The particular dialect or language that a person chooses to use on any occasion is called a code 1583 Copyright © 2022 The Author IDEAS is licensed under CC-BY-SA 4.0 License Issued by English study program of IAIN Palopo IDEAS Journal of Language Teaching and Learning, Linguistics and Literature ISSN 2338-4778 (Print) ISSN 2548-4192 (Online) Volume 10, Number 2, December 2022 pp. 1583 - 1600 Linking Adverbials in Indonesian EFL Students’ Essays: A Corpus-Driven Study Ikmi Nur Oktavianti1, Japen Sarage2 ikmi.oktavianti@pbi.uad.ac.id 1, 2 English Education Department, Universitas Ahmad Dahlan, Yogyakarta Received: 2022-10-12 Accepted: 2022-12-31 DOI: 10.24256/ideas.v10i2.3061 Abstract Writing is the activity of arranging and organizing ideas that need some logical connectors to make the ideas cohesively structured. There are several plausible means for cohesiveness, including linking adverbials (LAs). This study examines the LA categories found in EFL students’ academic essays and explores how EL students use them. The data were collected from students’ essays compiled as a learner corpus, Learner Corpus of Academic Writing, comprising 52,404 words. The present study employed LancsBox as the corpus tool and The Louvain Corpus of Native English Essays as the control corpus to assist the analysis. The study results show that all primary categories of LAs are found in the corpus. However, the subcategory transition to another topic is absent from the corpus. In the frequency of use, causal resultative LAs outnumber all other categories, followed by additive and adversative LAs. As for the least frequent LAs, the study identified sequential types. In comparison to the native writer corpus, it is found that there are shared similarities, including the categories and the subcategories of LAs found in both corpora. Regarding the frequency of use, the most frequent types are relatively similar with different positions, namely causal resultative and additive LAs, while the least frequent type is sequential. These results suggest more considerations in designing writing materials, especially in regard to transition markers belonging to the sequential category. Keywords: academic writing; corpus; linking adverbials; EFL Introduction When learning a new language, one of the important skills to master is writing. Several aspects should be considered to constitute a text, including the semantic unity of the ideas conveyed in the text, so textual continuity is created or called cohesion (Flowerdew & Mahlberg, 2007; Grisot, 2018; Halliday & Hasan, 2014; http://u.lipi.go.id/1457703302 Ikmi Nur Oktavianti, Japen Sarage Linking Adverbials in Indonesian EFL Students’ Essays: A Corpus-Driven Study 1584 Yang & Sun, 2012). For creating textual cohesion, alongside coordinators and subordinators, linking adverbials are essential since they indicate the relation between parts of the text (Biber et al., 2021; Lei, 2012; Yin, 2014). Linking adverbials (LAs) connect clauses, sentences, and paragraphs (beyond clauses), while coordinators and subordinators link clauses within sentences (Biber et al., 2021; Liu, 2008). According to Liu (2008), LAs can mark several different semantic relationships of discourse units, such as additive (that is, likewise), adversative (however, on the other hand), causal/resultative (as a result, consequently), and sequential (first, meanwhile). This four-way classification system is crucial since it represents the fundamental textual relations and enables the writing to function as a text (Halliday & Hasan, 2014). Liu (2008) underlined two significant reasons for focusing on LAs: (1) the importance of LAs in language use and learning and (2) the inadequate description of LAs in grammar books and textbooks. Thus, LAs are salient features that should be considered in writing (Biber et al., 2021; Lei, 2012; Liu, 2008). The importance of LAs has also been a part of the discussion in second and foreign language writing (Gao, 2016; Lei, 2012). Previous studies have demonstrated that LAs are crucial in second and foreign-language writing because they help learners build logical links of different units in a text (Ahmad & Way, 2020; Gao, 2016; Leedham & Cai, 2013; Lei, 2012; Wang, 2022). Furthermore, different disciplines or registers might rely on LAs distinctively (Yin, 2016). However, these studies showed that second and foreign-language learners found some difficulty in using LAs. Some studies proved that ESL and EFL learners overused or underused particular LAs in comparison to native writing (Ishikawa, 2010; Lei, 2012; Nizar & Munawaroh, 2019), which can weaken their writing (Ahmad & Way, 2020). In addition, there was also register-inappropriate use, e.g., using colloquial LAs in the academic context (Liu, 2008). Therefore, this study explores the use of LAs in students’ writing and compares it to native writing to find out the discrepancies in LAs usage. Previous studies on LAs have been focused on written and spoken academic contexts. In written academics, those studies can be categorized into the studies of native writing, ESL learners writing, EFL learners writing, the comparison between ESL/EFL learners to native writing, and the comparison of novice academic writers to