Rhetorical Move and Genre Knowledge Development of English and Indonesian Abstracts: A Comparative Analysis

A plethora of research has shown that genre analysis through move analysis is a practical approach to identify the complexity of writing research articles (RAs). However, little is known about the genre knowledge development that is manifested in abstract discourse patterns. This study aims to determine whether or not there is an influence from the level of education with the development of genre knowledge, especially in the field of writing research abstracts. Using Hyland’s (2000) five-move analysis model, this study analyzed the comparison and identity of abstracts of theses and dissertations in English and Indonesian. From the analysis, it can be seen that there are some differences and similarities in the manifestation of abstract discourse patterns in English and Indonesian final paper abstracts. In terms of genre knowledge, its development could be reflected through the level of study, in this case, from master’s to doctoral degrees. As evident in the dissertation abstracts in both languages, Move 1 (Introduction) evinces richer varieties in step realizations than in master’s theses. The analysis also indicates no crucial differences in genre knowledge development across languages and fields of study. Further comparative research on this particular topic with more subject of data is suggested.


INTRODUCTION
Abstract in a research paper is a measure of appropriateness which is often the main and first consideration for journal editors to accept or reject an article (Kurniawan, et al., 2019) and serves as a key determinant of whether or not international readers find the research paper relevant with their studies (Fauzan et al., 2020;Kafes, 2012). It also represents the entire content of the article, as it covers everything from the background to conclusions presented in a concise manner (Samraj, 2002;Swales, 1990). In a similar vein, Tocalo (2021) contends that research article abstracts have gained mounting significance, especially owing to the increased use of online scholastic web indexes, because they pave the way for research reports to be accessible to a global academic world.
In academia, writing styles can vary depending on the influence of the 'writer's mother tongue' (see Pérez-Llantada, 2014). In addition, differences in the system of language use between the mother tongue and English lead to variations in the realization of the language in organizing the patterns of discourse where there are variations that are accepted or that are not following the rules. Therefore, there is a need for follow-up to understand discourse patterns or abstract writing patterns in research articles (Pho, 2008). In other words, as part of the academic community, in order for writers to successfully get their research published, they need to acquire some knowledge of genre.
Genre knowledge is considered a field of study that deals with understanding texts in a particular discourse context (Berkenkotter & Huckin, 1995;Johns, 1997;Swales, 1990;Tardy, 2009). Hyland (2000, p. 197) suggested that the ability of writers to create genre-appropriate texts that "'display repeated rhetorical responses to similar situations, with each generic act including some degree of novelty and judgment' constructs their 'antecedent knowledge of texts'". As a result, in order to develop that ability, writers must draw on their awareness and knowledge of prior texts, which is established through explicit genre analysis in the classroom as a part of the ESP/EAP study (Swales, 1990;Tardy, 2009). Implicit in this notion is the underlying assumption that academic experience will influence someone's genre knowledge (Rounsaville, 2014, p. 334). It is also assumed that genre analysis through the method of move analysis is an effective method for recognizing the complexities of writing research articles. Move analysis is an analytical approach to text identification (Swales, 1990(Swales, , 2004 and has been used to evaluate various types of genres, such as academic papers, theses, and essays. Move analysis can be described as a textual analysis of the communicative goals and their constituent steps to explain how the discourse pattern is manifested in the writing being investigated. Many researchers have explored rhetorical move patterns in abstracts, which categorically fall under three contexts. The first one is the analysis that involves only one scientific discipline (e.g., Amnuai, 2019aAmnuai, , 2019bSabila & Kurniawan, 2020). The next one is research that compares the rhetorical move structure across disciplines (e.g., Gani et al., 2021;Juanda & Kurniawan, 2020;Nikpei, 2016;Ramadhini et al., 2021) as a way to see the difference between several disciplines. Lastly, some research compares abstract discourse patterns written by writers with various backgrounds (e.g., Mohsenzadeh, 2013;Putri et al., 2021;Suryani & Rismiyanto, 2019). Even though they have different focuses in data, their main goal is to see what move or discourse pattern is used in each text. Analysis on genre knowledge development, on the other hand, provides an insight that the development of genre knowledge is formed by the exposure received by the author through variations of writing course (Artemeva & Fox, 2010;Driscoll et al., 2020) and transnational boundary-crossing practices (Rounsaville, 2014).
Although studies that focus on the manifestation of abstract discourse patterns have been widely conducted (see , scant attention has been paid to the development of genre knowledge manifested in abstract move patterns. Clearly, additional research is needed. The urgency of this additional research is based on the role of a person's educational level in developing their academic writing skills. Occupying this lacuna, this study sought to determine whether or not there is an influence from the level of education with the development of genre knowledge, especially in the field of writing research abstracts. This study also aims to compare and identify commonalities and differences between abstracts written in English and Indonesian. This study used Hyland's (2000) five-move analysis model as the main instrument because it has been predominately used in many research studies that focus on the manifestation of discourse patterns (Amnuai, 2019a(Amnuai, , 2019bKurniawan et al., 2019;Suryani & Rismiyanto, 2019). Using this model, the findings from this study are expected to produce a clearer picture of (1) the development of abstract writing skills from postgraduate to doctoral dissertation levels along with the factors that influence it, and (2) differences in the manifestation of abstract discourse patterns at each level and language.

