Meta-discourse markers in the book reviews published in ISI and non-ISI journals of applied linguistics

Abstract Meta-discourse markers constitute a significant part of a reader friendly text. This study focused on the types and frequency of meta-discourse markers in applied linguistics book reviews (BR) published in ISI and non-ISI journals. To this end, meta-discourse markers were analyzed in 86 BRs selected from six journals (three ISI and three non-ISI journals). According to this model, meta-discourse markers include two main groups: interactive and interactional elements. Using AntConc text concordance software program, the meta-discourse markers were checked in the corpora. To check the type and number of meta-discourse markers, frequency analyses were carried out. Chi-square test results and frequency analyses showed that significant differences existed between meta-discourse markers used in BRs published in ISI and non-ISI journals and BR authors used more meta-discourse markers in ISI BRs. Additionally, comparing to non-ISI BRs, ISI BR authors used more evaluative patterns and interactional elements to develop their academic writing. Implications, limitations, and suggestions for further studies were discussed in detail.


PUBLIC INTEREST STATEMENT
Book review articles are a type of academic genres featuring summarized text books wherein authors adopt a critical stance to boil down key points of a textbook. To render a dry full-text book into a reader-friendly and coherent prose, an author needs to utilize a wide range of hedging and linguistic resources (meta-discourse markers) to project his attitude into the discourse and express his viewpoint toward the content and audience of the textbook. This research compared the way of using meta-discourse markers in BR published in ISI journals (journals indexed in Web of Science) and non-ISI journals to see if the quality and quantity of using metadiscourse markers play any role in publication of BR papers in credible journals. Results of this study raise scholars', in particular emerging scholars', attention about the importance of meta-discourse markers in successful publications in prestigious journals.

Introduction
As an elusive term, meta-discourse markers refer to elements that authors use to manage the text and create a reader-friendly and interactive text (Dahl, 2004). In a more precise definition, K. Hyland (2005) alluded to meta-discourse markers as " various linguistic tokens employed to guide or direct a reader through a text so that both the text and the writer's stance is understood" (p. 18). Meta-discourse markers serve different purposes in a text: elaborating on the authors' attitude (Fuertes-Olivera et al., 2001), monitoring and shaping the texture and discourse (Richards & Schmidt, 2010), and maximizing the intelligibility of communication and giving credence to the communities of practices (Zarei & Mansoori, 2011). That said, meta-discourse markers are used in a text to express the meaning and show the viewpoint from which an idea is reflected upon (K. Hyland, 2005).
A growing interest into doing research about meta-discourse markers in different genres has been observed during the last decade and many studies focused on meta-discourse markers to scrutinize pragmatic, and aspects related to beyond the sentence level. Given the status of metadiscourse markers in creating an intelligible and interactive text, increasing interest has been shown in analyzing and examining these elements in different writings (e.g., research article (Dahl, 2004), PhD dissertations, textbooks (Crismore, 1989, etc). Among different genres studied thus far, few publications have focused on the rhetorical patterns and meta-discourse markers in book reviews.
Book review papers fall in academic genres (K. Hyland and Diani (2009). Book review (BR hereafter) articles perform a leading role in selecting the books to purchase (Carvalho, 2002), and helping graduate and novice researchers keep abreast of the latest published books in the field of study (K. Hyland, 2000b;Suárez & Moreno, 2008). Reading full texts could be a demanding job for most of the students, BRs could tackle this problem, provide necessary information, and help selecting informative books that worth reading (Carvalho, 2002). Further, K. Hyland (2000b) explicated that BRs are "a site where the interpersonal stakes are much higher" given the inherent evaluative and interactional nature of this genre (p. 41).
BRs, as Swales and Feak (1994) argued, possess intentionally evaluative characteristics and authors should be cautious in terms of the language and the rhetorical markers to be used in reviewing the books. Given the evaluative nature of BR papers, BRs are recognized as interactive genres that carry social implications. In a similar vein, K. Hyland and Diani (2009) maintained that, "negotiating social interactions in review genres [such as BRs] can therefore mean charting a perilous course between critique and collegiality, minimizing personal threat while simultaneously demonstrating both disciplinary membership and an expert understanding of the issues" (p. 8).
The passion of publishing papers in ISI journals motivated this study to see if the type and frequency of meta-discourse markers play a distinguishing role and affect the quality of BR papers to certify them for ISI journals. Merits of ISI journals encourage scholars to have the ambition of publishing papers in journals indexed in web of science (ISI journals). For example, comparing to non-ISI journals, most of ISI journals offer meticulous review process, high quality copy editing and free of charge publication which convince academia to deem ISI indexed papers as top and more valid investigations. That said, most of universities and academia to consider papers published in ISI journals as more reliable yardstick to grade scholars and offer job promotion. With all these arguments backing up why scholars are interested in publishing in ISI journals and the scarcity of the research on meta-discourse markers in BR papers in mind, the current study aimed at examining the types, use and the frequency of meta-discourse markers in the BRs published in ISI and non-ISI journals of applied linguistics.