professional academic writers. To begin with, some previous research on LAs in written academic English focused on single-language data, i.e., native English/ESL/EFL. In the native English context, Liu (2008) reported the findings of the investigation of LAs in five registers (spoken, academic, fiction, news, and others), yield variations in the distribution of LAs. Overall, LAs in fiction are less frequent than in speaking and academic writing, and the news shows the least use of LAs. Similarly, Pipatanusorn & Wijitsopon (2019) explored the use of LAs in native English writers of various genres in present-day English. This study claimed that LAs occur most frequently in academic writing and least frequently in fiction and media texts, with additive LAs occurring most frequently across text types. The results IDEAS, Vol. 10, No. 2, December 2022 ISSN 2338-4778 (Print) ISSN 2548-4192 (Online) 1585 emphasize the distinct characteristics of each text type or genre. Regarding academic prose, Techarsatitwong (2015) examined LAs of results in written academic English as compiled in COCA and showed that the most common LA is thus, followed by therefore, and hence respectively. This study also demonstrated the use of so, which is frequent in written academic English in the middle position. Some research has also focused on the discrepancies in LAs usage in the ESL and EFL context. Regarding the ESL context, Appel & Szeib (2018) studied the use of LAs among learners from various L1 backgrounds, which showed that L1 Arabic writers overuse additive LAsadverbials, contrastive LAs are overused by L1 Chinese writers, and appositional LAs are overused by L1 French writers. This study indicates that each linguistic background might have its consequence aligned with the use of LAs. As for the EFL context, Leedham & Cai (2013) showed that Chinese students dominantly use certain LAs, e.g., besides, on the other hand, and they tend to disregard informality and use LAs in the sentence-initial position. Earlier research on LAs also compares the use by native speakers and non-native speakers of English. In the ESL context, Dutra et al. (2019) examined the use of LAs of Brazilian ESL learners and English native writers and claimed the differences between them concerning the syntactic position and meaning of the LAs. Whereas in the EFL context, previous studies identified the mismatches of LAs use of native and EFL learners. Ishikawa (2010) reported the discrepancies in LAs usage among native and Asian non-native writers in terms of the quantity of essay writing and the tendency to overuse additive LAs by Asian non-native writers. Similarly, Liu (2013) found that Chinese EFL learners overuse LAs in their speaking and writing, though they use LAs differently in those registers. The study also revealed a higher frequency of adverbials in speaking than in writing. Furthermore, Jamil et al. (2014) proved that L1 Pakistani writers use additive, summative, and contrastive LAs compared to English native writers. Later, Ha (2016) showed differences in LA use between English native writers and L1 Korean writers since L1 Korean writers overused the sequential and additive categories. In the Asian context, Nizar & Munawaroh (2019) showed a relatively similar distribution of LAs between native writers and Indonesian EFL learners. However, the latter overused additive LAs and underused causal LAs. Like Nizar & Munawaroh, Ahmad & Way (2020) demonstrated that Malaysian learners only utilized a limited set of LAs in their essays and tended to overuse additive LAs. On the contrary, Lumbangaol (2022) reported a lack of LAs in students’ writing due to their low writing proficiency in English. Similarly, in the Japanese context, Nakayama (2021) claimed that Japanese learners tend to overuse sentence-initial LAs and underuse sentence-medial/final LAs. The comparison of English native use of LAs to non- native writers in the Asian context was also made for textbooks. Phoocharoensil (2017) revealed that the LAs used in the Thailand EFL textbooks differ from those in the corpora, e.g., Corpus of Contemporary American English. LAs in the corpora seem more informative based on the frequency and authentic academic English. In addition, the use of LAs of results in the textbooks does not cover all possible Ikmi Nur Oktavianti, Japen Sarage Linking Adverbials in Indonesian EFL Students’ Essays: A Corpus-Driven Study 1586 common uses found in native-speaker corpora. LAs have been analyzed based on the perspective of professionalism. Lei (2012) examined LAs in EFL Chinese dissertations and professional writers and identified 33 LAs overused by doctoral students, and some of them were misused. Meanwhile, there are 25 LAs underused by doctoral students. The study also revealed that doctoral students relied more heavily on fewer LAs than professional writers. Cho (2020) compared the LAs used by master students and professional writers. The study showed that master students overused 23 LAs and underused 7 LAs. Also, there was some misuse regarding some LAs, e.g., on the other hand with in contrast or on the contrary. Some of the aforementioned previous studies have been conducted by focusing on native or second/foreign language data only, and