Rhetorical Move Analysis
A rhetorical move is a unit that relates to both the writer's purpose and the way in which the writer's message or meaning is conveyed (Nikpei, 2016). This communicative goal can be realized in the form of sentences, paragraphs, or several sentences and paragraphs. In other words, for example, if the communicative goal of the writer is to describe the research method, several constituent steps need to be considered, such as conveying information about research subjects, research instruments, and research data analysis procedures. Based on this definition, rhetorical move analysis can be interpreted as a textual analysis of communicative objectives and their constituent steps to understand how the discourse pattern is manifested in the writing that is being analyzed (Zamani & Ebadi, 2016). In addition, discourse pattern analysis aims to determine the trends in the practice of certain writing styles and the variations that appear between the texts being analyzed.
Move analysis is an analytical approach to text identification (Swales, 1990(Swales, , 2004 and has been used to evaluate various types of genres, such as academic papers, theses, or essays. Move analysis can be described as a textual analysis of the communicative goals and their constituent steps to explain how the discourse pattern is manifested in the writing being investigated. There has been plenty of research that focuses on rhetorical move analysis. Even though they had different focuses in data, they shared a common denominator: to see what move or discourse pattern is used in each text. Researchers are interested in the analysis of rhetorical move structure that involves only one scientific discipline. Amnuai (2019a), for instance, focused on analyzing the rhetorical move structure and grammatical features of abstracts of the two genres (master's theses and research articles or RAs) in the field of information and communication technology (ICT). The study investigated how the abstracts of the two genres are rhetorically realized and identified three grammatical aspects, namely tense, voice and that-complement clauses. The study revealed that differences were found in the frequency of occurrence of Introduction and Conclusion moves. Other aspects, such as voice usage, in the two genres are completely different. The study's findings provide insightful knowledge of rhetorical patterns and the use of lexico-grammatical features in each move in the abstracts.
The analysis that involves several disciplines is typical between researchers. Nikpei (2016), for example, tried to find out whether the research thesis abstracts in applied linguistics (TESOL students) and molecular biology in Tehran University are compatible with Bhatia's (1993) pattern of abstract moves. The study sought to discover any significant differences in the moves employed in the abstracts of two different disciplines' theses. The study found that abstracts in both disciplines have a similar structure with Bhatia's (1993) four-moves structure. However, certain variations of occurrences have been noted concerning those moves.
Other researchers compared abstract discourse patterns written by writers with various backgrounds, such as the mother tongue, academic writing proficiency level, or region. Suryani and Rismiyanto (2019), in their study, employed Hyland's five rhetorical move model to examine the rhetorical moves of English bachelor thesis abstracts written by Indonesian students from the non-English department. The study concluded that only a few English bachelor thesis abstracts follow Hyland's five move model. The study also revealed that the method move was always present in all of the English bachelor thesis abstracts. In a similar vein, Mohsenzadeh (2013) focused on how different discourse communities achieve their communicative purposes. The study found that differences between English and Persian literature book prefaces were in their obligatory move. In the English corpus, the obligatory move found was a general statement about the book, while the obligatory moves in the Persian prefaces were found to be the significance of the topic.
To resume, the existing body of literature has extensively explored rhetorical move patterns in abstracts of various paper types and taking into account a variety of contexts (inter alia fields of study, language, the writer's background). However, little to none explored how genre knowledge development could be realized in the manifestation of rhetorical move patterns in a research abstract. It is this research void that motivated the present research.

