Author Engagement in the Literature Review of Research Articles Published in International and Iranian Local Journals: Some Insights from Appraisal Theory

Introduction

One big challenge writers face when writing academic articles is to take up ‘subjective’ or interpersonal meanings of academic discourse. In other words, most novice writers find it challenging to settle themselves properly in interaction with the knowledge they are to present. For instance, they are not certain whether to use explicit subjective expressions like ‘in my opinion’ or expressions of modality such as ‘it appears that’ (Hoye, 2008).

The way writers construct their evaluative stance in their writings is, to a great extent, associated with their choice of words and expressions. Each word or expression conveys a different stance. To construct an argument, developing an evaluative stance is, therefore, an indispensable aspect of academic writing. Hence, focusing on evaluation in advanced academic writing courses can lead to giving out some thoughtful pedagogical interventions and the cultivation of an awareness of how this feature is applied.  Despite the importance of authors' evaluation of the subjects of their articles, the lack of evaluation in student academic writing has been massively noticed and tackled by different scholars (e.g. Hood, 2004; Hyland, 2006).

As with other academic genres, research articles, due to their importance in presenting and spreading knowledge, have been the subject of many researches, and a large  number of studies investigated authors’ evaluation in different sections of research articles such as discussion and introduction (e.g. Hood, 2004; Martínez, 2003; Myers, 1989, etc.). However, one section of research articles that has not been given considerable attention, particularly in relation to engagement of authors, is literature review (LR). For academic writers to be able to publish their manuscripts in academic journals, the ability to write LR section based on the conventions of academic genre is inevitable. Using Martin and White's (2005) Appraisal theory (AT), as its theoretical framework, this study aims to analyze the distribution of engagement resources and identify linguistic resources the authors use to indicate their engagement in the LR section of the articles published in international and Iranian local journals. Since in LR section of articles, writers basically use linguistic resources to express their stance toward the topic they study and refer to other voices, the current study aimed to analyze the engagement resources authors apply in this section of research articles. Familiarity with engagement resources can give researchers an awareness of how to present their stance and refer to other voices in a systematic way in LR section of their articles.

Martin and White's (2005) AT, an offshoot of Halliday's systemic functional linguistics, constitutes the theoretical framework of this study. Appraisal has been defined as “the semantic resources used to negotiate emotions, judgments and valuations, alongside resources for amplifying and engaging with these evaluations” (Martin, 2000, p. 145). In fact, in AT, evaluative linguistic resources are utilized “for negotiating our social relationships, by telling our listeners or readers how we feel about things and people, in other words, what our attitudes are” (Martin & Rose, 2003, p. 19). For appraisal, three interacting domains were recognized by Martin and White (2005): ‘attitude’, ‘engagement’ and ‘graduation’. Attitude, according to Martin and White (2005, p.35) "is concerned with our feelings, including emotional reactions, judgments of behavior and evaluation of things, and graduation, grades phenomena whereby feelings are amplified and categories blurred". Engagement, which is the focus of this study, is “a reader-oriented aspect of interaction which concerns the degree of rapport which holds between communicative participants” (Hyland & Jiang, 2016, p. 29).

Figure 1. Diagrammatic Representation of Appraisal Theory, taken from Martin and White (2005).

As illustrated in Figure 1, engagement is divided into monoglossic and heteroglossic assertions. The distinguishing feature between them is that unlike heteroglossic assertions, which recognize and allow for dialogistic alternatives, monoglossic assertions are facts which do not allow for referencing other voices and dialogisticpotentials in written texts. Sentence 1 below is an example of a monoglossic assertion. In this sentence, there is no linguistic resource to invoke alternative voices while in sentence 2, as an example of a heteroglossic assertion, the expression ‘in my view’ shows that there are other viewpoints in addition to the author’s viewpoint.

1. The banks have been greedy.

2. In my view, the banks have been greedy.

Heteroglossic or dialogic assertions are further categorized into dialogic contractiveand dialogic expansive resources. Contractive resources can be considered as a subtype of heteroglossic resources due to the fact that they contain multiple voices. However, what distinguishes them from other heteroglossic resources is that they do not allow for certain dialogic alternatives, or alternatives which are not in accord with them. In other words, “they explicitly refer to utterances and viewpoints of external voices, and they close down the space for dialogic alternatives” (Martin & White, 2005, p. 103) (e.g. sentences3-5 below).

Contractive dialogical evaluation can be done either through the use of disclaiming or proclaiming resources. Disclaiming resources serve the purpose of presenting a contrary position through aligning the reader with the opposing views and is done through denial (negation) and countering. Negation, according to Martin and White, is a resource for invoking alternative positive positions into the dialogue, hence acknowledging them so as to reject them. Countering resources, on the other hand, include formulations which represent the current proposition as replacing or supplanting and thereby ‘countering’, a proposition which would have been expected in its place (2005, p. 118 ). Sentence 3 is an example of negation, and sentence 4 is an example of the use of countering resources.

3. You don’t need to give up potatoes to lose weight.

4. Although he ate potatoes most days, he still lost weight.

In sentence 3, the author invokes the expected alternative proposition that you need potatoes to lose weight but through using ‘don’t’ rejects it directly. In example 4, the author using ‘although’ is invoking the alternative position and in this way he is acknowledging it; however, in the next sentence, he is rejecting it.

Proclaiming assertions, made by presenting compelling or valid propositions, can be realized through using three types of resources: concurrence, pronouncement and endorsement resources. Through the use of the concurrence resource (e.g. ‘typically’ in sentence 5), the author overtly states that he agrees with the voice presented and at the same time excludes the voices with which he is not in agreement.