some were done by comparing second/foreign language data to native language repertoire. In the Indonesian context, LA research has been less notorious (Lumbangaol, 2022; Nizar & Munawaroh, 2019) and should be done more frequently to comprehensively describe the learners’ language development and mastery. This research specifically focuses on academic writing because this genre is more complex than other genres or registers (Biber & Gray, 2016), and LAs are the common features in academic prose (Biber, 2006; Biber et al., 2016; Biber & Gray, 2016). This study investigates LAs to answer the following questions: (1) What are LAs used by EFL learners in academic writing? and (2) How are the LAs used by EFL learners? Furthermore, this research employs a corpus-based approach meaning that the language data is large which can support a more comprehensive analysis. By conducting a corpus study on the use of LAs, the results can map the learners’ language development, especially related to LAs. These can inform writing instructors and materials writers to provide better exposure to the use of LAs. Method This is mixed-method research because there are qualitative data (sentences containing LAs) and quantitative data (frequency, percentage, token). This study has several steps: corpus design, data collection (corpus construction, LAs identification), and data analysis. Participants The study's participants are sixth-semester students of the English Education Department at Universitas Ahmad Dahlan enrolled in the Academic Writing course. The rationale of the participant selection: the students have completed previous writing courses and learned basic writing materials and techniques. In addition, the course syllabus aligns with the study's needs since the outcome is an academic essay. Research Instrument The instruments in this study are Language Background Questionnaire (LBQ) and a writing task. LBQ provides the linguistic backgrounds of the participants related to IDEAS, Vol. 10, No. 2, December 2022 ISSN 2338-4778 (Print) ISSN 2548-4192 (Online) 1587 the use of the L1 and L2. The background is necessary to describe the profile of the learners whose writing will be compiled as a learner corpus in the present study. This study adopts Flege et al.’s (Flege et al. 1999) LBQ. The second instrument is a writing instruction to collect students’ academic writing. Writing instruction was developed to collect students’ written production from several topics related to education and technology for the learner corpus ranging from 1500-2000 words. The learner corpus design is generally modified from (Granger, 2008, 2012), which is presented in figure 2. Figure 1. Learner corpus design Based on the design, the focal points are learners and task variables. Thus, the design of the learner corpus in this study is displayed in table 1. Table 1. Learner corpus design Learner Task variables General L2-specific General L2-specific Age: 19-21 Learning context: FL Medium: written Task type: non- exam Gender: Female, male Proficiency: Intermediate, B1 written reports, and essays Field: education, technology Conditions: timed Region: Indonesia L2: Indonesian Genre: Academic Mother: local language(s), Indonesian Other FL: n/a Data Collection and Identification of Linking Adverbials The data collection procedure involves the distribution of writing assignments to the participants with several topics they can select. The academic essay should be written in 1500-2000 words. The writing assignment will be timed and submitted online in a Google Drive folder. The compiled texts were cleaned for the exact Ikmi Nur Oktavianti, Japen Sarage Linking Adverbials in Indonesian EFL Students’ Essays: A Corpus-Driven Study 1588 measurement of corpus size and the accuracy of query results and collected as a learner corpus named Learner Corpus of Academic Writing (LCAW) with 52,404 words. The corpus files were converted into .txt format and loaded into a corpus tool, LancsBox (Brezina et al., 2020). Regarding the identification, this study utilizes Liu’s (2008) list comprising 110 LAs. All 110 LAs were the search keyword to identify the LAs in the learner corpus and the KWIC feature in LancsBox was used. However, one LA might carry multiple meanings, thus, the searches were followed up with a manual check to ensure the identified units are LAs and not a conjunction. This is also explained by Liu (2008) as an important step to validate the data. Data Analysis The analysis focuses on semantic types of LAs in the students’ academic writing and the comparison to native corpora to answer the research questions. The analyses include the classification of semantic functions and the frequency of use. The present study employs Liu (2008)’s classification of semantic functions. The analysis will be confirmed by employing inter-rater agreement to diminish bias in classifying LAs. This step was commonly done in the corpus-based analysis (Oktavianti & Prayogi, 2022; Wang, 2022). Lastly, the results will be compared to those in native corpora of students’ essays, Louvain Corpus of Native English Essays (LOCNESS) (Granger, 1998), which comprises university-level essays of 324,304 words. Results This section presents the types of LAs found in the LCAW. The present study identified all major categories of LAs based on Liu (2008)’s classification; they are emphatic, apposition, adversative, causal resultative, and sequential LAs. Linking Adverbial Categories in LCAW The followings are all LA categories found in the learner corpus. 