Genre Knowledge Development
Genre knowledge, according to Tardy (2009), can be classified into four domains: formal knowledge, rhetorical knowledge, process knowledge, and disciplinary knowledge. Formal knowledge can be defined as one's understanding of the textual features of a genre, such as norms, discourse pattern structures, and linguistic features. Rhetorical knowledge relates to one's understanding of the relationship between the genre's goals to be achieved and the contexts involved in it; in the context of this research, how the writer positions himself in conveying the contents of his writing. Process knowledge is related to efforts to complete the goals to be achieved, such as involving the ability to manage time and the ability to find related sources. Knowledge of scientific disciplines, or disciplinary knowledge, is something that is pertained to the study at hand. In other words, this knowledge is more to the writer's understanding of the disciplines in which s/he is involved.
The development of genre knowledge is formed by the exposure received by the writer through variations of writing and transnational boundary-crossing practices. Artemeva and Fox (2010), in their study, examined the purpose of students' antecedent genre knowledge and their development through an illustrative excerpt of an engineering genre-competence evaluation. The findings of this assessment suggested that students' antecedent genre knowledge does not guarantee their successful writing performance in the genre. Filling the gap, Driscoll et al. (2020) explored genre as a critical factor of understanding and promoting writing development. Their study aimed to describe how a writing course influences students' development of genre knowledge and metacognition and how these factors impact writing performance. The findings indicated that students make significant gains in their writing over the course. Rounsaville's (2014) study, however, focuses on a case study of how one student's distinctly transnational genre knowledge developed in response to a series of relocations between Italy and the United States. The study found that the student's literacy history reveals the particularly transnational character of applying prior genre knowledge within new writing situations.
In light of the previous studies explained above, this study aims not only to analyze the quality of discourse patterns displayed by the authors in their master thesis and doctoral dissertation abstracts but also to explore how their genre knowledge develops along with their experiences during their master to doctoral degrees. The importance of this study is centered around the potential role of a person's educational level in relation to his/her academic writing skills.

Research Design
Since this study aims to analyze and compare the realizations of abstract discourse patterns from the master theses and dissertations of lecturers at a state university in Bandung and their genre knowledge development, a comparative qualitative design was employed. Specifically, the manifestation of rhetorical organization in abstracts written in English and Indonesian were compared. This study used two main instruments. Hyland's (2000) abstract discourse patterns model as the first instrument was adopted to analyze the abstract discourse patterns (see Table 1).
Researchers have also been focusing on linguistic realizations, such as tense and voice, in abstracts, because they offer valuable insight into written genres (Esfandiari, 2014;Hanidar, 2016;Tseng, 2011;Tu & Wang, 2013;Zhang et al., 2012, as cited in Amnuai, 2019a. Following those previous studies, the linguistic realization in each language was also compared, focusing on verb tense and voice used in each abstract. For research on genre knowledge development, this study used the second instrument of an interview, following Rounsaville (2014) and Suherdi et al. (2021). The interview was undertaken to explore the acquisition of genre knowledge through different levels of study, i.e., master's and doctoral degrees.  Hyland's (2000) model.

Move
Step M1 Introduction S1 Arguing for topic significance S2 Making topic generalization S3 Defining key term(s) S4 Identifying gap M2 Purpose Stating the research purpose M3 Methods S1 Describing participant/data sources S2 Describing instrument (s) S3 Describing procedure and context M4 Findings Describing the main results M5 Conclusion S1 Deducing conclusion S2 Evaluating the significance of the research S3 Stating limitation S4 Presenting recommendation and implication