1. 5.  ERT istypically prescribed by physicians experienced in treating Gaucher disease.

Pronouncement resources refer to assertions in which the author intervenes overtly into the text so as to assert or emphasize the value of the proposition. Through the use of endorsement resources, authors make reference to external sources which are presented as unquestionable, valid or correct.

Expansive resources explicitly “ground the proposition in an individualized, contingent subjectivity, that of some external source” (Martin & White, 2005, p.114). Put it differently, “they ground the viewpoint conveyed by the proposition in an explicit subjectivity thereby signalling that it is individual and contingent and, therefore, but one of a range of possible dialogic options” (Martin & White, 2005, p. 113) (e.g. sentences 6-8 below).

Expansive dialogical space is realized through ‘entertainment’ and ‘attribution’ resources. Entertainment resources like, ‘less likely’ in sentence 6 below, refer to utterances where an individual himself presents the proposition within a range of possible alternatives or possibilities through one of the following linguistic resources: it seems, the evidence suggests, apparently, perhaps, probably, maybe, it’s possible, in my view, I suspect that, I believe that, probably, it’s almost certain that, as well as modal verbs such as may, will and must.

6. Adolescents from more cohesive families are less likely to smoke or use substances.

Attribution resources, by contrast, evoke an external voice to carry the subjectivity of the proposition. Attribution resources offer two ways in which the authorial voice is positioned in relation to the external source introduced either by expressing alignment with it - acknowledging, typically through the use of a reporting device like ‘Mejia (2002) asserts’ in sentence 7 below or by establishing a distance from it through some reporting verbs, particularly through the use of the verb ‘claim’.

7. Mejia (2002) asserts that language is a symbolic source signifying linguistic skills, cultural awareness, and particular abilities.

Acknowledgment, as the underlined part of sentence 8 shows, can be realized through citing the name of authors without using expressions like, ‘he asserts’.This way of realization of acknowledgment is common in LR section of studies.

8. Children learning their L1 set up numerous categories of objects prior to the produc­tion of any words (Clark, 1993, p. 44).

Figure 2 is the diagrammatic representation of heteroglossic resources.

Figure 2. Diagrammatic Representation of heteroglossic Categories of AT, adapted from Martin and White (2005, p.134).

To Martin and White (2005), heteroglossic resources can also be classified from the perspective of the sources of authors’ voices. The resources such as endorsement and attribution, which authors use to make reference to external voices are called extra-vocalizing resources. For instance, in sentence 7 above, the author used acknowledgement resource ‘Mejia (2002) asserts’ to make reference to an external voice (Mejia’s (2002) voice). On the other hand, the resources such as disclaiming, pronouncement, concurrence, and entertainment which authors use to convey their own voice are referred to as intra-vocalizingresources. In sentence5 above, as an instance of intra-vocalizing resource, the author used the concurrence resource ‘typically’ to express his own voice not an external one.

Literature Review

AT has been applied, as the theoretical framework of research, by many scholars (Hyland & Jiang, 2016; Salom & Monreal, 2014; Xie, 2016, etc.) to different genres such as essays, articles, etc. An exploratory observation of evaluation using Appraisal Theory in undergraduate students’ argumentative essays was conducted by Wu and Allison (2003). Forty undergraduate students’ essays ranging from high-rated to low-rated ones were selected as the corpus of the study. The purpose was to find out how students convey evaluative meaning in the claims they make in their essays, and whether differences in using evaluative choices of wordings have any effect on the overall quality of essays. The findings of the study indicated that differences in the degree and manner of using evaluative resources do not have any role in the overall quality of the essays. The findings of Wu and Allison’s (2003) study can be explained by Coffin’s (2009) claim that the ways of using evaluating resources are not given considerable attention by novice writers, and this might be due to the fact that scant attention is paid to pedagogical integration of evaluation resources in academic writing courses.

Hood (2004) using AT, developed by Martin and Rose (2003), studied the use of attitude resources in published articles and the texts produced by Hong Kong undergraduate students. He found that attitude of authors is predominantly expressed through the use of appreciation resources. Hong Kong students, however, made their texts more personal and subjective by making more frequent use of judgment and affect than the authors of published articles. Hood’s (2004) study can be considered as a seminal article using AT as a means of investigating evaluation in texts. However, only one aspect of evaluation, that of attitude, was taken into consideration.

Another study, which, like Hood (2004), focused on attitude as one of the dimensions of AT, is that of Liu and Thompson (2009). In their study, the use of attitude resources by Chinese students in writing argumentative texts in both English and Chinese were studied. A significant difference was found in terms of using affect and judgment in English and Chinese students’ essays. This difference was attributed to the difference between Western and traditional Chinese writing rhetoric. Regarding the use of appreciation, however, the study revealed no significant difference.

Engagement, as one aspect of AT, was put in the spotlight by a number of researchers. For instance, Wu (2007) investigated the use of engagement resources in high and low-rated Geography undergraduate essays written by Singaporean students. The results of Wu’s (2007) study revealed that monoglossic assertions are by far more prevalent in low-rated essays. Furthermore, in general, the writers of lower-rated essays tended to use contractive resources more often than expansive ones. However, writers of high-rated ones showed a greater tendency to use expansive resources. In particular, the use of pronouncement resources was more frequently observed in low-rated essays resulting in a stronger tendency to insist on the validity of one’s proposition. Disclaiming resources such as countering and denying; on the other hand, were rarely used by the writers of low-rated essays. By contrast, among the contractive resources, the writers of high-rated essays made frequent use of endorsement resources to show their alignment with a source voice. The findings of Wu’s (2007) study sheds some light on the use of evaluative language in essays and offer valuable new insights into the management of the array of evaluative language choices available when writing essays to achieve particular rhetorical effects. Wu’s (2007) study, however, was limited to just one academic discourse, i.e. geography essays; therefore, any generalization should have been taken cautiously.

