatively significant number claimed to do this at high frequencies (7% always, 31% often) and 24% do it sometimes (See Table 2, S13). This supports previous studies where the use of L1 can provide support as a "psychological tool" to assist learners when there is "a cognitive overload" (Bruen & Kelly, 2014, p.4). It also refutes the view that the use of L1 leads to a cognitive load (Nawal, 2018). Learners may have extra work to translate the ideas subsequently, but at least they can feel that translating enables them to put something on paper.

Perceived effects of the use of translation on the writing quality
As for the effect of translation, most of the participants (around 60%) believed that translation could improve different aspects of their written output at significant frequencies except when it came to grammar (Table 3). The components of the overall writing quality and variety of vocabulary (S5 & S8) received slightly more positive responses than the content and idiomatic expressions (S6 & S7). Despite this, only around 30% of the total participants reported always or often thinking favorably of translation in their responses to the statements. In explaining their answers to S5, some students said translation made their English expressions unnatural. In general, the findings are reminiscent of Kobayashi and Rinnert's research results (1992). The study found that the more proficient students tended to prefer to write directly in L2 partly because they could have more natural language and better grammar. Regardless of this, the use of unnatural expressions in translation may either suggest a lack of knowledge of L2 idiomatic expressions or a lack of translation skills. Further research is necessary to examine this issue. The tendencies of the present study's participants to see the use of translation in writing somewhat unfavorably are inconsistent with the finding that most of them still often reverted to translation in their L2 composition (see Statement 3, Table 1). Here, as indicated by the students who were contacted for further information after the survey, they may have just been forced to employ translation to assist them to write rather than choose to apply it because it was an effective strategy for them. In other words, translation is still an essential means of scaffolding for the students. As resorting to translation is often inevitable, there is an apparent need to train advanced learners with the ability to translate. Such a skill may help them turn translation into a more effective tool of scaffolding as students evolve strategies to mediate incompatibilities across languages. Besides developing an awareness of L1-L2 differences, the translation practice may aid students to learn more specific L2 features. According to Colina and Lafford (2018), it can illuminate various pragmatic aspects of texts that will assist students to understand and create high-quality texts in L2. This is possible as translation allows an in-depth engagement with meaning, enabling learners to interact intimately with textual features.

Students' feelings towards their use of translation strategies
In terms of students' feelings, Table 4 shows that not many students (28%) reported always or often feeling positive when using the interlingual strategy and 21% just felt so occasionally. When asked about their reasons for their positive feelings, most students said they felt assisted (31%) or confident (17%). Although not many suggested a favorable feeling, fewer students claimed to frequently or occasionally feel bad when using translation to help them in writing (17% and 24% respectively). *One out of 23 students did not answer Q14 However, a closer look at the data shows that just 21%of the students favored translation without negative feelings. In comparison, the majority (61%) of the total participants were ambivalent by reporting both positive and negative feelings. While they might feel assisted in some way with their writing, they felt concerned about their translation quality or that using the strategy may have adverse effects on their language learning. The latter confirms a finding that language students tended to view translation as disrupting their language acquisition (Kern, 1994;Liao, 2006). Table 5 demonstrates examples of students' dilemmatic feelings towards translation, which were elicited by Statement 14 and Statement 15 of the present study's questionnaire. Such ambivalence was also found in Murtisari's study (2016) among students across different levels of L2 competence, but to a smaller degree. Besides reflecting students' lack of translation skills, the students' feelings seem to suggest an entrenched resistance among language learners against the use of translation in language learning. Instructors' frequent unfavorable attitudes towards translation, although often not explicitly shown, and the pressure to develop L2 fluency seems to have developed some kind of fear for its adversative effects. Often driven by misconceptions about translation, this may instead be disadvantageous to students in making the most of their natural learning resources. Therefore, it is imperative to develop awareness among students and language practitioners that the use of L1, including translation, is natural in additional language learning. It is not reasonable to expect students to complete tasks involving difficult elements in L2 when they lack mastery in the language. Rather than suppressing the use of translation, which is often unavoidable, it is better to teach students how to use it judiciously by appropriately integrating it in language teaching early on in language education. In this way, Indonesian students may reap the potential benefits of translation for their language learning.

Conclusion
This study demonstrates that translation is a crucial supporting tool among advanced EFL learners in writing, although it was employed in lesser degrees. Although advanced learners are more competent language users, they still have gaps in L2 mastery with which they need support. Therefore, as writing in L2 is a challenging task, resorting to translation seemed to be frequently inevitable as a compensatory strategy. However, they often did not seem to be able to use the strategy effectively. There also seemed to be a concern that the translation may get in the way of their language acquisition. It is, therefore, imperative to teach language learners the cross-language mediation skill and foster awareness that translation is a valid tool to support L2 acquisition. These will not only assist them to use translation strategies more successfully but also equip them with a crucial social skill for communication in real-life situations. With a small number of participants, this study is not generalizable. However, it has highlighted crucial issues surrounding the use of a translation that seems to have received very little attention in language education. To conclude, it will be more fruitful, in borrowing Cohen's (2001) words, to "[get] translation out of the closet into the open where it can be utilized more explicitly and ideally, [and] more beneficially as well" (p.105) than deny the inevitable.