1. Additive LAs The present study identified all subcategories of additive LAs, namely emphatic, apposition, and similarity comparative LAs. a. Emphatic The examples of emphatic LAs found in LCAW are besides, furthermore, moreover, and in addition, as follows. (1) Besides, teachers should be ambitious and professional to support the process of teaching and learning to their students and enhance students’ competence. (2) Furthermore, some students who are from rural background do not have access to the Internet or require smartphones. (3) Moreover, the applications on their phones, like Instagram, TikTok, Twitter, Facebook, Snapchat, can distract their focus while studying. IDEAS, Vol. 10, No. 2, December 2022 ISSN 2338-4778 (Print) ISSN 2548-4192 (Online) 1589 (4) In addition, teachers who have creative thinking skills can also motivate their students to be more creative in dealing with various areas of life. b. Apposition LAs Apposition LAs are also found in LCAW, e.g., in other words, for example and namely; below are some examples from the corpus. (5) In other words, less monetary investment, but the results can be better than the other options. (6) For example, if you study abroad and want to get a job, online education is a good choice. (7) There are also other factors that become obstacles in implementing online learning, namely health factors. c. Similarity comparative LAs Another additive type identified in the corpus is similarity comparative LAs, such as likewise and similarly, as shown below. (8) Likewise, people who do not have the opportunity to continue their education because of limited funds, they still have the opportunity to learn through the online learning system. (9) Similarly, the Facebook organ does not initiative visibly shared new online registration to all friends in an individual’s network, encouraging social conformity. 2. Adversative LAs All types of adversative LAs can be identified in LCAW. Below are examples of each subcategory: proper adversative and contrastive adversative. a. Proper adversative Some examples of proper adversative (while and however) LAs are presented below. (10) While it is indisputable, there is still plenty of schools which still apply corporal punishments. (11) However, the benefits of adding the 21st-century skills to your curriculum are great. b. Contrastive adversative The second subcategory of adversative, contrastive adversative LAs are also found in LCAW, as presented below. (12) In fact, a lot of contents are presented on social media with various themes. (13) Actually, I think that this online learning method is considered effective and helpful. c. Correction The third subcategory of the adversative LAs is the correction type (e.g., Ikmi Nur Oktavianti, Japen Sarage Linking Adverbials in Indonesian EFL Students’ Essays: A Corpus-Driven Study 1590 instead and rather) which can be seen in examples (14) and (15). (14) Instead, it should be about declaring your approach to using technology. (15) Rather than missing important class sessions due to weather conditions, students can always attend by participating in discussion boards or chat sessions and watching lectures or reading materials. d. Dismissal This subcategory occurs restrictedly (e.g., despite), as shown in example (16). (16) Despite all the disadvantages, a new influencer Fadil Jaidi has become popular on social media. 3. Causal/Resultative LAs LAs of the causal resultative category are also found in LCAW, which is divided into two subcategories: general causal and conditional causal. Some examples are shown below. a. General causal General causal LAs found in LCAW include as a result, consequently, therefore, and some others. (17) As a result, using an online education platform allows a better work balance and study, so there is no need to give up. (18) Consequently, those unavailing things that the students do will ruin their mindset and critical thinking. (19) Therefore, every educational institution continues to improve the quality of skills for every educator b. Conditional clausal Conditional cuasal LA identified in LCAW is then as exemplified below. (20) Then, we also have to ascertain whether the economy can be profitable or even detrimental 4. Sequential LAs Lastly, there are some sequential LAs identified in LCAW, consisting of nearly all subcategories, namely enumerative/listing, simultaneous, and summative LAs. The examples are as follows. a. Enumerative/listing The first sequential subcategory is enumerative, which shows a greater number of type representations. Some examples are written in (21)—(25). (21) Firstly, be flexible in time and place because online learning system is different from ordinary schools. (22) Secondly, it can trigger cyberbullying. (23) Finally, digital minimalism is how we use technology wisely and consciously to help our activities. IDEAS, Vol. 10, No. 2, December 2022 ISSN 2338-4778 (Print) ISSN 2548-4192 (Online) 1591 (24) Lastly, an