The authors’ use of attitude and graduation, as two constituents of appraisal theory, were investigated by Jalilifar, Hayati, and Mashhadi (2013) in the introductions of 80 English language teaching (ELT) research articles written in English by Iranian and international writers. The research aimed to study how Iranian and international writers use attitude and graduation resources in their articles. The findings indicated that international writers made more frequent use of attitude and graduation resources than Iranian writers. They attributed their findings to the strong tendency of Iranian writers not to show conformity to the conventions of their international discourse community. Jalilifar et al. (2013) limit their study to identifying explicit attitude resources; however, implicit or invoked attitude resources were not taken into consideration.

The relatively scarce use of attitude resources by Iranian authors of academic texts was also noted by Gholami, Rafsanjani Nejad, and Looragi Pour (2014). Investigating the conclusion section of Iranian undergraduate students’ essays, Gholami et al. (2014) found that few incidences of attitude resources are employed by Iranian students in these sections. In Gholami et al.'s (2014) study, attitude resources were not classified according to a systematic comprehensive framework; rather, the traditional attitude resource classification made by Crismore, Markkanen, and Steffensen (1993) was employed. Therefore, not all types of attitude resources were taken into account.

Like Wu (2007), Geng and Wharton (2016) conducted a study to investigate how engagement resources are used by L1 (first language) Chinese and L1 English writers in their discussion section of doctoral dissertations. Their findings revealed no significant difference across the patterns of evaluative language choices, leading to the conclusion that L1 cannot be seen as an influencing variable in making evaluative language choices.

Exploring the frequency of appraisal resources in 40 English nutritional science articles in four popular sources, WebMD, Better Nutrition, Science Daily, and New York Times, was the aim of a study conducted by Babaii, Atai, and Saidi (2017). All three categories of appraisal theory (attitude, engagement, and graduation) were taken into consideration in their analysis. Attitude resources were found to be the most frequent resources and engagement resources were the least. Appreciation resources were the most frequent of attitude resources, force rather than focus resources were used more frequently in graduation ones, and the frequency of heteroglossic resources was higher than monoglossic in engagement resources. The findings indicated that the authors of science articles use appraisal resources to inject their feelings about their discoveries. Being limited to the field of Nutrition and to a small sample of 40 articles, Babaii, Atai, and Saidi’s (2017) findings cannot be generalized to other fields of studies, especially those of soft science.

More recently, Tavassoli, Jalilifar, and White (2018) applied AT to investigate the attitudinal representations of commentary articles published in two British newspapers, The Guardian and The Telegraph, which have different political orientations towards the Syrian refugee crisis. Their findings reflected the welcoming stance of the left-leaning Guardian and the lack of welcome from the right-leaning The Telegraph towards the Syrian refugees.

While, like afore-mentioned studies, many other studies (e.g. Brooke, 2014; Hood, 2004; Hyland & Jiang, 2016; Loi, Lim, and Wharton, 2016) investigated different aspects of evaluation such as attitude graduation and engagement in their corpus-based works, there has not been any study using AT for the analysis of LR section of academic articles published in international and Iranian local journals. In view of this gap, the focus of the present study is on the LR section of research articles published in international and Iranian local journals.

Methodology

The Corpus

LR section of eighty psychology and applied linguistics articles as representatives of humanities and mechanical engineering and medicine articles as representatives of hard science articles published between 2008 and 2017 in the international and Iranian local journals formed the corpus of the study. The reason for the selection of these disciplines is because medicine and mechanical engineering, based on Becher’s (1989) taxonomy and definition of hard and soft science discipline, have a more scientific, less humanistic content while psychology and applied linguistics have more humanistic content than scientific (Becher, 1994). Therefore, considering this difference, a different distribution of engagement resources is also expected in these categories of science. Forty of the articles were chosen from international journals and the rest were selected from Iranian ones. The international journals include six applied linguistics, five psychology, three mechanical engineering, and four medicine journals. Iranian local journals consist of four applied linguistics, four psychology, three mechanical engineering, and three medicine journals. Ten articles from international journals of each science and ten from local journals of each science were randomly selected.

The international journals chosen for the study were Language Teaching Research, Journal of Language & Education, Language Learning, Applied Linguistics, Research in the Teaching of English, System, Perspectives on Psychological Science, Journal of Family Psychology, Psychological Science in the Public Interest, Journal of Homosexauality, Journal of Experimental Psychology, Psychological Science, International Journal of Mechanical Sciences, Archives of Civil and Mechanical Engineering , Progress in Energy and Combustion Science, European Journal of Internal Medicine, The Journal of Experimental Medicine, Molecular Therapy, and The New England Journal of Medicine.

Iranian local journals included Journal of English Language Teaching and Learning (University of Tabriz), Iranian Journal of Applied Linguistics (IJAL), Iranian Journal of Language Teaching Research (Urmia University), Journal of Research in Applied Linguistics (RALs), Iranian Journal of Psychiatry, Iranian Psychological Association, Practice in Clinical Psychology, Iranian Journal of Psychiatry and Behavioral Science, Iranian Journal of Hydrogen & Fuel Cell, Iranian Journal of Materials Science and Engineering, Iranian Journal of Science & Technology, Iranian Journal of Public Health, Iranian Journal of Basic Medical Sciences (IJBMS), Iranian Journal of Medical Sciences (IJMS), and Archives of Iranian Medicine.