Review of the related studies
Meta-discourse markers in advanced writings elucidate writers' attempt to "present and negotiate propositional information in ways that are meaningful and appropriate to a particular disciplinary community" (K. Hyland, 2004, p. 136). Steffensen and Cheng (2016) studied the role of teaching meta-discourse in writing skill and they found that correct use of meta-discourse markers improves writing significantly.
Different models of meta-discourse markers have been conceptualized. VandeKopple (1985) proposed an early model of meta-discourse markers to classify different types of interpersonal and textual markers. Based on this model, seven categories were defined for meta-discourse markers: text connectives (e.g., however); (2) code glosses (e.g., this means that); (3) illocution markers (to conclude); (4) narrators; (5) validity markers (hedges, emphatics, and attributors); (6) attitude markers (surprisingly); (7) commentaries (you might not agree with that. However, this taxonomy was not user friendly and its subcomponents showed considerable overlaps and vagueness (K. Hyland, 2005). To ease the problem, K. Hyland (2005) proposed a framework to classify meta-discourse markers into two broad types: interactive and interactional markers. Interactive markers (K. Hyland, 2005, p. 49) refer to the elements that "shape and constrain a text" and assist authors in writing more convincingly. Interactional markers are used to display the authors' emotion and attitudes, and influence the readers. K. Hyland (2005) pointed to interactional markers as "writer's efforts to control the level of personality in a text and establish a suitable relationship to his or her data, arguments, and audience (p. 139)." Regarding the contribution of meta-discourse markers in writing intelligible and well-organized text, a growing body of recent investigations has focused on these elements in diverse fields of research. K. Hyland and Tse (2004) examined meta-discourse markers in different fields of study: computer science, engineering, business studies, applied linguistics, and public administration. Results of their study showed that authors of soft-sciences use more meta-discourse marker. They concluded that "meta-discourse" is an aspect of language which provides a link between texts and disciplinary culture, assist in defining rhetorical context by revealing some of the expectations and understandings of the audience for whom a text was written" (K. Hyland & Tse, 2004, p. 175). Types and functions of meta-discourse markers in different genres including newspapers, Persian news reports (Hashemi & Golparvar, 2012), academic lectures (Eslami & Eslami-Rasekh, 2007), have been studied thus far. Metadiscourse markers in different sections of research articles (RA) have been the focus of previous studies (e.g., meta-discourse markers in abstracts by K. Hyland (2000a) and Samarj (2005), discussion and conclusions (Yang & Allison, 2003). The focuses of these studies were mainly the role of metadiscourse markers in different parts of RA.
In 2004, Dahl conducted a study into the meta-discourse markers in RA by English and French writers. Findings of that study showed that the language and the culture of the authors could affect meta-discourse markers. Estaji and Vafaeimehr (2015) studied the type and frequency of meta-discourse markers in introduction and conclusion of electrical and mechanical engineering RA. Using Chi-square and manual corpus analyses, they concluded that significant differences existed in frequency and types of meta-discourse markers in these RAs.
Meta discourse-markers in thesis writing have been the topic of many investigations. For example, Marandi (2003) found that only some meta-discourse markers in different chapters of master theses were used differently.
