Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 138 Article History: Submitted: 4 April 2022 Reviewed: 23 June 2022 Edited: 5 July 2022 Article Accepted: 8 July 2022 Metadiscourse Markers: Frequency and Errors Made by EFL Students Junaity Soften Sine1*), Rosdiana Mata2 1Institut Agama Kristen Negeri Kupang, Indonesia 2Politeknik Negeri Kupang, Indonesia *)Corresponding author email: junenighty@gmail.com DOI: https://doi.org/10.18196/ftl.v7i2.14403 Abstract To present a skillfully written abstract, an author should apply some strategies. One of them is the application of Metadiscourse Markers, which provides readers with linguistic features that will help them quickly get the gist of the entire research. Studies have examined the metadiscourse markers in abstracts of theses and published article journals written by Indonesian, but only a few used a corpus-based approach and analyzed the errors simultaneously. This study, therefore, aimed at finding the frequency of Metadiscourse Markers and grammatical errors made by Accounting Department students of Kupang State Polytechnic in their final papers’ abstracts. A corpus-based approach was applied by using AntConc Version 3.5.9 concordance tool. The study showed that interactive markers occurred more frequently than interactional resources. Moreover, grammatical errors were related to applying punctuation (comma), plural forms, verb tenses, and modality. This study suggests the teaching material includes the features of Metadiscourse Markers in abstract writing and further emphasizes the grammar on topics where students performed errors. Keywords: Metadiscourse Markers; Interactive Markers; Interactional Markers. =========================================================================== Introduction English has been acknowledged as an international language for scientific publication. Therefore, writing in English is unavoidable for those getting higher education at universities, https://journal.umy.ac.id/index.php/FTL/issue/view/903 https://doi.org/10.18196/ftl.v7i2.14403 https://crossmark.crossref.org/dialog/?doi=10.18196/ftl.v7i2.14403&domain=pdf&date_stamp=2022-07-30 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 139 even in countries where English is neither the first nor national language like Indonesia. In Indonesian universities, it is mandatory to write an abstract in both Bahasa Indonesia and English for students' final research papers. Reports have shown how Indonesian students thought writing abstracts in English was difficult (Yanto & Sulistiyo, 2017). In addition, students suggested that their English course include abstract writing techniques. Abstract functions to briefly portray the whole research; thus, it must attract readers to read more. This function means an author should effectively present their abstract. Because of this, research ideas in the abstract should be written by considering readability and acceptability in a particular discipline. There are techniques to learn to convey research information properly in academic articles, especially abstracts, such as using metadiscourse like transition words, hedges, boosters, attitude markers, or engagement markers (Doró, 2015). Moreover, accommodating ideas in a foreign language requires much effort, particularly in academic English. Foreign language learners must have large vocabulary lists to properly support their thoughts and arguments. However, it is more than knowing the words; knowing how to use them in writing accurately. A study investigated language, structure, and content challenges faced by undergraduate students in Afghanistan in writing academic articles. One finding concerning language was students’ incapability to appropriately use discourse markers and the repeated use of English words (Noori, 2020). Students could not express and support their ideas because they lacked academic phrase banks. Consequently, they repetitively used the exact words. Studies have examined academic abstracts, for example, the investigation of rhetorical moves and linguistic realization produced by English and Spanish authors in Ecuador and North America. Neither Native English nor Non-Native English researchers use complete structures of the five-move model (Hyland). Spanish authors wrote the purpose, method, and product organization, while Native English writers added an introduction to their abstract. He further discussed that different written abstracts among writers were influenced by "lingua- cultural convention practiced in each discourse community." This practice included the discourse constructions adopted and accepted by journals and academic communities (Viera, 2019). Another research investigated syntactical errors when non-native English students wrote their thesis abstracts. This study investigated how Acehnese undergraduate students organized https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 140 their thesis statement, idea, and metadiscourse markers to introduce their research papers. The findings showed that students used quasi-inductive ways to introduce thesis statements instead of inductive or deductive ones. Moreover, the way they organized their idea was not linear (circular). The last finding was intriguing because they found four rarely-produced discourse markers: comparison, persuasion, emphasis, and opinion (Novawan et al., 2020). The topic of metadiscourse applied in research articles and academic writing has been previously discussed. These researches can be categorized into several categories, such as research comparing different academic genres (research articles, theses, book reviews, and student essays), native-speaking influence, various