The criteria for the selection of the journals were popularity, rank and the indices of the journals from which the articles were selected.

Engagement resources within the LR section of research articles were marked and labeled manually so as to form the basis of the analysis. As stated earlier, engagement resources of AT were selected for the coding in this study. Therefore, first of all, the decision was made as to whether the proposition in question is monoglossic or heteroglossic. Then, the subcategories of heteroglossic resources were recognized based on Martin and White’s (2005) taxonomy. An example of coding the data is elaborated using sentence 9 below.

9. Varcocile is an abnormal enlargement of the testicular vein (pampiniform venous plexus) above and around the testes. [Mono]. This abnormality is probably one of the most frequent reasons for low sperm production and decreased semen quality. [entertainment] (Iranian Journal of Basic Medical Science, 2016; 19, p. 1279).

The first sentence of the example above contains no heteroglossic linguistic resource to make reference to other voices; hence, it is regarded as monoglossic (Mono). The second sentence, by contrast, is heteroglossic because the word ‘probably’ invokes alternative viewpoint by indicating that the proposition is within a range of possible alternatives and is, hence, regarded as an entertainment resource.

After conducting a pilot study, it was found out that the appropriate unit for the coding of engagement resources in this study is sentence. The reason lies in the fact that sentence, according to Halliday and Matthiessen (2004), is regarded as “the most extensive domain of grammatical structure” (p.436). If no subcategory of heteroglossic resources were found in a sentence, then the sentence was considered monoglossic; otherwise, the sentence was viewed heteroglossic and might have contained one or more categories of heteroglossic resources. Additionally, the total frequency of each category and subcategory of engagement was recognized with AntConc freeware concordance program. Through putting the identified resources in the world list of the program, it shows the total frequency of the resources and the contexts in which they occur. Figure 3 shows the frequency of the modal verbs (166) as entertainment resources in articles published in local journals.

Figure 3. An Example of AntConc Freeware Output

Finally, to ensure the reliability of the coding process, an expert familiar with AT was asked to code engagement resources of the articles for the second time. And the inter-coder reliability was calculated through test-retest quantity-frequency method, and yielded a level of consistency of 92%.

Data Analysis

The Distribution of Engagement Resources in Published Articles

Table 1 represents the distribution of engagement resources per 1000 words in the articles published in local and international journals. Monoglossic assertions contain 20.42% of total engagement resources. In the local and international journals, they constitute 26.67% and 15.36% of cases respectively.

Table 1. The Distribution of Engagement Resources in the Articles Published in Local and International Journals

As Table 1 indicates, one of the major differences across articles published in international and local journals lies in the distribution of monoglossic resources. LR sections of articles published in local journals contained more monoglossic resources in their LR section (26.67% and 15.36% in local and international selected articles respectively). Table 2 represents the distribution of monoglossic resources used in the soft and hard science journals.

Table 2. The Distribution of Monoglossic Resources in Journals of Soft and Hard Sciences

As Table 2 shows, mechanical engineering, as a representative of hard sciences, has the highest percentage of monoglossic resources (11.40 per 1000words). On the whole, monoglossic resources were more frequently used in articles published in hard science journals.

Regarding hetroglossic resources, the frequency of contractive and expansive resources is presented in Table 3. Contractive resources constitute 36.82% of cases while expansive resources make up 42.76% of the whole resources. As Table 3 shows, in the articles published in both local and international journals, the occurrence of expansive resources in LR section is more frequent than that of contractive ones. Table 3 also shows that articles published in international journals made more frequent use of both contractive and expansive resources.

Table 3. The Distribution of Contractive and Expansive Resources LR Section of Articles Published in Local and International Journals

The Distribution of Contractive Resources in Published Articles

Contractive resources are realized through two subcategories: disclaiming and proclaiming resources. Table 1 above shows the differences in the distribution of disclaiming resources across articles published in local and international journals. Firstly, it is evident in Table 1 that the authors of the articles published in international journals use more disclaiming resources (countering and denying) than the local ones. Another marked difference across local and international journals, as shown in Table 1, is in the distribution of proclaiming (pronouncement and concurrence) resources. It is evident from Table 1 that the articles published in international journals employed more pronouncement and concurrence resources (7.53%) than the local ones (4.50%).

The number of contractive resources in the selected disciplines per 1000 words is presented in Table 4. The reason why the results of endorsement resources are presented separately from other proclaiming resources is that unlike the other proclaiming resources (pronouncement and concurrence), endorsement resources are extra-vocalizing ones. That is to say, through endorsement, the intersubjectivity of propositions is presented and attributed to an external voice (White, 2005). As such, the dialogic space is somewhat narrowed down. As Table 4 presents, the frequency of disclaiming resources in articles published in Psychology and Applied Linguistics journals is higher than those of journals of other disciplines (8.35 per 1000 words) and (7.79 per 1000 words), respectively.

Table 4. Disclaiming Resources in LR Section of Journals of Soft and Hard Sciences

Regarding proclaiming resources, as presented in Table 5, concurrence and pronouncement resources  are used more frequently in articles published in hard science journals than those of soft science ones.

Table 5. Proclaiming Resources (Concurrence & Pronouncement) in LR Section of Journals of Soft and Hard Sciences

Table 5 shows that mechanical engineering and medicine articles contain 32.49% and 27.91% of concurrence and pronouncement resources respectively. Articles published in applied linguistics made the least frequent use of incidences of concurrence and pronouncement resources compared to other articles (2.31 per 1000 words). As Table 6 shows, unlike other items of proclaiming resources (concurrence and pronouncement), articles published in soft science journals used endorsement resources more than those of hard science ones. Articles published in local mechanical engineering journal used only 1.30 endorsement resources per 1000 words.