Comparing to other genres, analyses of meta-discourse markers in BRs published in different journals have not been the focus of many lines of inquiry. One of the early attempts to unfold the role of meta-discourse markers, Tse and Hyland (2007) examined the frequency and pragmatic purpose of meta-discourse markers in a corpus of 84 BR articles from three contrasting disciplines and interviews with editors and reviewers. This study showed that meta-discourse markers along with solid background of the subject matter knowledge lead into developing a well-structured and coherent book review. Patterns of meta-discourse markers also depended on the field of the study. For example, in philosophy authors tend to use more extensive metadiscourse markers to express judgmental stance toward the content of the book. In biology, however, meta-discourse markers are used to enhance the readership of the paper and help readers of the book.
Later one, Moreno and Suárez (2008b) designed a framework for comparing evaluation resources across academic texts. The corpus of the study was limited to 20 recently published book review papers in English in prestigious humanities and social sciences. They used a comparative framework and analyzed evaluative markers in BR articles on propositional, and meta-discourse planes. The study highlighted the role of cross-cultural differences in using metadiscourse markers and the effects of culture on using rhetoric markers in BR articles. Using a qualitative design, Moreno and Suárez (2008a) studied 40 BRs published in English and Spanish and found that frequency and type of meta-discourse markers in criticizing the books depended on the language and the social factors.
Mention should also be made of the strand of studies into meta-discourse markers in diverse languages (e.g., Falahati, 2008;Jalilifar, 2011;Zarei & Mansoori, 2011) that mainly lend credence to the impact of contexts on the use of meta-discourse markers in writing. Junqueira (2013) compared meta-discourse markers in book review in Portuguese and English and he found that English BRs were more critical than the Portuguese ones. Junqueira and Cortes (2014) conducted a study on metadiscourse in BRs in English and Brazilian Portuguese. They used a corpus analysis approach to this study and found that interpersonal meta-discourse markers in English BRs were more than the ones in Brazilian Portuguese BRs. In the context of Iran, Famian and Kargar (2013) examined the BRs of linguistics books published in Iran based on K. Hyland's (2005) model. They discovered that the authors used more negative attitude markers in criticizing the books and the gender of the authors was a determinant factor in the types of meta-discourse markers used in BRs.
In a study by Gezegin (2016) checked meta-discourse markers in academic BR in Turkish and English BR articles. Using a corpus analyses method, this study reflected the significant use of meta-discourse markers in English BR articles. The study confirmed positive role of hedging in toning down the sentences in academic genres that was highlighted feature of English BR papers.
Ebrahimi (2018) compared meta-discourse markers used in reading comprehension books published by Oxford publication and similar books published in Iran. This research indicated that comparing to the books published in Iran, more meta-discourse markers are used in the books published by Oxford University. That said, Oxford reading comprehension textbooks are more reader-friendly.
In a recent study, Tabatabaee Lotfi et al. (2019) used Hyland's model of meta-discourse markers to examine the effect of first language and compare the use of rhetorical patterns in two different cultural settings of Chinese and Iranians. Results indicated that Iranian and Chinese learners showed significant differences regarding the use of meta-discourse markers in argumentative essays. This finding verifies the role of context in application of meta-discourse markers.
In another study by Shafique et al. (2019), native English researchers and Pakistani researchers' writings were examined to see how meta-discourse markers were used. This study showed that Pakistani writer used interactive markers in their writings.
As discussed earlier, meta-discourse markers in the genre of BR papers published in ISI and non-ISI journals seems an open line of inquiry as the review of earlier studies indicated a lacuna in this regard and this research has been motivated by this gap.