disciplines, and other genres. For example, Kawase (2015) exposed how research article writers employed more metadiscourse markers than thesis writers, arguing the nature of research articles as a professional genre that insists on presenting oneself in the academic world. However, research article and abstract thesis writers used more interactive markers than interactional markers (Ashofteh et al., 2020; Pandey, 2020; Al-Shujairi et al., 2016; Mazidah, 2019). Some other studies reflected how native speakers and non-native speakers of English used metadiscourse markers (Hussein et al., 2018; Kaies, 2019; Nugroho, 2019). These comparative studies found that interactional markers were employed more frequently by native speakers of English. The challenge in employing metadiscourse markers is not concerned with native or non-native status but more with the language competence of the writers, as proved by Gholami et al. (2014). They confirmed the correlation between the TOEFL score and metadiscourse markers. Research contrasted and compared the production of metadiscourse markers between abstracts written by students from different disciplines and found that humanity research abstracts tended to use markers to convince rather than engage the readers (Mansouri et al., 2016). This finding was also supported by Jin and Shang (2016) and Suntara and Chokthawikit (2018), in which interactional markers used were mostly attitude markers. Moreover, there has been corpus-based research that examined the topic in various academic writing, such as book reviews (Bal-Gezegin, 2016), psychology books (Karami & Poor, 2020), opinion and sports articles in newspapers (Nugroho, 2020), introduction and conclusion sections of the master dissertations (Bouchemet, 2019) and students’ essay (Alqahtani & Abdelhalim, 2020). However, few studies have examined both corpus-based https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 141 analysis and the misuse or errors in deploying metadiscourse markers (Gholami et al., 2014), especially in the Indonesian academic context. Therefore, this study was conducted to contribute to the corpus-based analysis of metadiscourse markers found in Indonesian students' abstracts in Accounting and the textual analysis of errors they made, which was not done before. This study applied a corpus-based approach to finding the frequency of metadiscourse markers in students’ abstracts and errors made when applying metadiscourse markers. Literature Review The fact that a text is written for readers should allow any writer to think about what ideas to offer and develop them carefully. An abstract is an introductory text that gives readers the gist of the research as attractive as possible, so they decide to find out more in the whole article. This function is clarified as a screening device (Huckin, 2001). This term was expressed by Cargill and O’Connor as the only part of an article read by busy readers before they are convinced that it is essential for them to read further (as cited in Jobling, 2014). By considering the vital role of an abstract as a part of an academic text, linguistic devices are used to make the writing more effective. One of these is metadiscourse which was first coined by Harris (1959) and developed by Williams (1981), Kopple (1985), Crismore (1989), and the most current one by Hyland (2005). It refers to a strategy applied by writers to guide readers to grab the information in a text (Hyland, 2005). He further revised metadiscourse categories by proposing two dimensions, namely interactive and interactional. The interactive dimension functions to construct and organise a text that will help readers while reading. The writers anticipate the presence of other academics that go through the text with particular background knowledge. The second dimension focuses on the writer's invitation to readers to interact during the reading process. Resources in this category are deployed to share writers' opinions about their insight and ideas. Table 1 shows this metadiscourse model. https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 142 Table 1. An Interpersonal Model of Metadiscourse Source: Hyland, 2005:49 Category Function Examples Interactive Help to guide the reader through the text Resources Transitions express relations between main clauses in addition; but; thus; and Frame markers refer to discourse acts, sequences or stages finally; to conclude; my purpose is Endophoric markers refer to information in other parts of the text noted above; see Fig; in section 2 Evidential refer to information from other texts according to X; Z states Code glosses elaborate prepositional meanings namely; e.g., such as; in other words Interactional Involve the reader in the text Resources Hedges withhold commitment and open dialogue might; perhaps; possible; about Boosters emphasize certainty or close dialogue in fact; definitely; it is clear that Attitude markers express the writer's attitude to the proposition unfortunately; I agree; surprisingly Self mentions explicit reference to the author(s) I; we; my; me; our Engagement markers explicitly build a relationship with the reader consider; note; you can see that Applying the corpus-based method is considered helpful in analyzing metadiscourse markers that occur in abstracts. Corpus Linguistics is defined by McEnery and Hardie (2012) as the way to examine a collection of texts – corpora – which is stored in a computer. The corpora are composed of texts produced naturally by language users. In this research, learner corpora were used. Previously, Mocanu (2015) and Liu and Huang (2017) have conducted corpus-based studies of metadiscourse in abstracts