Table 6. Endorsement Resources in LR Section of Journals of Soft and Hard Sciences

The Distribution of Expansive Resources in Published Articles

The result of analyzing expansive resources used in the articles is as follows. Resources used for entertainment, as a dialogically expansive category, make up 33.30% and 33.47% of the resources in  articles published in local and international journals respectively, showing very frequent occurrence of this category in LR sections of articles. Nonetheless, as Table 1 suggests, articles published in international journals have made more frequent use of entertainment resources (9.99 per 1000 words) than those of local ones (7.88 per 1000 words).

Table 1 suggests that in LR section, the authors of articles published in both local and international journals do not employ distancing resources very often in presenting the authorial voice, and attribution of a claim to another author is predominantly done through the use of acknowledgement resources. The frequency of distancing resources in articles published in local journals (0.22 per 1000 words) is lower than that in articles published in international ones (0.38 per 1000 words). As for acknowledgment, the situation is different: articles published in local journals constitute a higher percentage of acknowledgment resources than those in international ones. The distribution of subtypes of expansive resources in journals of the disciplines in question is displayed as follows.

As Table 7 suggests, entertainment resources are much more frequently used in psychology and applied linguistic articles which are the representatives of soft sciences compared with that of mechanical engineering and medicine articles which are the representatives of hard sciences.

Table 7. Entertainment Resources in LR Section of Journals of Soft and Hard Sciences in Question

What is striking in Table 7 is the marked difference between the distribution of entertainment resources used in LR section of articles published in local and international psychology journals. Entertainment resources used in psychology articles published in international journals make up the highest percentage among all cases (14.45 per 1000 words). However, in psychology articles published in local journals, there are far fewer entertainment resources (8.60 per 1000 words). Table8 shows that attribution resources, especially acknowledgement ones were used frequently in all journals of the hard and soft science in both local and international ones.

Table 8. The Distribution of Attribution Resources in LR Section of Journals of Soft and Hard Sciences

Linguistic Representations of Engagement Resources

Linguistic Representation of Monoglossic Resources in LR Section of Published Articles

The results of data analysis indicate that the authors of articles published in local journals put forward their propositions monoglossically which is, according to Richardson (2003), one means of getting credibility and revealing a considerable extent of commitment toward their propositions so as to convince the readers.

It is almost impossible to claim that monoglossic resources are realized through certain linguistic representations. In fact, a sentence containing no linguistic resource allowing for diverse viewpoints is considered as a monoglossic assertion. Nonetheless, some linguistic devices which mark monoglossic assertions are listed below.

Table 9. Linguistic Representations Marking Monoglossic Resources in Articles Published in Local and International Journals

Most of the resources in Table 9 were used by writers of articles published in both international and local journals to report the previous studies related to their work without using dialogic diversity in presenting their propositions.

In sentence 10 below, a mongolossic resource is used to serve the functions of claiming centrality and reporting previous studies. And then, to report the findings of the previous studies, an endorsement resource ‘reveal’ was used.

10. Gass et al. (1999) investigated the effects of task repetition on the linguistic output of L2 learners of Spanish [Monoglossic]. The results revealed that task repetition had an effect on the overall proficiency, accuracy in the use of estar (a Spanish copula verb), and lexical sophistication [endorsement].

(Iranian Journal of Applied Linguistics, 2016; 19(2), p.3-4).

Linguistic Representation of the Authors' Use of Heteroglossic Resources in Published Articles

Disclaiming resources, as stated before, are realized through denying and countering assertions. Table 10 indicates the linguistic resources used for disclaiming assertions used by the writers of the analyzed papers in LR sections of international and local journals.

Table 10. Linguistic Representations for Disclaiming Resources in Articles Published in Local and International Journals

The resources presented in Table 10 are, as sentence 11 shows, used in LR section to introduce propositions as a way to present beliefs and opposition to the beliefs. They are, as sentence 11 shows, also used to evaluate the state of the field by indicating the gap or, in Kwan’s (2006) term, negational strategy.

11. Despitea high rate of initial response, the development of resistance to crizotinib almost invariably leads to tumor relapse [counter].

(Cancer Immunol Res, 2015; 3(12), p. 1334).

In sentence 11, the writer uses conjunctive adverbials ‘despite’, as a counter resource to show his/her proposition as more reliable to readers, at the same time he rejects implicitly one that would be expected

12. There isno systematic study focusing on the effect of annealing and pH values on the surface morphology and structural properties of sol-gel derived TiO₂ nonocrystals. [denial]. (International Journal of Science and Technology, 2014; 38A2, p.188)

Here in sentence 12, the writer used the denial resource’ no’ to invoke the proposition that there is to be a systematic study focusing on the effect of annealing and pH values, but meanwhile rejected the positive voice that there is an expectedly systematic study focusing on the effect of annealing and pH values on the surface morphology and structural properties of sol-gel derived TiO₂ nonocrystals.

In some cases, as in sentence 13, the interplay between countering and denying resources was used to serve the purpose of evaluating the state of the field.

1. 13.  Although many outcome studies have documented the effectiveness of couple therapy in treating both relationship distress and individual pathology, these studies have not evaluated the progression of change over time.[+counter, +deny].