Objectives and research questions
BRs are normally short critical reports of the newly published books and because the authors should write in brief, the types of meta-discourse markers used in this genre play a key role in writing convincingly. However, as mentioned earlier, most of related studied into meta-discourse markers focused on research articles, textbooks, theses writing, and some of the genres including BRs, particularly BRs in applied linguistics, call for more investigations (Burneikaitė, 2008). To address this scant of attention to studies into meta-discourse markers in BRs, the researchers relied on K. Hyland (2005) model for meta-discourse markers to explore the use, the types, and the frequency of meta-discourse markers in BRs published in two major types of journals: ISI and non-ISI journals.
This study could depict of a picture of how and what types of meta-discourse markers authors use to criticize and evaluate a book in a very concise manner and persuade the readers to read the whole book. By doing this, the researchers examined if there are any significant differences in types and frequency of the meta-discourse markers in ISI and non-ISI journals. Hence, this study was an attempt to answer these questions: (1) How does the use of meta-discourse markers employed in BRs differ in ISI and non-ISI journals of applied linguistics?
(2) Are there any significant differences between BRs published in ISI and non-ISI journals in terms of the frequency and type of the meta-discourse markers?

Corpus
According to Nwogu (1997), variation of disciplines generates inconsistency in rhetorical structures. Given that, the corpus of this study was restricted to BRs published in ISI and non-ISI journals of applied linguistics. The data of this study were gathered from six journals (3 ISI and 3 non-ISI). ISI journals were indexed in Thomson Reuters and non-ISI journals were indexed in other database including Scopus, Google scholar, Sage, etc. To control for variations due to chronological order of publication, the researchers selected 86 BRs published from 2013 to 2015. Some of the sampled journals publish BRs every other issue and some of them published BRs in all of the issues.
Each book review was between two and seven pages at most and they were all about the issues related to the language teaching and assessment. The following Table 1 displays more details about the corpus of the study. Samples of the BRs were from different subjects including language testing and assessment, second language acquisition, discourse analyses, sociolinguistics, etc. No restriction was made in terms of the number of authors in BRs, both single authored, and multiple authored BRs were listed. Further, both native speakers' and non-native speakers' BRs were included in the list of corpus of this study.

The model
The researchers adopted K. Hyland (2005) model describing the types of meta-discourse markers as functional elements helping the authors "express a viewpoint and engage with readers as members of a particular community" (K. Hyland, 2005, p. 37). The cornerstones of this model are the author-reader relations and the manner of creating internal cohesion and coherence. Two main subcomponents of this model were interactive and interactional resources.

Interactive resources include: (a) transitions, (b) frame markers, (c) endophoric markers, (d) evidential, and (e) code glosses. The interactional markers are: (a) Hedges, (b) booster, (c) attitude markers, and (d) elf mentions and engagement markers.
While interactive meta-discourse markers are employed to create a well-organized and interactional meta-discourse markers are used to help the authors to divulge their feelings and emotion towards a text.

Interactive meta-discourse markers are divided to the following categories
(1) Transitions: are the markers that help smooth flow of the writing (e.g., and, but, thus) (2) Frame markers: help topic shift in the discourse and label the information is a specified frame (my purpose, in conclusion).
(3) Endophoric markers help the reader find the reference of what is mentioned in different parts of the text (e.g., as mentioned in chapter 2, as seen in the table).
(4) Evidential markers point to the source of evidence (according to, Y mentioned that) (5) Code glosses are used to clarify and explicate the authors' intended idea (that is, put it more specifically)

Interactional meta-discourse markers
(1) Hedges are used to reflect that the authors shift the responsibility of suggested meaning (maybe, perhaps, possible) (2) Boosters shows the authors confirm the presented ideas (e.g., in fact, definitely) (3) Attitude markers show the authors' attitude about the text (interestingly, fortunately).
(4) Self-mention markers are used to announce the presence of the authors in the texts. These markers are used to create an intimate relationship between the readers and the author (I, me, our, we) (5) engagement markers are used to show how the authors addresses the readers in a text (look at, consider, note that)

Procedure of the study
Meta-discourse markers include large categories. To have a clear picture of the types of metadiscourse markers to be checked for, the classifications proposed by K. Hyland (2005) were adopted in this study based on which a list of the most recurrent meta-discourse markers was developed. Through word for word analyses, 18 BRs selected from the journals (three from each corpus) and a list containing 640 meta-discourse markers was derived to ease the analyses of the corpora. In the following examples of the meta-discourse, markers are displayed: Transitions: whereas, on the contrary, however, additionally, since.
Frame markers: the other, regarding, in conclusion, secondly Endophoric markers: as seen, above mentioned, earlier, the following list Evidential: based on, according to X, Engagement marker: refer to …, it should be noted that … ., notice that …, .
All BRs were checked electronically by AntConc 3.2.1 software. Then 20% of the cases were randomly chosen to examine manually the type and use of the meta-discourse in the context. To check the reliability of the analyses Kappa index was measured (.85). To calculate the frequency of the meta-discourse markers in the BRs, the following method was used:

F per 1000 words = meta-discourse markers ÷the total number of the words in each BR˟ 1000
To measure the differences in meta-discourse use in BRs published ISI and non-ISI journals, Chisquare test was employed. According to Hatch and Farhady (1982), if the obtained Chi-square is more than 3.84 at one degree of freedom, the observed difference is considered significant. In the current study, one degree of freedom was set for the comparisons and the 0.05 level of significance was considered.