Data Collection
The data for this study were collected from the master thesis and dissertation abstracts of lecturers in a state university in Bandung. A total of eight abstracts in two languages (English and Indonesian) from four lecturers were purposively selected. As this study focused on exploring genre knowledge development, the number of abstracts under examination was purposely restricted in order to enhance the researchers' understanding of each participant's genre knowledge development. The selection of the participants was based on the criteria below: • The participants are two lecturers of the English department and two lecturers of the Indonesian department who have obtained a doctoral degree and have written a thesis and dissertation. • All participants obtained their degrees in universities in Indonesia (domestic graduates). • The abstracts are written in their respective field of study (English and Indonesian).
Interview as the second instrument covered three main sections in which each section consisted of several questions. The first part was intended to find out the background and personal identity related to the topic of this study. The second part was to unveil the acquisition of genre knowledge seen from Tardy's (2009) four domains of genre knowledge. In addition, the second part was also tailored to spot the development of genre knowledge influenced by personal, social, and socio-historical factors at the postgraduate levels. The third part was devised to solicit the information at the master and doctorate levels and the process of transitioning from the former to the latter. The protocol of the interview was based on several theoretical studies. Questions related to genre knowledge were based on Tardy's (2009), Devitt's (2015), and Tardy et al.'s (2020) study were used as the foundations of questions about factors contributing to the 'writers' genre knowledge development. Hyland's (2000) five-move abstract model was adopted as the guideline for data analysis. Introduction move establishes the context of the paper and motives for the research or discussion; the Purpose move indicates and outlines the intention behind the paper; the Method move provides information on design, procedures, assumption, approach, and data; the Product or Findings move states the main findings and the arguments; and the Conclusion move interprets results, draw inferences, points to the application or broader implications (Kurniawan et al., 2019). The procedure started by breaking down all abstracts into sentences. Then, the sentences were labeled according to the guidelines. Data analyzed manually using the Hyland (2000) model was transferred to Excel to make it easier to pinpoint trends and variations in the emergence of abstract discourse patterns in each type of scientific work. Abstracts written in English were assigned as E1 and E2, while abstracts written in Indonesian were assigned as I1 and I2. The numbers refer to master theses and dissertations, respectively. The exact process was carried out to determine trends and variations in the realization of the language.

Data Analysis
For research data from interviews, this study used the interview data analysis technique from Creswell (2012). In the interview session, the answers to each question were transcribed and inserted into a table. The answers of each participant were compared and concluded afterward. The answers were also used to deduce the overall conclusion on the development of the genre knowledge of the participant. The semistructured interviews were applied since there was some adjustment depending on the results of the textual analysis and the participants' answers to each question. Data display from the abstracts and interviews are coded as D (for data) throughout the paper. The moves being studied are bolded in the data display.

RESULTS AND DISCUSSIONS
This section delineates the results concerning the manifestation of move step occurrences and its language realization. This section also aims to address the purpose of the study: (1) to determine whether or not there is an influence from the level of education with the development of genre knowledge, especially in the field of writing research abstracts, and (2) to compare and identify whether there are differences between abstracts written in English and Indonesian

The Occurrences of the Moves and Steps
This sub-section discusses the occurrences of the moves and steps in each language and level of the study. The data obtained are translated into percentages. From the data obtained, there were 122 moves in total. There were 19 moves in E1, 33 moves in E2, 25 moves in I1, and 45 moves in I2. As seen in Figure 1, the number of Move 2 -Purpose was the least compared to the other four moves in all final papers. This could mean that the writers of the final papers did not focus on the outlines and intention behind the papers that come in the Purpose move. The other move that occurred the least was Move 5 (Conclusion). Meanwhile, Move 4 (Findings) had the highest number of occurrences for I2 and E1. For I1 and E2, Move 3 (Methods) and M1 (Introductions) occurred the highest, respectively.
The proportion of each manifested step received little consideration in previous research, as the majority were more concerned with the salience of the moves (Hardjanto, 2017, Kafes, 2012, Pho, 2008, as cited in Kurniawan et al., 2019). For that reason, this analysis would also focus on explaining move occurrences across the field and the level of study presented in Figure 2.  Since Move 2 (Purpose) and Move 4 (Findings) do not have any steps, according to Hyland (2000), only the rest of the moves were further examined with the focus of manifestation of their steps. This study analyzed the similarities and differences in the occurrences of the steps.
Based on Figure 2, the similarities showed that Step 3 of Move 1 (Defining the key term(s)) is the least manifested step, followed by Step 1 of Move 1 (Arguing for topic significance). Meanwhile, Step 3 of Move 3 (Describing procedure and context) occurred the most in all final papers. Interestingly, both languages did not use any Move 5, Step 2 (Evaluating the significance of the research), and Step 3 (Stating limitation). In the case of differences in each language, there were cases that illustrated more evidence of difference.
Step 2 of Move 1 (Making topic generalization) and Move 3 (Describing instrument(s)), for starters, were less performed in the abstracts In contrast, Step 1 of Move 3 -Describing participant/data sources was less used in abstracts written in English than abstracts written in Indonesian. The following excerpt depicts the manifestation of Step 1 of Move 3.