(Journal of Family Psychology, 2015; 29(5), p. 658)

In sentence 13, the writer used the countering resource (although) to indicate that his/her proposition is more reliable to readers and at the same time rejected the one that would be expected through the denial resource ‘not’.  In fact, through the use of the denial resource ‘not’, the author invoked that expectedly these studies have evaluated the progression of change over time, but at the same time rejected it by using 'not’.

Proclaiming (pronouncement and concurrence) resources were found couched in the linguistic representations presented in Table 11.

Table 11. Linguistic Representations for Realizing Proclaiming (Pronouncement and Concurrence) Resources.

As sentences 14 and 15 below show, through the linguistic resources presented in
Table 11, writers provide background information, as a function of LR section, with a high level of credibility and at the same time acknowledge that there are other voices that may have different points of view.

14. Notably, oxide growth under the sensor or weld relaxation can cause rocking and mechanical instability and measuring strain on a curved surface will require special calibration. [pronounce]. (International Journal of Mechanical Sciences, 2016; 110, p. 190).

15. Normally, teachers are responsible for adjusting materials to the immediate needs of their learners.[concurrence].(Journal of language and education, 2016, 2(4), p.15).

In sentences 14 and 15, the writers use pronouncement and concurrence resources ‘notably’ and‘normally’ to present her voice emphatically as opposed to the alternative voices acknowledged by the writer as related to the credit of his/her voice.

Entertainment assertions are realized through the linguistic resources presented in
Table 12 were observed all over the LR sections of the articles analyzed in the present study.

Table 12. Linguistic Representations for Entertainment Resources in Articles Published in International and Local Journals

As sentence 16 shows, authors use the resources presented in Table 12 to convey their own propositions within a variety of other possibilities.

16. Studies on CALL suggest that by integrating technology into language learning, learners are exposed to more authentic input as well as more opportunities to participate in a socio-cultural context.[entertainment]. (Journal of English Language Teaching and Learning, 2015; 16, p. 41)

The function of the verb ‘suggest’ in sentence16 is that the writer, after considering studies on CALL, intends to give his/her own proposition, not an external-sourced one. In addition, it shows his/her proposition is one of a variety of possibilities. Therefore, in example17, the writer uses a word expressing evidence-based postulation, i.e. ‘suggest’ as an entertainment resource to convey his/her voice.

Endorsement assertions, as one way of making reference to other voices, are used by the author to explicitly support the external-sourced voices. Therefore, in the articles analyzed in the current study, they were either realized by a supporting expression preceded or followed by a reporting verb, or by a verb which shows the author’s support. The resources presented in Table 13 were used for the purpose of endorsement. (The verbs were also used in other tenses and in passive voice).

Table 13. Linguistic Representations used for Marking Endorsement Resources in Published Articles

The verb ‘assert’ is a linguistic resource used for the purpose of acknowledgement, but here the author supports the external voice by using ‘Providing some interesting data’. Sentence17 is an example of endorsement assertion.

1. 17.  In a study, Alsbeih et al. showedthat the presence of oxygen as a radio sensitizer could significantly increase biological damage to ionizing radiation.[endorsement].

(Iranian Journal of Medicine Science, 2016; 41 (3), p. 180)

In sentence 17, the verb ‘showed’, as an endorsement resource, indicates that, the writer explicitly considers the external voice as valid and reliable. In other words, using the verb ‘showed’, the writer engages in citation by supporting and agreeing with the voice he makes reference to.

Attribution assertions were the most frequent resource used among all the heteroglossic resources in articles published in local journals and were realized through the following linguistic resources.

Table 14. Linguistic Representations for Realizing Attribution Resources in Articles Published in Local and International Journals

As sentence 18 shows, by using acknowledgment resources, authors adopt a neutral stance towards the external voice.

18. Erikson (1968) believed that having a fully formed of sense of self is essential to being able to form intimate relationships.[acknowledgement] (Iranian Psychological Association, 2014; 8(2), p. 80)

In sentence 18, the writer used the acknowledgement resource ‘Erikson (1968) believed’ to present an external voice, and at the same time adopted a neutral stance towards the voice s/he makes reference to.

As sentence 19 shows, by using distancing resources, authors took a doubtful stance towards the external voice.

19. Previous studies on the acquisition of tense-aspect have claimed that the development of tense-aspect morphology in L2 acquisition is strongly influenced by the inherent semantic aspect of the verbs to which the inflections are attached. [distance]

(Language Learning, 2015; 65(4), p.795)

In sentence 19, the situation is different from sentence 18. Here the writer used the distancing resource ‘claim’to assume a doubtful stance towards the external-sourced voice. By using the verb ‘claim’, the writer explicitly showed that the external voice makes a proposition that the citer feels may be invalid and so needs to stand back from.

The salient point regarding the use of attribution resources is how they are realized rather than how often. While the articles published in the journals in question use attribution resources almost equally in terms of frequency, the linguistic realization of attribution resources varies in articles published in hard science such as medicine, from those in soft science journals.  In the former, authors used only citation as the attribution resource while in the latter attribution is done either by using only citation or by citation and specific wording. To clarify the point, sentences 20 and 21 show the way attribution is linguistically realized in articles published in applied linguistics and medicine journals.

20. Both Vygotsky (1978) and Bakhtin (1986) believed that the development of mental functioning in the individual is the result of learning conversations with others.

(System, 2017; 64, p. 2)

21. Chronic infection is maintained by small numbers of parasite cysts localized in the CNS and contained by the residual T cell response (Suzuki et al., 1988).

(The Journal of Experimental Medicine, 2016; 213, p.3042)

In sentence 20, attribution is realized through the word ‘believe’, while in sentence 21, attribution is manifested through citation.