Results and discussion
To answer the first question "How are meta-discourse markers used in BRs published in ISI journals of applied linguistics?", the meta-discourse markers in ISI and non-ISI journals was checked in AntConc program. To provide a clearer picture of the analyses in the BRs, the whole number of the words and the number of the meta-discourse markers in each corpus was displayed in Table 2. In total, 87/7 meta-discourse markers were used per 1000 words in the corpus of ISI journals and 73/ 7 meta-discourse markers were used in the corpus of non-ISI journals. The frequency of metadiscourse markers was also calculated for each journal (see Tables 3 and 4).
The number of meta-discourse markers in ISI journals was larger than the ones employed in non-ISI journals. This suggests that ISI BR authors made use of more meta-discourse markers to write an evaluative summary of the newly published books. That is, regarding the credibility of ISI journals and high standards of these journals for academic publications, ISI writers utilized metadiscourse markers in a larger amount and probably in a more appropriate manner to develop a well-organized BR and get it published in prestigious journals indexed in ISI web of knowledge list (Jalilifar & Moazzen, 2014).
In both corpora, the frequency of the interactive markers was higher than the interactional ones. The results of this study showed that BR authors using more interactive markers to organize the discourse of their BRs tried to convince their reader and help them read the text more understandably and quickly. This could be justified by the very specific format of BR requiring author to follow word limitation of the journal and develop a brief summary of the most important points discussed in the book.
To meet the particular authorship guidelines of BR writing, author need to prioritize the readers' interest, needs, and accommodate the writing style to put the arguments as acceptably as possible (K. Hyland, 2005).
On the contrary, interactional markers are used to create interaction between the author and the reader and help the authors signal their identity and position in the writing (K. Hyland, 2005). Hence, authors' reliance on interactional markers implies the primacy of the authors' ideas than the   convenience of the readers. Inevitably, the issue of word limitation in BR makes the authors overshadow their personal stance, focus on the readership of the BR, and make use of fewer interactional makers.
The results of this study also showed that comparing to non-ISI journals, authors of BR in ISI journals used larger number of interactional markers (F = 15). Allegedly, non-ISI BR authors of the selected corpus showed slight tendency towards explicating themselves and used fewer interactional markers in their writing. This might be due to dependency of the meta-discourse on authors' socio rhetorical context). Most of BRs published in non-ISI journals are written by Asian authors from Asian cultural background whose propensity towards implicitness has been vouched in earlier studies (Jalilifar et al., 2013;Tirkkonen-Condit, 1996). However, English native speakers penned most of BRs in ISI journals containing more interactional markers. This piece of finding was in congruence with Moreno and Suárez (2008a) and Giannoni (2006) demonstrating that English writers use more personal markers authors to evaluate books.
To answer the second research question: Are there any significant differences between BRs published in ISI and non-ISI journals in terms of the frequency and type of the meta-discourse markers, frequency of the meta-discourse markers was measured and Chi-square tests were conducted to measure whether there existed any significant differences among ISI and non-ISI journals in terms of meta-discourse markers.
As shown in Table 4, BRs in ISI journals presented higher frequency in all types of markers and the transition markers displayed the maximum frequency per 1000 words in both ISI (F = 60)and non ISI journals (F = 53/3). Transition markers in BR were used to reveal the semantic connection among the ideas and sentences. See this example: "Moreover, regarding challenges and concerns for L2 writing pedagogy, both teachers pointed to the lack of writing instruction. In addition, both teachers claimed that they never created or even acted out the role of a writing teacher" (non-ISI journal, IJLTR). Evidential markers in non-ISI journal were less frequent than ISI journals. After transitions, Endomorphic markers were the most frequently used elements in both corpora. In particular, the word "chapter" was most repeatedly used in most of the BRs. Here are examples: "Chapter 5 is devoted to the speech act theory that studies how language is used to perform communicative acts (RELC, non-ISI journal).