related to economics. The first examined 130 accounting research article abstracts produced by Romann from 2006 to 2014. She confirmed her hypothesis that publication's maturity correlated to the use of interactional resources. There was an increasing number of metadiscourse elements deployed. Hedges, boosters, and attitude markers had the highest frequency. The second study concerned interactional metadiscourse employed by Chinese authors in 289 abstracts derived from prestigious economic journals in China. They found that hedges and boosters were frequently used while self-mentions and engagement markers were underused. Both studies related their findings to the cultural and historical changes which have made academics aware of their contribution to a more global scientific community. https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 143 In Indonesia, corpus-based research on metadiscourse was done by some academics, such as Hamid (2019) and Nugroho (2019). The first study aimed to determine the types of interactive metadiscourse markers used by postgraduate students in their finding and discussion section. The result showed that transition and endophoric markers were dominant because the writers first wanted to connect their arguments to create coherence in their text and refer to previously mentioned information. The following study was a contrastive analysis of Indonesian and American students applying metadiscourse features in their undergraduate thesis abstracts. Using Crismore et al.'s model (1993), his findings showed that Indonesian used more markers than Americans; however, the person marker is most frequent. In deploying metadiscourse markers, it is assumed that students may well commit errors. At this point, it is necessary to identify students' errors, categorize them, and provide suggestions for more effective language teaching. This process is called Errors Analysis (EA). The final purpose of doing EA is to evaluate and treat errors (Yaghi & Abdullah, 2015). They further explained that the classification of errors had been developed over time and one of the current taxonomies proposed by Otchi (2005) in Yaghi and Abdullah (2015). He classified grammatical errors into five main errors: verb errors, noun ending errors, article errors, wrong words, and sentence structure. The research on EA in abstracts written by Indonesian students was done by Maria (2020), Fitria (2020), and Madkur (2013). The first study pointed out that Word Order was the most frequent error committed by students. She further proposed two solutions: applying an indirect method to teach grammar and prioritizing to teach grammar structures. The latter studies found that errors made in English translation abstracts were related to three aspects of writing: grammar, spelling, and punctuation. Another interesting research was conducted by Sharif and Hassani (2016), who analyzed errors made by Iranian writers in their English translation abstracts. The most frequent errors made were in articles and verb agreement. Method This study relied on a quantitative approach followed by a qualitative analysis. A corpus-based method was applied in this study and followed by textual analysis. As explained previously, corpus-based is a methodology in which language is stored in a computer to be https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 144 processed by a concordance tool. In the process of analysis, concordance tools such as WordSmith (Scott, 2013), AntConc (Anthony, 2005), ConcGram (Greaves, 2009), Sketch Engine (Kilgarriff, Rychly, Smrz & Tugwell, 2004) can be applied (Ngula, 2018). To answer the research questions concerning occurrences and errors students made in employing Metadiscourse markers, a concordance tool, AntConc Version 3.5.9 (Anthony, 2020), was used. The application is free to download and, therefore, more familiar to the researchers to be applied. Furthermore, the texts were then analyzed to describe the errors made by students. Learner corpus is collected from students who write texts for their language course activities (Hunston, 2002). In this study, the data source was 50 English abstracts written by students of Kupang State Polytechnic in East Nusa Tenggara Province. These abstracts were collected as learner corpus for this study. The population in the Accounting Department was 170 students. The sample was 50 students for the academic year of 2020 using a simple random sampling technique. This learner corpus contained 17149-word tokens with 1595 word types. To collect the data, two-step procedures were taken. First, the researchers collected the students' abstracts from Kupang State Polytechnique. The next was converting the texts to .txt format, which is acceptable for AntConc Version 3.5.9. The metadiscourse categorization used was Hyland's (2005), which includes interactive and interactional features. This model was used because it has been the latest revision of the proposed metadiscourse model. A keyword list was presented in analyzing the data to find metadiscourse resources in the corpora. After the categorization, the words were examined whether or not students made errors in applying them using concordance and collocates tools. Another English lecturer reevaluated the findings to ensure validity. Table 2. Demographic Information of Students Faculty/Study Program Male Students Female Students Total Accounting 8 42 50 https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 145 This study used corpora – texts collected from students in Accounting Department, Politeknik Negeri Kupang. The female students in this department outnumbered male students, influencing the number of abstracts