Discussion

The results of the current study revealed the distribution of engagement resources used in articles published in local and international journals across hard and soft science journals. The findings indicated that monoglossic resources were used more frequently in articles published in local journals than those published in international ones. In other words, Iranian writers show a tendency to give taken for granted facts that ostensibly allow no space for the negotiation of meaning between the writer and the readers in LR sections of their articles. Regarding heteroglossic resources, intra-vocalizing resources (countering, denial, pronouncement, and entertainment) were found more frequently in articles published in international journals. However, extra-vocalizing resources (endorsement and attribution resources) used in local journals outnumbered those used in international ones. Furthermore, monoglossic resources were used more frequently in articles published in hard science journals compared to those published in soft science journals. In terms of heteroglossic resources in the journals of the disciplines in question, disclaiming resources were observed more frequently in articles published in soft science journals while proclaiming resources (pronouncement and concurrence) were used more frequently in articles published in hard science journals. The frequency of entertainment and endorsement resources was higher in soft science journals, and attribution resources were distributed almost equally in both hard and soft science journals.

The finding of this study denoting the abundant use of monoglossic resources by Iranian writers is consistent with the findings of some studies (e.g. Abdi, 2002; Atai & Sadr, 2008) which found significant discrepancies between articles published in Iranian local journals and those published in international ones in terms of using hedges, modalities, and other discourse markers. Nevertheless, the tendency of Iranian authors to use monoglossic resources can be attributed to the point that teachers offering writing for academic purposes do not focus on the dialogic nature of genre in general and academic writing in particular. While writing, the authors of articles published in local journals might be preoccupied with textual features and pay less attention to the dialogic features of academic writing (Jalilifar, Hayati, & Mashahadi, 2013). The authors of articles published in international journals, on the other hand, are more willing to turn monoglossic assertions into heteroglossic ones through the use of clause modulation. Sentence 22 can elaborate on this issue.

22. LDM therapy may mediate its antitumor effect by inhibiting regulatory T cells. [entertainment]. (The Journal of Experimental Medicine, 2016; 213(13),
p. 2968)

In sentence 22, the writer used the entertainment resource ‘may’ as a voice modification technique to indicate that his voice is dialogic and among other possible voices. As a result, the frequency of monoglossic assertions in articles published in international journals is low. To shed some light on the issue, the authors of articles published in international journals employ engagement resources to modify the voice and make it dialogic, and as a result more tentative (Hyland & Jiang, 2016).

It is noteworthy that the discrepancy in the distribution of monoglossic resources across articles published in soft and hard science journals in their LR sections is less than that between articles published in local and international journals. This finding puts attest to the idea that first language (L1) writing conventions have more influence over the use of monoglossic or heteroglossic propositions than the nature of the discipline (Atai and Sadr, 2008). This claim is consistent with Li and Wharton’s (2012) idea who consider context, rather than discipline, as an influencing factor on the writers' choice of metadiscourse markers. In choosing metadiscourse markers when putting forward their arguments and propositions, writers of articles published in international journals tend to engage their readers in their arguments by using heteroglossic resources more frequently; therefore, their texts are more reader-focused. Writers of articles published in Iranian journals, in contrast, tend to write less reader-focused texts (Jalilifar, Hayati, & Mashahadi, 2013). In reader- focused texts, according to Hyland and Jiang (2016), writers interpersonally negotiate with their readers and take into account the readers’ reaction when making their claims; that is to say, “the writers acknowledge and connect to others, recognizing the presence of their readers, pulling them along with their argument, focusing their attention, acknowledging their uncertainties, including them as discourse participants, and guiding them to interpretations” (Hyland, 2005, 176). In fact, by using heteroglossic resources and dialogic expressions more frequently, international writers put more emphasis on reader awareness through opening up more space for alternative voices and consequently creating a writer-reader interaction while putting forward their propositions. They use mononoglossic resources only if they are quite certain and have a clear sense of their readers.

The frequent use of monoglossic resources by Iranian authors may be attributed to the writing conventions of Iranian writers and their tendency to put more emphasis on convincing and consensus rather than negotiation and argumentation (Jalilifar, Hayati, & Mashahadi, 2013).

The overall use of intra-vocalizing resources (disclaim, pronouncement, concurrence, entertainment) in articles published in local journals is less frequent than those of published in international ones (52.84% vs. 63.94%). Nevertheless, the use of extra-vocalizing resources (endorsement, acknowledgement, distancing) is more frequent than the ones found in articles published in international journals (47.16% vs. 36.06%). This suggests that writers of articles targeted for international readership tend to convey their own voice in LR section more than those of local ones. According to Liu’s (2013) findings, the writers of low-rated essays used all subcategories of extra-vocalizing resources (endorsement, acknowledgment, and distance) more than those of high-rated ones. The most likely explanation for the high frequency of extra-vocalizing resources in LR section of articles published in Iranian local journals stems from Iranian writing conventions. The tendency of Iranian writers is not to express oneself explicitly (Jalilifar, Hayati, & Mashahadi, 2013), and in many cases, they take mongolossic resources or propositions made by other scholars for granted. The international writers, by contrast, use intra-vocalizing resources more frequently to present their own voice along with the propositions made by other scholars. Furthermore, the type of intra-vocalizing resources they use depends largely on the extent of certainty they have towards their propositions. For example, if the degree of certainty is low, entertainment resources such as ‘may’ is used as in sentence 22 above. Unfamiliarity with linguistic representations of intra-vocalizing resources is another possible explanation for lower frequency of intra-vocalizing resources, and this, in turn, is due to the fact that in their writing courses, appraisal resources for making persuasive propositions and argumentations are not taught explicitly (Jalilifar, Hayati, & Mashahadi, 2013).