Zou's chapter on China is in fact different from all other chapters, as it presents a more empirical investigation based on interviews of four policy-makers" (Language teaching and research).
In the subcategory of interactional markers, hedges represented the highest frequency (see Table  4). K. Hyland and Tse (2004) found that hedges were frequently applied in academic dissertations to distinguish facts from the authors' personal claims. This piece of finding confirmed Junqueira and Cortes' study (Junqueira & Cortes, 2014) who found that hedges were the most frequent elements in English and Brazilian Portuguese BRs.
The lowest frequency belonged to the engagement markers in the both corpora. Finding of this study indicated that BR authors inclined to stick to the transferring the ideas expressed in the books to be reviewed rather than expressing their views through engagement markers. Additionally, BR authors of both corpora used less number of attitude markers in comparison to other interactional elements. Yet, in comparison to non-ISI BRs, attitude markers appeared to be more frequent in ISI BRs.
Results of Chi-square tests at 0.05 level of significance showed significant differences between ISI and non-ISI BRs in terms of transition markers, endomorphic markers hedges, attitude markers, and self-mentions. However, no significant differences were observed in the use of other meta-discourse markers as their Chi-square values were less than the critical value of 3.84 at the 1 degree of freedom and 0.05 level of significance. In total, there was a significant difference between ISI and non-ISI journals in the frequency and use of meta-discourse markers (4.50) (see Table 5).

Conclusion
BR is one of the most complicated genres demanding high capacity in writing a critical review in which the authors should sustain a balance between simple description of the main themes and weighting up the pros and cons of the book (K. Hyland, 2000b). This study focused on the frequency and use of meta-discourse markers across BRs published in ISI and non-ISI journals. Using the meta-discourse model by K. Hyland (2005), the researchers compared meta-discourse markers in six journals (three ISI and three non-ISI). Frequency analyses and Chi-square tests showed variations in how BR authors used meta-discourse markers to get their academic writings published in ISI and non ISI journals of applied linguistics. Findings of this study suggested that ISI BR authors used more meta-discourse markers (both interactive and interactional elements) than their counterparts in non-ISI journals. This might be due the norms of the journal, the cultural background of the authors, the proficiency of the authors, and or even the audience of the BRs. Contrary to the researchers' expectations, the number of interactional markers were fewer than interactive markers in BR. This implies that BR authors of the selected samples withheld evaluative patterns and or judging their colleagues and preferred to present more description of the book, though interactional markers are not necessarily employed to praise and or criticize. (K. Hyland, 2000b).
The findings of the current inquiry might carry some implications for academic, in particular, BR writing. BR is one the most important and complicated genres requiring a high level of language proficiency and critical thinking and evaluative patterning. To write a wellorganized and acceptable BR, authors need to know how to use meta-discourse markers appropriately. This study might contribute to familiarity of the authors with the significance and status of the meta-discourse markers in academic writing. Via direct and or indirect consciousness about the correct use of meta-discourse markers, authors can improve their writing abilities and finally become a recognized member of their community of practice (Hyland, 2009).
Studies into meta-discourse markers in different genres might illuminate the use of different techniques to develop cohesive and coherent texts. More importantly, the comparison of these elements in credible and type 2 or non-ISI journal could contribute to the novice authors' understanding about the standards of writing and appropriate use of meta-discourse markers to get academic writings published in prestigious journals. This study, similar to other studies, has some limitations. The researchers had access to limited number of journals and only 86 BR were selected from these journals. Future studies should pick more BRs from larger number of journals. Additionally, only applied linguistics journals were examined in this study and it is suggested that future studies focus on other fields of study comparing the BRs of hard sciences and other soft sciences (e.g., psychology, history, etc). In this study the researchers did not considerate the impact of some moderating variables including the language and cultural back ground of the BR authors, it is recommended that future studies deal with these factors in meta-discourse analyses.