taken. Findings Frequency of Metadiscourse Markers The analysis showed that transition and frame markers were students' most frequent interactive metadiscourse resources. Table 3 presents the frequency of each feature of the interactive dimension. Table 3. Frequency of the Interactive Metadiscourse in students’ abstracts Transitions Addition and 385 in addition 2 furthermore 1 Comparison But 9 however 5 In contrast 2 while 9 whereas 1 meanwhile 8 Consequence due to 10 because 30 So 27 therefore 3 nevertheless 3 thus 4 Total 15 499 Frame markers Sequence parts third 1 then 5 at the same time 1 Label text stages concluded 15 concludes 2 conclusion 4 result 3 results 85 Announce discourse goals aim 1 aimed 1 aims 33 purpose 8 Indicate topic shifts Total 12 159 Endophoric markers Evidential according 8 Code glosses namely 28 such as 7 Total 2 35 Total Interactive Markers 30 701 https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 146 Transition markers performed most frequently, with 499 occurrences in the texts. In this feature, additive resources were dominant, with 385 appearances. Whereas markers used to show causative relations came next with 77 occurrences. After that, 34 times, students used markers that function to compare ideas. The frame markers ranked second with 159 hits, and 109 occurrences of ‘conclude’ and ‘result’ indicated passage stages. The next feature was the application of markers to announce the goals of a text with 43 hits. Moreover, the frequency of features that function to sequence text parts ranked last with seven occurrences. No data was found that indicated shifts in topics. The last two features were code glosses and evidential. Features of code glosses ranked third with 35 appearances in students’ corpora. ‘Namely’ was deployed more frequently (28 hits) than ‘such as’ (7 hits) to elaborate information. Evidential marker, ‘according’, hit eight occurrences in the corpora. There were not any endophoric markers found in the students’ text collection. Table 4 presents 209 hits of interactional metadiscourse markers and shows that Hedges and Attitude Markers are the two most frequent resources deployed in students’ abstracts. There were nine hedges with 74 occurrences on students' abstracts. Can was the most frequent, followed by ‘indicate’ with 16 occurrences. Most students deployed epistemic verbs – verbs showing how one is confident about an assertion – such as ‘could’, ‘would’, ‘need’, and ‘must’, while probability adverbs (‘generally’) and epistemic expression (‘most’) were the slightest variations of hedges used by students. The second most frequent interactional marker was attitude markers, and 60 occurrences with eight words were found. The most frequent resources were ‘good’, ‘consider’, and ‘well’. On the other hand, ‘appear’, ‘suggestion’, ‘important’, ‘appropriate’, and ‘better’ were less frequent markers presented on the corpora. Generally, students were more familiar with using the variation of attitudinal adjectives compared to attitude verbs or adverbs. Engagement markers came next, with 28 occurrences in the text collection. ‘Seen’ and ‘consider’ were used by students more than other resources. Self-mentions and boosters were the least frequent markers deployed by students. https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 147 Table 4. Frequency of the Interactional Metadiscourse in students’ abstracts Hedges can 43 could 1 would 1 need 2 must 3 indicate, indicates 16, 1 generally 1 most 6 Total 9 74 Boosters always 2 in fact 1 evidenced 1 Prove, proved, proven, proves 2, 1, 1, 2 find, found 5, 1 Total 9 16 Attitude markers considered, considering 12, 1 appear 1 suggestion, suggestions 2, 4 important 1 appropriate 2 good 25 well 11 better 1 Total 10 60 Self mentions I 1 researcher, researchers 8, 6 writer 3 author, authors 3, 2 Total 6 23 Engagement markers seen 14 considered, considering 12, 1 need to 1 Total 4 28 Total Interactional Markers 37 206 Grammatical Errors in Applying Metadiscourse Markers Interactive markers Students made errors in deploying Transition and Frame markers. The examples are shown in the table 5. The problem students faced in writing transition marker resources generally dealt with punctuation. For example, there were 27 hits in the corpora, with six commas before ‘so’, five wrote ‘so’ at the beginning of a sentence, and 16 did not use commas. Another case found was about the verb agreement. They wrote ‘it because’ instead of ‘it is because’. Students overgeneralized ‘due to’ as the synonym of ‘because’; therefore, they wrote it at the beginning of the sentence. https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 148 Table 5. Errors in Using Transition Markers No. Error Correction 1. “It would be useful in the long term and also pay attention to expenditure efficiency in order to maintain savings but also pay attention to any program realisation." (PNK 5) ‘And’ should connect parallel words (e.g., noun and noun, not a noun and verb) ‘but also’ is used when there ‘is not only’ previously used. 2. “Due to the implementation stage, there are still activities that are not carried out as determined.” (PNK 12) ‘Due to’ is similar to ‘caused by’, so it cannot be used at the beginning of a sentence. 3. “… statement, because” (PNK 46) “… is considered very poor, because the …” (PNK 41) “… maximised, because there are …” (PNK 19) ‘Because’ functions to connect two clauses, so there should be no comma here. 4. “Still, for Dinas Pendidikan dan Kebudayaan Kota Kupang, they need to pay more attention to the allocation of expenditure funds so as not to cause significant variance and to the harmony of spending in order to increase capital funds.” ( PNK 5) ‘So’ is used to connect two independent clauses, so it should be written with a comma before. 