D3: Subjek coba dalam penelitian tahap pengembangan adalah (1) ahli pembelajaran sastra, ahli evaluasi pembelajaran bahasa, dan praktisi pembelajaran puisi yang memiliki kualifikasi di bidangnya untuk menguji validitas buku model PMP-PAK (Penilaian Autentik-Kolaboratif); (2) guru dan siswa di sekolah daerah (rural school) dan berada pada kluster menengah untuk menguji realibilitas rubrik penilaian pada buku model PMP-PAK serta untuk mengetahui respons pengguna terhadap buku model PMP-PAK.
[The research subjects in the development stage were (1) literary learning experts, language learning evaluation experts, and poetry learning practitioners who have qualifications in their fields to test the validity of the PMP-PAK (Authentic-Collaborative Assessment) model book; (2) teachers and students in rural schools and are in the middle cluster to test the reliability of the assessment rubric on the PMP-PAK model book and to determine the user's response to the PMP-PAK model book.] (Dissertation 2, I2, Move 3, Step 1, Sentence 12) Another interesting point to analyze is the embedded steps occurrences. In particular, in Move 3, there were some cases of embedded steps. There was a combination of two steps, namely Step 1 (Describing participants/data sources) + Step 2 (Describing instrument(s)), Step 3 (Describing procedure and context) + Step 2 (Describing instrument(s)), and Step 3 (Describing procedure and context + Step 1 (Describing participants/data sources) + Step 2 (Describing instrument(s)). Interestingly, these cases of embedded steps manifested a lot more in abstracts written in English. The following excerpts depict the manifestation of embedded steps. D6: The method used is primarily qualitative and the data were collected from two sources: the Indonesian data were taken from the library of a private university in Jakarta and the American data were directly downloaded from two American websites and were classified into prepositional phrases and clauses using the concordance software Antconc 3.4.1.0 developed by Anthony (2015). (Dissertation 2, E2, Sentence 5)

The Realization of Linguistic Features
This subsection discusses the details concerning the linguistic features realization concerning verb tenses and voices used in each move. Table 2. Linguistic features realization.

Move
Tense

Ac (100%)
Pa (100%) Table 2 shows the linguistic features realization concerning verb tenses and voices used in each move. Present tense and past tense are coded as Pr and Pa, while active voice and passive voice are coded as Ac and Pa. The bolded number shows the dominance in occurrences in each category. Based on Table 2, in terms of verb tense, past tense dominated the occurrences in abstracts written in English, especially in Move 3 (Methods), Move 4 (Findings), and Move 5 (Conclusion). The following excerpts present the use of past tense in abstracts written in English. The dominant use of past tense, especially in Methods and Findings moves, has been reported in some previous studies (Alhuqbani, 2013;Tseng, 2011;Zhang et al., 2012, as cited in Amnuai, 2019b. Zhang et al.'s (2012) study, however, found that present tense is primarily used in Vietnamese RA abstracts (Amnuai, 2019a). Meanwhile, comparing it to the abstracts written in Indonesian, the latter did not use any verb tense since the Indonesian language did not have overt tense markers in its linguistic realization. This further emphasizes the difference between English and Indonesian abstracts.
Interestingly, master thesis abstracts written in English used the present tense to present context and motives for the research realized in Move 1 (Introduction). The following excerpt is the use of present tense in Move 1 (Introduction).
D9: Therefore, the preliminary study's findings argue for a need to utilize a dialogic feedback process to solve the aforementioned problems. (Dissertation 1, E2, Move 1, Sentence 8) In terms of voice usage, it can be concluded that active voice dominated the occurrence in all moves (except for Move 3 (Methods) in English theses and Move 5 (Conclusion) in Indonesian dissertations). Lorés' (2004) study mentioned that texts are easier to understand when using the active voice (Amnuai, 2019b). The use of active voice in abstracts implies that the authors tried to focus on the agent rather than the action in the sentence.
D10: The preliminary study reported that the prospective teachers found it difficult to review back their teachings and to identify the focus of their teacher supervisors' comments. (Dissertation 1, E2, Move 1, Sentence 4) D11: Peneliti memilih cerita klasik "Hikayat Raja Kerang" sebagai sampel cerita klasik.