English teacher must have collaboration skills so they can work with many people within and outside the country. (25) Next, replace applications that have been deleted or restricted to use and move to other activities, such as reading comics, painting, or other social activities. b. Simultaneous The second subcategory of sequential LAs, simultaneous LAs, are limitedly found in LCAW. (26) Meanwhile, the cons are many people consider this activity to be a place to flex. (27) At the same time, the second most users are from 20 to 24 years old. c. Summative Summative subcategory examples are seen in examples (28)—(30). (28) In conclusion, if we can study optimally with the materials given by the lecturers, surely we can get the same knowledge as in face-to-face learning. (29) In short, people will be attracted if the viral content is what is useful to them. (30) To sum up, social media has become so addictive to many users from various ages. It is evident that nearly all categories of LAs are found in the learner corpus under study, LCAW. However, there is one missing/absent subcategory, which is the transition to another topic. This subcategory of LAs belongs to the category of sequential LAs. It is surprising to see the absence of this subcategory since the corpus was compiled from university-level students’ essays. Nonetheless, the result might indicate a lack of LA mastery of the learners that should be revisited and taken into account. The Use of Linking Adverbials by EFL Learners After discussing the findings of LAs types/categories in the learner corpus, this section focuses on the frequency of use and the distribution of each LA in LCAW. Table 2 presents the frequency of use individually. Table 2. Frequency of LAs in LCAW Category LAs Token Frequency Percentage (1) Additive Emphatic additionally 1 0,11 again 1 0,11 also 6 0,67 and also 4 0,44 not only… but also… 10 0,00 but also 1 0,11 as well 5 0,55 besides 20 2,22 Ikmi Nur Oktavianti, Japen Sarage Linking Adverbials in Indonesian EFL Students’ Essays: A Corpus-Driven Study 1592 in addition 37 4,11 in addition to 13 1,44 further 3 0,33 furthermore 5 0,55 moreover 8 0,89 not to mention 1 0,11 of course 59 6,55 too 32 3,55 Subtotal 206 21,75 Apposition that is 1 0,11 in other words 4 0,44 for example 46 5,11 for instance 3 0,33 namely 27 3,00 Subtotal 81 8,99 Similarity Comparative likewise 2 0,22 Similarity 1 0,11 Subtotal 3 0,33 Subtotal Additive 290 31,52 (2) Adversative Proper Adversative at the same time 3 0,33 while 43 4,77 however 50 5,55 nevertheless 1 0,11 though 22 2,44 Subtotal proper adversative 119 13,21 Contrastive Adversative actually 27 3,00 in fact 10 1,11 in reality 2 0,22 on the other hand 6 0,67 Subtotal contrastive adversative 45 4,99 Correction instead 11 1,22 on the contrary 3 0,33 rather 12 1,33 Subtotal correction 26 2,89 Dismissal after all 2 0,22 despite 3 0,33 despite that 1 0,11 Subtotal dismissal 6 0,67 IDEAS, Vol. 10, No. 2, December 2022 ISSN 2338-4778 (Print) ISSN 2548-4192 (Online) 1593 Subtotal adversative 196 21,75 (3) Causal Resultative General Causal as a result 4 0,44 because of 5 0,55 consequently 1 0,11 hence 1 0,11 so (sentence-initial) 49 5,44 so 197 21,86 therefore 45 4,99 thus 9 1,00 Subtotal general causal Conditional Causal then 9 0,98 Subtotal conditional causal 9 0,98 Subtotal Causal Resultative 320 34,52 (4) Sequential Enumerative/listing first 18 2,00 firstly 5 0,55 first of all 2 0,22 second 14 1,55 secondly 6 0,67 third 8 0,89 thirdly 1 0,11 finally 5 0,55 last 2 0,22 lastly 7 0,78 last but not least 2 0,22 next 2 0,22 then 18 2,00 and then 2 0,22 Subtotal enumerative 92 10,21 Simultaneous at the same time 1 0,11 meanwhile 1 0,11 Subtotal simultaneous 2 0,22 summative In conclusion 11 1,22 In short 3 0,33 To sum up 6 0,67 Subtotal summative 20 2,22 Subtotal sequential 114 12,65 Ikmi Nur Oktavianti, Japen Sarage Linking Adverbials in Indonesian EFL Students’ Essays: A Corpus-Driven Study 1594 Based on Table 2, it is clear that various LAs found in the corpus and the frequency indicate that the students already make use of LAs in writing their essays. Table 3 presents the token frequency and percentage of each linking adverbial type. Table 3. Frequency of LA type in LCAW No Category of LAs Token Percentage 1 Causal 320 34,78 2 Additive 290 31,52 3 Adversative 196 21,30 4 Sequential 114 12,39 Table 3 demonstrates that the most frequent type found in students’ essays is causal resultative LAs (34,78%), followed by additive (31,52%) and adversative (21,30%). Meanwhile, the least frequent type is sequential LA, with 12,39%. Causal resultative adverbials outnumber all types of LAs in the learner corpus under study. This present study identified some significant gaps in the token-type ratio regarding token and type comparison, as shown in figure 2. Figure 2. Token and Type Frequency of LAs in LCAW Figure 2 displays the comparison of tokens and types of LAs in LCAW. Some significant TTR (Type-Token Ratio) scores are based on the graph, such as general causal LAs. This LA type has 8 types with 311 tokens meaning that the TTR is 2,57%. The low percentage of TTR shows the low variety of the words under this category. Another category with a pretty noticeable TTR percentage is emphatic and