5. “It because all activities have been able to be realised according to the predetermined targets." (PNK 32) ‘It is because’, not ‘it because’, to introduce cause and effect. 6. “… on time. Therefore it is hoped that …” (PNK 27) Use a comma after ‘therefore’ to connect two independent clauses. 7. “Thus it is necessary to improve …” (PNK 49) Put a comma after ‘thus’. When students presented rhetorical moves – introduction, purpose, methodology, result, and conclusion – in an abstract, it seemed difficult to differ when using the present or past tense, as seen in the table 6. https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 149 Table 6. Errors in Using Frame Markers No. Error Correction 1. “From the research conducted, it was concluded that the accountability of direct cash assistance in the Maulafa village was quite …” (PNK 10) A conclusion should be written in the present tense, not the past tense. 2. “This study's results indicate that based on the t- test …” (PNK 37) The result of the study should be presented in past form. 3. “This study aims to analyse and understand the accounting information system's application for …" (33 hits) To state, the purpose of the research should be in the past tense. 4. “The purpose of this study is to determine …” (PNK 2) To state, the purpose of the research should be in the past tense. The structure of an abstract is constructed using frame markers such as label stages and present goals in the texts. According to the abstract writing convention, a conclusion should be in the present tense, while stating the purpose of the study and presenting results should be written in the past form. Interactional markers The markers applied to open an interaction with readers are called interactional markers. A hedge is one of such markers that shows how a writer proposes ideas where they are not fully committed or open to alternatives. Hedges were also problematic to students. Table 4 clearly shows the problems. The first problem concerns the degree to which a writer is confident about their statements or opinions. Students could not differentiate between' must' and ‘should’. Secondly, they use present tense to state research goals and results, usually in past form. Finally, students also made an error in mentioning themselves in plural form. https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 150 Table 7. Errors in Using Hedges No. Error Correction 1. “… the Rote Ndao district government must maintain its expenditure efficiency in the future.” (PNK 2) ‘Must’ is used to express “unavoidable obligation,” whereas ‘should’ express a recommendation regarding someone’s responsibility or duty. 2. “The results of this study indicate that local taxes …” (PNK 31) Mostly, ‘indicate’ was used by students to report the result of the study. Therefore, past tense must be used (‘indicated’). 3. “From the results of this study, the hypothesis testing and determinant coefficient test, prove that there is …” (PNK 44) Past tense is applied when the author wants to report the result of the study. This sentence should also not have a comma before ‘prove’. 4. “The purpose of this study is to determine …” (PNK 2) Stating the purpose of the research should be in the past tense. 5. “Suggestions from researchers are that it is hoped that PT. ASDP Kupang Branch …” (PNK 4) It should be ‘researcher’ or ‘author’ without –s because it is not a plural form. “Researchers limit the problem to the Analysis of the Effectiveness and Efficiency Implementation Keluarga Harapan Program in Lasiana Village.” (PNK 27) Discussions An abstract is the first part that will impress readers before they decide whether to read the article further or not. The clarity and well-structured abstracts may well create the impression. It means the writer can present their research by employing linguistic features considered beneficial to meet these requirements. Metadiscourse Markers function to serve this purpose effectively. https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 151 The Interactive Metadiscourse Markers This study found that the interactive metadiscourse markers were more frequently used than the Interactional markers in students' abstracts (see Figure 1). Consistent with those of Al- Shujairi et al. (2016) and other previous studies mentioned earlier, the more frequent appearance of the interactive instances indicated that the writers consider the readers, so they organize the flow of the information efficiently. One possible reason for such a frequent occurrence is the characteristics of these markers to link ideas and, simultaneously, help readers follow the research's brief step-to-step process. Figure 1. Percentage of metadiscourse markers The students, of course, wanted to make their abstract apparent and easy to interpret. Consequently, they mainly deployed Transition and Frame markers. Transition markers were dominant in writing abstracts because they add, compare, and show causative relations in texts. Students applied and showed an addition in their abstracts. Only a few used the words ‘in addition’ to show additive function in texts. To compare ideas, they used lexical variations such as ‘however’, ‘in contrast’, ‘while’, ‘whereas’, and ‘meanwhile’. Students frequently applied ‘because’ and ‘so’ for causative relations. Some others used ‘due to’, ‘therefore’, ‘nevertheless’, and ‘thus’. It is entirely various. The most frequent resources were ‘and’, ‘because’, and ‘so’. This result seems consistent with the previous research that found extensive transition markers in