The Development of Genre Knowledge
This subsection discusses the influence of the level of education on the genre knowledge development that is realized in the writing of research abstracts. The initial hypothesis of this study is that there is an influence from the level of study towards the development of genre knowledge manifested in their research writing (Alamargot & Fayol, 2009;Berninger et al., 2002;McCutchen, 1996McCutchen, , 2000. From the analysis, it can be concluded that abstracts in both languages have some development in genre knowledge manifested in abstract move patterns, especially in Move 1 (Introduction). As seen in Figure 2, the steps in Move 1 (Introduction) are more varied in dissertations such as E2 (E1 21% < E2 30%) and I2 (I1 8% < I2 22%).
Interestingly, the writers in each field of study learned to write research abstracts mainly through self-study and peer discussion. The development of genre knowledge through this procedure is further supported by Freedman's (1987) study that demonstrated genre learning as a gradual occurrence that comes without, or with little, explicit instruction. Artemeva (2005Artemeva ( , 2008Artemeva ( , 2009) also noted that genre knowledge is a result of the combined effects of various condiments that include, but are not limited to, classroom and workplace practices.
Comparing the genre knowledge development in the respective fields of study (English and Indonesian), there are no significant differences across the two languages. This implies that no notable differences were identifiable in research writing and learning in any field of study in Indonesia. The participants added that they adopted the culture of academic writing, especially final papers, primarily as part of the assignments from their lecturers. They also regarded final papers as a requirement to graduate. To recap, the abstracts in both languages under examination evinced some signs of genre knowledge development as manifested in abstract move patterns. This development; however, seems to converge across two distinct languages and two different fields of study, Indonesian and English.

CONCLUSION
The main conclusions can be summarized based on the study. In terms of move occurrences across the field and level of study, the number of Move 2 (Purpose) was the least compared to the other four moves in all final papers in both languages. Meanwhile, all final papers had different moves with the highest occurrences. Move 4 (Findings) had the highest number of occurrences for I2 and E1. For I1 and E2, Move 3 (Methods) and M1 (Introductions) occurred the highest, respectively. In terms of manifestation of step occurrences, final papers in both languages showed similarities in Step 3 of Move 1 (Defining the key term(s)), with the lowest occurrences out of all steps. Final papers in both languages also rarely used Move 5, Step 2 (Evaluating the significance of the research), and Step 3 (Stating limitation).
Looking into the linguistic features, past tense dominated the occurrences in abstracts written in English. As discussed in the results section, the dominant use of past tense, especially in Methods and Findings moves, is supported by some previous studies. However, present tense is used chiefly in Vietnamese RA abstracts, nearly twice as much as passive voice. Concerning the use of voice in abstracts, both languages show similarities in the use of active voice. The use of active voice in abstracts implies that the authors tried to focus on the agent rather than the action in the sentence.
From the study, it can be seen that the development of genre knowledge could be reflected through the educational level, in this case from master's to doctoral degree. It is shown in the dissertations in both languages, Move 1 (Introduction) had more various steps in its manifestation. The genre knowledge was realized and developed through self-study and peer discussion. Since there were no significant differences between the development of genre knowledge in the two fields of study, this suggests that research writing and learning may be uniform in any field of study in Indonesia. Indeed, such a claim warrants further inquiry.
Despite the limited sample used in this study, the findings could be helpful to reveal the characteristics of English and Indonesian languages and fields of study. It is also believed that the findings could help students in the future to write abstracts in research writing. However, given the small sample used in this study, it is difficult to generalize the results to all situations. A larger corpus will provide a complete contrastive picture of abstracts' rhetorical and grammatical features in English and Indonesian. Pedagogically speaking, this study resonates with Kurniawan and Lubis (2020)'s recommendation that data-driven genre-based pedagogy (Peacock, 2002;Quinn, 2014) is to be entertained in teaching (English and Indonesian) academic writing for publication purposes.