proper adversative, with 8,67% and 4,2%, respectively. Furthermore, the similarity comparative subcategory is negligible in terms of the token frequency and type frequency. Although the additive category is quite frequent in general, the similarity comparative subcategory is included as one of the least frequent LAs in LCAW. These 196 81 3 119 45 26 6 311 9 92 2 20 0 17 5 2 5 4 3 3 8 1 14 2 3 0 0 50 100 150 200 250 300 350 F re q u e n cy Category of LAs Token Type IDEAS, Vol. 10, No. 2, December 2022 ISSN 2338-4778 (Print) ISSN 2548-4192 (Online) 1595 findings show that, even though the students have used the LAs in their writing, there is a problem in regard to the variants of the LAs indicating the restricted repertoire of LAs. This study also compares the frequency of LAs found in LOCNESS as the control corpus. Table 4 displays the token frequency and percentage of each primary type of LAs in the corpus. Table 4. LAs in LOCNESS No Type of LAs Token Frequency Percentage 1 Adversative 1808 37,95 2 Causal 1318 27,67 3 Additive 1189 24,96 4 Sequential 449 9,42 Based on Table 4, the most frequent type found in LOCNESS is adversative type (37,95%), followed by causal resultative (27,67%), additive (24,96%), and the least frequent is sequential type (9,42%). In comparison to LCAW, there are some points to underline. Table 3 presents the frequency of LAs in LCAW and proves that the most frequent type is causal resultative (34,78%). This result differs from that in LOCNESS because LOCNESS exhibits the use of adversative as the most frequent one (37,95%), and causal resultative occupies the second place with 27,67%. Additive type in LCAW occupies the second position (31,52%), while in LOCNESS it occupies the third position (24,96 %). However, LCAW and LOCNESS shared the same results for the least frequent types, i.e., sequential LAs. As for the type distribution of LAs in LCAW and LOCNESS, figure 2 displays the comparison. 17 5 2 5 4 3 3 8 1 14 2 3 0 18 9 3 7 8 3 6 10 4 19 2 5 2 0 2 4 6 8 10 12 14 16 18 20 F re q u e n cy Type in LCAW Type in LOCNESS Ikmi Nur Oktavianti, Japen Sarage Linking Adverbials in Indonesian EFL Students’ Essays: A Corpus-Driven Study 1596 Figure 3. Type frequency of LAs in LCAW and LOCNESS Figure 3 shows that many types have similar numbers in the category and subcategory of LAs, either in LCAW or LOCNESS, e.g., emphatic, general causal, and simultaneous. However, certain types have a significant gap in type, e.g., apposition, contrastive adversative, and enumerative. The transition to another topic subcategory is not found in LCAW, but there are 2 types representing this subcategory in LOCNESS. In other words, transition to another topic subcategory is absent in LCAW, which should be revisited and considered pretty seriously. However, in a more general sense, sequential LAs are less familiar for foreign learners and native writers, as proven by LCAW and LOCNESS. Discussion The study’s results demonstrate the tendency of certain use of LAs in the academic context. Students in this study were identified to use causal resultative, emphatic, and adversative LAs compared to the other categories. The current study highlights the importance of specific LAs categories, such as causal resultative, additive, and adversative, and those LAs missing from the corpus (i.e., transition to another topic). Regarding the findings from other relevant studies, this study can confirm the importance of a particular LA type, i.e., additive. The findings show that additive LAs are the highly frequent LAs (although they are not the most frequent in the present study) following some previous research, e.g., Ahmad & Way (2020), Appel & Szeib (2018), Ha (2016), Jamil et al. (2014), Nizar & Munawaroh (2019). These studies show that people connect ideas in the discourse by utilizing addition or additive markers. However, the current study shows that the additive type is the second most frequent LAs, following causal resultative LAs. This phenomenon is surprising,g but considering the texts compiled in this corpus, the widespread use of causal resultative LAs is sensible to provide robust arguments. This study also proves the use of adversative LAs as the third most frequent type in students’ writing, which is in line with the study of Appel & Szeib (2018), Ha (2016), and Jamil et al. (2014). Adversative LAs are important to express contrary opinions, facts, comparisons, and corrections that might be necessary for an academic context. As proven by Liu (2008), academic prose uses more of this LA type than spoken language and fiction registers. However, according to Lei (2012), adversative LAs are the most problematic connectors since they might perform multiple functions and students tend to use them in inappropriate registers. In regard to the less frequent LAs, Ishikawa (2010) found that sequential LAs tend to be underused by EFL learners, which is also relevant to the finding of this study. The present study identifies sequential LAs as the fourth most frequent LAs type. This can be related to the nature of EFL learners with limited exposure to the target language, especially regarding academic language. Many studies have proven the lack of specific linguistic features in academic prose