abstract writing (Mazidah, 2019; Hamid, 2019). https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 152 As previously mentioned, frame markers are used to set the structure of an abstract. Features of frame markers that indicate stages and introduce discourse goals were the most occurring ones. Both features are essential for an abstract because the nature of abstracts gives the main ideas of the whole research. ‘Conclude’, ‘result’, ‘aim’, and ‘purpose’ were words deployed by students to mark stages and state goals. The least occurring markers were code glosses and evidential. Code glosses provide readers with detailed information or elaborate on the writer's ideas. Students preferred to use ‘namely’ over ‘such as’ to give additional information in their abstracts. Furthermore, ‘according’ was the only word used to support their opinion. This result may be explained by the fact that the word count for an abstract usually ranges from 150 to 250 words, which makes elaborating and quoting ideas unnecessary in abstract writing. The Interactional Metadiscourse Markers The Interactional dimension is said to function to express the writer's perspective and provide space for readers' knowledge. Students used hedges more than boosters. This finding is in line with a previous study indicating that students rarely used boosters (Suntara & Chokthawikit, 2018). The study, which examined 60 abstracts written in Public Health Journals in Thailand, clarifies that if the frequency of boosters is low, hedges might have been used more frequently. They further found that students used hedges mostly in conclusion moves; only a few used them in the background and objective moves caused by the strategy to let readers be involved in interpreting the result, which is still open for discussion. Using hedges implies that they invite readers to consider their view and are still open to alternatives while using boosters means closing down alternatives. ‘Can’ is the most occurring interactional marker used mainly by students to compose conclusions and results. Meanwhile, they used boosters – ‘always’ and ‘in fact’ – to provide a convincing background of the study and ‘prove’, ‘evidenced’, and ‘found’ to highlight certainty in the result of their study. Hyland believed that providing a balance portion of hedges and boosters is essential to show commitment to an assertion and still give readers a vital position in the interaction process while reading (Rodina, 2007). https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 153 Other frequent interactional markers were attitude markers. Out of ten words portraying students' attitudes towards opinions, they mainly used four attitude adjectives. ‘Good’ and ‘well’ were frequent of all features in Attitude Markers. ‘Good’ was frequently written with the word ‘category’ in the corpora showing practical expression. Meanwhile, the attitude adverb ‘well’ collocates with ‘carried out’, ‘running’, and ‘planned’ to describe the quality of the actions. Although features of Self Mentions were dominated by third-person pronouns such as ‘researcher’, ‘writer’, and ‘author’, the appearance of ‘I’ once in students' corpora is an intriguing finding. “As a researcher, I collect the data using interviews, documentation, and literature study.” (PNK 41) First-person pronouns in most academic writing in Indonesia are not allowed due to the issue of subjectivity in a scientific field, including in Politeknik Negeri Kupang. Most students use a third-person pronoun to mention themselves in their abstracts to meet this academic convention. Another feature in the interactional dimension is the engagement marker. When writers want to drag their readers to the thinking process, which leads them to the writer's interpretation, they use directives such as ‘see’, ‘note’, and ‘consider’ or obligation models such as ‘should’, ‘must’, and ‘have to’. In this study, ‘seen’, ‘considered’, ‘considering’, and ‘need to’ were found to be applied by students. Mocanu (2015) and Liu and Huang (2017), in their studies, suggested the importance of interactional metadiscourse markers. They argued that the quality and maturity of published research could be seen from the increasing use of hedges, boosters, and attitude markers. Unfortunately, these markers are not easy to employ by EFL students. Students are merely stating their rhetorical moves without reflecting their insight and opinion. The Grammatical Errors in Applying Metadiscourse Markers Word order, concord, parallel structure, learner languages, and spelling were five grammatical errors examined and found in students’ abstracts (Madkur, 2013). He further https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 154 explained how language interference was also a phenomenon in his study, stating that a learner's foreign language portrayed their mother tongue. This study also examined concord, parallel structure, and learner language errors. Gholami et al. (2014) mentioned this misuse as interlingua errors in students' argumentative essays when applying metadiscourse markers. In applying transition markers, grammatical errors are mainly related to punctuation (commas) in sentences. A comma should connect two independent clauses, such as ‘so’. Students consistently miswrote commas. It may well be due to the influence of the function of a comma in Bahasa Indonesia. A possible reason for this might be that students translated their abstracts from Bahasa Indonesia directly to English without considering the grammatical rule in English. This finding confirmed the previous finding of Fitria (2020), where students made errors concerning punctuation aspects in 29% of the