in the EFL context that should be considered more seriously (Ahmad & Way, 2020; Bychkovska & Lee, 2017; IDEAS, Vol. 10, No. 2, December 2022 ISSN 2338-4778 (Print) ISSN 2548-4192 (Online) 1597 Nakayama, 2021; Oktavianti & Sarage, 2021). The findings of the current research are also contradictive to some previous studies. For example, Nizar & Munawaroh (2019) claimed that causal LAs are underused by the learners, which does not correspond to the results of this study showing that causal resultative LAs are the most frequent LAs in students’ essays. Another thing to consider is the use of adversative LAs. Ahmad & Way (2020) pointed out that adversative LAs are the least frequent LAs, which contradicts this study’s result claiming that adversative LAs occupy the third position (not the last one). This might be plausible due to some aspects, including the level of learners’ proficiency and the text types of the corpus. The results of this study also complement those of the previous ones, i.e., similarity and sequential LAs. Regarding similarity comparative, this subcategory has the lowest frequency in LCAW and LOCNESS, indicating that this subcategory is less crucial in students’ writing (EFL learners or native writers). As for the absent subcategory, transition to another topic is absent in LCAW. At the same time, it is found in LOCNESS, showing that this subcategory is not familiar to the EFL learners or the learners do not have sufficient knowledge of the category. Based on the study findings, there are some points to underline concerning LAs teaching and learning. Students should be more aware of the category and variants of LAs since sufficient knowledge of LAs can help improve writing quality. Furthermore, the ideas can be connected more flawlessly and logically. Thus, the teachers should emphasize some categories to the learners, namely sequential LAs and similarity LAs. The missing subcategory should be taught and intensively introduced to the learners, so this category can be part of the learners’ linguistic repertoire. Conclusion This study shows that students have utilized LAs to create logical connections in their writing. Overall, the EFL learners in this study have some knowledge of LAs and how to use them in their writing. However, some points must be considered, including the variants of the LAs. It is proven by the restricted type of particular LAs and the absence of the transition to another topic subcategory. It is then crucial to revisit the existing teaching materials so the students can be equipped with sufficient knowledge of LAs. In comparison to the native corpus, LOCNESS, it is evident that there are some differences in the frequency of LAs category and the type comparison also exhibits the fact of some limited variants of LAs in the EFL students’ writing. Unfortunately, the present study merely focuses on the LAs categories and subcategories and their frequency of use. The study overlooks other variables of study as the functions of LAs, the patterns of use, the register contexts, etc. Therefore, future research should be able to analyze more aspects, and it is also necessary to collect more data (larger corpus size). The study of LAs and other discourse elements is very potential to conduct, and the findings might contribute robustly to language teaching and Ikmi Nur Oktavianti, Japen Sarage Linking Adverbials in Indonesian EFL Students’ Essays: A Corpus-Driven Study 1598 learning. Acknowledgment The authors express their deepest gratitude to the institute of research and community service, Lembaga Penelitian dan Pengabdian Masyarakat, Universitas Ahmad Dahlan, for funding this research. References Ahmad, U. K., & Way, L. P. (2020). The Use of Linking Adverbials in Malaysian Students’ Argumentative Essays. International Journal of Social Science and Human Research, 03(12). https://doi.org/10.47191/ijsshr/v3-i12-08 Appel, R., & Szeib, A. (2018). Linking adverbials in L2 English academic writing: L1-related differences. System, 78, 115–129. https://doi.org/10.1016/j.system.2018.08.008 Biber, D. (2006). University language: A corpus-based study of spoken and written registers. J. Benjamins. Biber, D., & Gray, B. (2016). Grammatical complexity in academic English: Linguistic change in writing. Cambridge University Press. Biber, D., Johansson, S., Leech, G. N., Conrad, S., & Finegan, E. (2021). Grammar of Spoken and Written English. John Benjamins Publishing Company. https://doi.org/10.1075/z.232 Biber, D., Reppen, R., Schnur, E., & Ghanem, R. (2016). On the (non)utility of Juilland’s D to measure lexical dispersion in large corpora. International Journal of Corpus Linguistics, 21(4), 439–464. https://doi.org/10.1075/ijcl.21.4.01bib Brezina, V., Weill-Tessier, P., & McEnery, A. (2020). LancsBox (v.5.x). Lancaster University. http://corpora.lancs.ac.uk/lancsbox. Bychkovska, T., & Lee, J. J. (2017). At the same time: Lexical bundles in L1 and L2 university student argumentative writing. Journal of English for Academic Purposes, 20, 38–52. Cho, Y. (2020). Linking Adverbials in Academic Writing on English Linguistics by Korean MA Students. 