abstracts. She also clarified the importance of mastering English grammar rules to produce high-quality English abstracts. The principle of modality is a problem students face. So far, students used ‘can’, ‘could’, and ‘would’ correctly. For example, ‘can’ is for presenting ability, ‘could’ is used to say that something can happen, and ‘would’ is used to describe something that certainly will happen in an imagined situation. The problem they found was using ‘must’ in a context where they can only recommend something to be executed, not an obligation. For example, when discussing what the government ‘must’ or ‘should’ do. In Bahasa Indonesia, the word harus is translated as ‘must’, so students were perplexed and used ‘must’ instead of ‘should’. In using frame markers resources, they made errors in applying tenses. For example, " ‘concluded’, " which indicates past form, " is incorrect because the writing conclusion should be in the present tense. In addition, the research purpose should be written in the past tense. Another highlight is applying a plural form to mention themselves as researchers (plural form), not a singular person in charge of their final research paper. In line with Sharif and Hassani (2016), the translated version of the EFL context's abstract should be carefully composed. They found grammatical errors in publishing journals in Iranian universities were dominantly related to the use of articles and lack of concord or agreement. Despite the difficulties in mastering, teaching English grammar should cover the most frequent errors. https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 155 Conclusion and Implication Metadiscourse markers are linguistic features that help authors to compose well- structured academic writing. Since abstracts are the first part of any scientific article to be read, it is essential to present clear and remarkable abstracts. Resources of Metadiscourse markers function to guide readers to find the entire research article. Interactive markers technically function to guide readers in a step-by-step process of the research abstract. Consequently, the interactive markers become more frequent than interactional markers. However, the interactional markers can depict the quality of the academic writing because they provide the author’s opinion, thought, and evaluation which are highly considered in the academic community. Moreover, grammatical errors highlighted the need to add extra attention to the abstract structure with details on punctuation, tenses, plural form, and modality. This study encourages English lecturers at universities to include the teaching of Metadiscourse Markers and their application to produce a better abstract. The future study may contribute by adding interviews with students and lecturers to enrich the comprehension related to pedagogical issues. References Al-Shujairi, Y. B. J., Ya’u, M. S., & Buba, J. A. (2016). Role of moves, tenses, and metadiscourse in the abstract of an acceptable research article. Mediterranean Journal of Social Sciences, 7(2), 379–386. https://doi.org/10.5901/mjss.2016.v7n2s1p379 Alqahtani, S. N., & Abdelhalim, S. M. (2020). Gender-based study of interactive metadiscourse markers in efl academic writing. Theory and Practice in Language Studies, 10(10), 1315– 1325. https://doi.org/10.17507/tpls.1010.20 Anthony, L. (2020). AntConc (Version 3.5.9). Waseda University. https://www.laurenceanthony.net/software Ashofteh, Z., Shirvan, M. E., & Golparvar, S. E. (2020). The move structure of abstracts in applied linguistics research articles in light of the distribution and functions of metadiscourse markers. Journal of Language and Linguistic Studies, 16(4), 2077–2096. https://doi.org/10.17263/JLLS.851035 https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 156 Bal-Gezegin, B. (2016). A corpus-based investigation of metadiscourse in academic book reviews. Procedia - Social and Behavioral Sciences, 232(April), 713–718. https://doi.org/10.1016/j.sbspro.2016.10.097 Bouchemet, T. (2019). Interactional metadiscourse in applied linguistics master theses: A corpus-based comparative study. 57 ,اإلنسانية العلوم مجلة. https://doi.org/10.34174/0079-000- 051-027 Crismore, A. (1989). Talking with Readers: Metadiscourse as Rhetorical Act. Peter Lang. Doró, K. (2015). The rhetoric structure of research article abstracts in English studies journals. Prague Journal of English Studies, 2(1), 119–139. https://doi.org/10.2478/pjes-2014-0013 Fitria, T. N. (2020). Error analysis of English abstract in International Journal of Economics, Business and Accounting Research (IJEBAR). Al-Lisan, 5(2), 164–181. https://doi.org/10.30603/al.v6i2.1335 Gholami, J., Nejad, S. R., & Pour, J. L. (2014). Metadiscourse markers misuses; A study of EFL learners’ argumentative essays. Procedia - Social and Behavioral Sciences, 98, 580–589. https://doi.org/10.1016/j.sbspro.2014.03.454 Hamid, S. F. (2019). Interactive Metadiscourse in Finding and Discussion Section of Theses at English Department Universitas Negeri Jakarta. 254(Conaplin 2018), 564–567. https://doi.org/10.2991/conaplin-18.2019.320 Huckin, T. (2001). Abstracting from abstracts. In M. Hewings (Ed.), Academic Writing in Context (pp. 93–103). University of Birmingham Press. Hunston, S. (2002). Corpora in Applied Linguistics. Cambridge University Press. Hussein, K. A. A., Khalil, J. A., & Abbas, N. F. (2018). Metadiscourse markers in master thesis abstracts of American and Iraqi English Theses. Arab World English Journal, 9(4), 347–360. https://doi.org/10.24093/awej/vol9no4.26 Hyland, K. (2005). Metadiscourse: Exploring Interaction in Writing. Continuum. Jin, X., & Shang, Y. (2016). Analyzing metadiscourse in the english abstracts of ba theses. Journal of Language Teaching and Research, 7(1), 210–215. https://doi.org/10.17507/jltr.0701.24 Jobling, M. (2014). [Review of the book Writing scientific research articles: Strategy and steps (2nd edition by M. Cargill and P. O’Connor:). Aquaculture International, 22(2), 963–964. https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 157 https://doi.org/10.1007/s10499-013-9689-7 Kaies, S. (2019). A corpus-based comparative study of the use of interactional metadiscourse markers in master theses general conclusions: The case of postgraduates of oum el bouaghi and iowa universities (University of Larbi Ben M’Hidi OEB). Karami, Y., & Poor, M. L. (2020). a comparative corpus-based analysis of using hedging, boostering, and self-mentioning metadiscourse markers in persian and English psychology books. International Journal of Research in English Education, 5(2), 13-28. Kawase, T. (2015). Metadiscourse in the introductions of PhD theses and research articles. Journal of English for Academic Purposes, 20, 114–124. https://doi.org/10.1016/j.jeap.2015.08.006 Kopple, W. J. V. (1985). Some exploratory discourse on metadiscourse. College Composition and Communication, 36, 82–93. Liu, P., & Huang, X. (2017). A study of interactional metadiscourse in English abstracts of Chinese economics research articles. Higher Education Studies, 7(3), 25. https://doi.org/10.5539/hes.v7n3p25 Madkur, A. (2013). Grammatical analysis on the abstract of scientific writing. Pedagogy: Journal of English Language Teaching, 1(1), 52–67. Mansouri, S., Najafabadi, M. M., & Boroujeni, S. S. (2016). Metadiscourse in research article abstracts: A cross-lingual and disciplinary investigation. Journal of Applied Linguistics and Language Research, 3(4), 296–307. www.jallr.com Maria, A. D. (2020). Grammatical errors analysis in writing english abstract of final report at Air Conditioning and Refrigeration Programme Study Of Politeknik Sekayu. Esteem Journal of English Education Study Programme, 2(1), 77. https://doi.org/10.31851/esteem.v3i2.4447 Mazidah, E. N. (2019). A comparison of the interactive metadiscourse in the abstracts of articles written by Indonesian and NES Scholars. Etnolingual, 3(1), 57–74. McEnery, T., & Hardie, A. (2012). Corpus linguistics. In Library of Congress Cataloguing in Publication Data (Vol. 53, Issue 9). Mocanu, M. (2015). An empirical analysis of metadiscourse in the abstracts of Romanian accounting research articles. Contabilitate Şi Informatică de Gestiune, 14(2), 362–377. https://journal.umy.ac.id/index.php/FTL/issue/view/903 https://ijreeonline.com/article-1-275-en.pdf https://ijreeonline.com/article-1-275-en.pdf https://ijreeonline.com/article-1-275-en.pdf Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 158 Ngula, R. (2018). Corpus method in language studies. In Perspectives on Conducting and Reporting research in the Humanities (pp. 205–223). University of Cape Coast Press. Noori, A. (2020). An Investigation of Afghan undergraduate English major students’ academic writing difficulties. Journal of Foreign Language Teaching and Learning, 5(2). https://doi.org/10.18196/ftl.5249 Novawan, A., Aisyiyah, S., Miqawati, A. H., Wijayanti, F., & Indrastana, N. S. (2020). Exploring the teachers' perspective on morality in an English as a foreign language pedagogy. JER| Journal of ELT Research. 5(1), 80–93. https://doi.org/10.22236/JER Nugroho, A. (2019). Exploring metadiscourse use in thesis abstracts: A cross-cultural study. Journal of English Language and Culture, 9(2), 113–127. https://doi.org/10.30813/jelc.v9i2.1689 Nugroho, A. (2020). Investigating the use of metadiscourse markers by American and Indonesian writers in opinion and business articles. Journal of English Language and Culture, 10(2), 75–87. https://doi.org/10.30813/jelc.v10i2.2138 Pandey, G. P. (2020). Metadiscourse use in thesis abstracts : A case of M . Ed. English. Nepal Journal of Multidisciplinary Research (NJMR), 3(2), 1–17. https://doi.org/https://doi.org/10.3126/njmr.v3i2.33019 Rodina, H. (2007). Metadiscourse: Exploring interaction in writing by HYLAND, KEN. In The Modern Language Journal (Vol. 91, Issue 3). https://doi.org/10.1111/j.1540- 4781.2007.00593_9.x Sharif, F. D., & Hassani, M. (2016). Error analysis of the grammatical patterns in the English translation of the abstracts of the persian articles in the iranian published academic journals. ELT Voices- International Journal for Teachers of English 7(1), 1–10. Suntara, W., & Chokthawikit, S. (2018). Interactional metadiscourse in research article abstracts: An analysis from public health journals. Language and Linguistics, 36, 31–52. Retrieved from https://so04.tci-thaijo.org/index.php/joling/article/view/154294 Viera, R. T. (2019). Analysis of abstracts in scientific papers written in English using corpora. Script Journal: Journal of Linguistic and English Teaching, 4(2), 112. https://doi.org/10.24903/sj.v4i2.323 Williams, J. (1981). Style: Ten lessons in clarity and grace. Scott Foresman. https://journal.umy.ac.id/index.php/FTL/issue/view/903 Journal of Foreign Language Teaching and Learning Volume 7, No. 2, July 2022 Available online at: https://journal.umy.ac.id/index.php/FTL/issue/view/903 e-ISSN: 2580-2070, p-ISSN: 2527-7650 159 Yaghi, E., & Abdullah, A. (2015). Review of Learners' Errors: an Error Analysis Perspective. https://www.researchgate.net/publication/312595230 Yanto, & Sulistiyo, U. (2017). Students ’ difficulties in writing their thesis in English at English education study program Jambi University. International Seminar and Annual Meeting BKS- PTN Wilayah Barat, 763–765. https://journal.umy.ac.id/index.php/FTL/issue/view/903