영학논집 (English Studies), 40, 193–219. Dutra, D. P., Orfanó, B. M., & Almeida, V. C. (2019). Result linking adverbials in learner corpora. Domínios de Lingu@gem, 13(1), 400. https://doi.org/10.14393/DL37- v13n1a2019-17 Flege, J. E., Yeni-Komshian, G. H., & Liu, S. (1999). Age Constraints on Second-Language Acquisition. Journal of Memory and Language, 41(1), 78–104. https://doi.org/10.1006/jmla.1999.2638 Flowerdew, J., & Mahlberg, M. (2007). Lexical Cohesion and Corpus Linguistics (T. Nevalainen & S.-K. Tanskanen, Eds.). John Benjamins Pub. Co. IDEAS, Vol. 10, No. 2, December 2022 ISSN 2338-4778 (Print) ISSN 2548-4192 (Online) 1599 Gao, X. (2016). A cross-disciplinary corpus-based study on English and Chinese native speakers’ use of linking adverbials in academic writing. Journal of English for Academic Purposes, 24, 14–28. https://doi.org/10.1016/j.jeap.2016.08.002 Granger, S. (1998). The computer learner corpus: A versatile new source of data for SLA. In S. Granger (Ed.), Learner English on computer. Longman. Granger, S. (2008). Learner corpora. In A. Lüdeling & M. Kytö (Eds.), Corpus linguistics: An international handbook volume I (pp. 259–274). Walter de Gruyter. Granger, S. (2012). Learner Corpora. In C. A. Chapelle (Ed.), The Encyclopedia of Applied Linguistics (p. wbeal0669). Blackwell Publishing Ltd. https://doi.org/10.1002/9781405198431.wbeal0669 Grisot, C. (2018). Cohesion, Coherence and Temporal Reference from an Experimental Corpus Pragmatics Perspective. Springer International Publishing. https://doi.org/10.1007/978-3- 319-96752-3 Ha, M.-J. (2016). Linking adverbials in first-year Korean university EFL learners’ writing: A corpus-informed analysis. Computer Assisted Language Learning, 29(6), 1090–1101. https://doi.org/10.1080/09588221.2015.1068814 Halliday, M. A. K., & Hasan, R. (2014). Cohesion in English. Taylor & Francis. Ishikawa, S. (2010). A corpus-based study on Asian learners’ use of English linking adverbials. Themes in Science and Technology Education, 3(1–2), 139–157. Jamil, E., Mahmood, M. A., & Hussain, Z. (2014). A Corpus-Based Analysis of Linking Adverbials in Pakistani English. Research on Humanities and Social Sciences, 4(7), 78–81. Leedham, M., & Cai, G. (2013). Besides … on the other hand: Using a corpus approach to explore the influence of teaching materials on Chinese students’ use of linking adverbials. Journal of Second Language Writing, 22(4), 374–389. https://doi.org/10.1016/j.jslw.2013.07.002 Lei, L. (2012). Linking adverbials in academic writing on applied linguistics by Chinese doctoral students. Journal of English for Academic Purposes, 11(3), 267–275. https://doi.org/10.1016/j.jeap.2012.05.003 Liu, D. (2008). Linking adverbials: An across-register corpus study and its implications. International Journal of Corpus Linguistics, 13(4), 491–518. https://doi.org/10.1075/ijcl.13.4.05liu Liu, G. (2013). On the Use of Linking Adverbials by Chinese College English Learners. Journal of Language Teaching and Research, 4(1), 149–155. https://doi.org/10.4304/jltr.4.1.149- 155 Lumbangaol, R. R. (2022). The Using Of Linking Words to Enhance Students’ Writing Ability at Universitas Potensi Utama. Journal MELT (Medium for English Language Teaching), 7(1), 01. https://doi.org/10.22303/melt.7.1.2022.01-13 Ikmi Nur Oktavianti, Japen Sarage Linking Adverbials in Indonesian EFL Students’ Essays: A Corpus-Driven Study 1600 Nakayama, S. (2021). A Corpus-Based Analysis of Japanese EFL Learners’ Linking Adverbial Use. The Educational Review, USA, 5(6), 164–171. https://doi.org/10.26855/er.2021.06.002 Nizar, A. A., & Munawaroh, I. B. (2019). Journal of Research on Applied Linguistics Language and Language Teaching. 2(2), 103–113. Oktavianti, I. N., & Prayogi, I. (2022). Discourse functions of lexical bundles in Indonesian EFL learners’ argumentative essays: A corpus study. Studies in English Language and Education, 9(2), 761–783. https://doi.org/10.24815/siele.v9i2.23995 Oktavianti, I. N., & Sarage, J. (2021). Lexical Bundles in Students’ Argumentative Essays: A Study of Learner Corpus. Indonesian Journal of EFL and Linguistics, 6(2), 510–534. Phoocharoensil, S. (2017). Corpus-based Exploration of Linking Adverbials of Result: Discovering What ELT Writing Coursebooks Lack. 3L The Southeast Asian Journal of English Language Studies, 23(1), 150–167. https://doi.org/10.17576/3L-2017-2301-11 Pipatanusorn, L., & Wijitsopon, R. (2019). A Corpus-based Study of Linking Adverbials in Written American English. 14(1), 47. Techarsatitwong, P. (2015). A Corpus-Based Study of Linking Adverbials of Result so, thus, therefore, and hence [Master Thesis]. Thammasat University. Wang, Y. (2022). Linking Adverbials in Argumentative Essays of Advanced EFL Learners: A Corpus-assisted Analysis. Korean Journal of English Language and Linguistics, 22(null), 396– 417. https://doi.org/10.15738/KJELL.22..202204.396 Yang, W., & Sun, Y. (2012). The use of cohesive devices in argumentative writing by Chinese EFL learners at different proficiency levels. Linguistics and Education, 23(1), 31–48. https://doi.org/10.1016/j.linged.2011.09.004 Yin, Z. (2014). Linking Adverbials in English [Ph.D. Dissertation]. Victoria University of Wellington. Yin, Z. (2016). Register-specific meaning categorization of linking adverbials in English. Journal of English for Academic Purposes, 22, 1–18. https://doi.org/10.1016/j.jeap.2016.01.004