The possible reason for more frequent occurrence of monoglossic resources in journals of hard science disciplines in comparison with those of soft science ones is that hard science articles present more objective and precise propositions in their LR sections (Hyland, 2008). Monoglossic resources are one way of conveying objectivity or certainty of information, and are mostly used by local articles in such journals. Heteroglossic resources such as concurrence and pronouncement resources can also be used to present accurate and precise information. This may be why the frequency of concurrence and pronouncement resources in hard science journals is higher than those of soft science ones.

The lower frequency of pronouncement and concurrence resources used in soft science journals is consistent with the expectation that soft sciences, being qualitative, openly allow for alternative points of view in all circumstances while presenting a proposition (Becher, 1994); therefore, the author of soft science articles use fewer pronouncement and concurrence resources which do not leave any space for alternative voices.

The lower frequency of disclaiming resources in articles published in hard science journals maybe because hard science disciplines, according to Dahl (2004), contain a stable and homogeneous knowledge base and structure for research reporting. Therefore, resources like countering and denying which invoke prior utterances or alternative views in order to refute or replace those views, are rarely used in hard science journals.

Endorsement and distancing resources are to be found more commonly in LR sections of soft science articles than in hard science ones. However, as shown in Table 11, the frequency of acknowledgment resources is almost equal in hard and soft science journals. The authors of soft science articles tend to use an evaluative stance when making reference to others’ voices while the authors of hard science articles show a tendency to remain more neutral. That is simply because the latter makes reference to more concrete and objective information and tend to show conformity to authoritative voices rather than making critical evaluation. Nevertheless, the frequency of endorsement resources in articles published in medicine journals is not significantly lower than those of articles published in soft science journals. This is because hard science disciplines use a great deal of research-related verbs (e.g. explore, find, prove, etc.), especially when they make reference to other’s findings (Hyland, 2000). The infrequent use of distancing and endorsement resources in LR sections of local mechanical engineering journals is more a matter of L1 writing conventions than discipline. According to Hyland (2000), a relatively frequent occurrence of endorsement resources is expected in hard sciences; however, endorsement resources are very infrequent in LR sections of articles published in local mechanical engineering journals. According to Hyland (2000, p. 360), “when compared with those published in soft sciences such as applied linguistics and sociology, research articles published in hard science journals, such as engineering and physics, because of epistemological orientation of hard science disciplines which see laboratory activities as key sites where knowledge is generated, display a greater use of research-related verbs (e.g. explore, find, prove, etc.)”. However, the frequency of endorsement resources used in the LR sections of articles published in international mechanical engineering journals is relatively high. This indicates that the infrequent use of endorsement resources is a matter of L1 writing conventions rather than the writing conventions of a particular discipline. Endorsement resources are normally used by the writers who support their claim by showing alignment with a source voice. As can be expected, the appropriate use of various textual sources to synthesize ideas is a skill that presents many difficulties to novice writers (Petric, 2007; Wu, 2007).

Conclusion

The distribution of engagement resources as one domain of AT in LR section of articles was analyzed in the present study. It was mentioned that using engagement resources efficiently can foster the argumentation and persuasion of LR section of research.

The results of the study yielded several patterns of the way engagement resources appear in LR section of articles published in local and international journals in four different disciplines. Compared to the authors of articles published in international journals, authors of articles published in local journals employed a higher percentage of monoglossic resources in their LR sections suggesting that the writers of articles published in international journals allow more space for dialogic alternatives (Martin & White, 2005). Expansive resources were used more than contractive ones in both local and international articles showing the tendency of writers to allow more space and possibilities for dialogic alternatives. Articles published in international journals made more frequent use of contractive and expansive resources.

In addition, it was found out that the writers of articles published in local journals tend to use intra-vocalizing resources in their LR sections less than their international counterparts. According to Lui (2013), using infrequent amount of intra-vocalizing resources is a characteristic of novice writers’ texts.

Furthermore, regarding citation in LR section, it was found out that authors of articles published in soft science journals tend to take a more evaluative stance than the authors of articles published in hard science ones whereas authors of articles published in hard science journals tend to take a more neutral stance and show conformity to the authoritative voice.

The findings of the current study can open a new horizon to those writers, especially novice ones who are not able to bring about rhetorical effects in writing argumentative and persuasive articles. It can also give the insights to novice researchers that LR section is not to be considered merely as a space for presenting external voices. The findings of this study show them how they can use intra-vocalizing resources to convey their own voices along with external-voices to support their voice. It has also implications for teaching writing in ESP and EAP courses. As stated before, one of the major differences between international and local writers is the fact that international writers tend to write reader-focused texts while local writers do not pay considerable attention to the awareness of readers. This problem is not limited to Iranian EFL contexts and has been reported by many studies in different contexts, especially Asian contexts (Geng & Wharton, 2016; Wu, 2007; Xie, 2016, etc.). That is being so, the findings of this study call for a change which takes into account the awareness of readers in teaching academic writing in ESP and EAP courses not only in Iran but in other EFL contexts. Writing, as Hyland and Jiang (2016) state, needs to be considered as an interaction or negotiation between the readers and the writers.

One possible reason for not writing persuasive and augmentative academic texts, in general, and LR section of articles in particular, by local writers is the lack of familiarity with the linguistic representation of appraisal resources, especially engagement ones. This challenge has also been addressed by many studies in different EFL contexts (e.g. Tirkkonen-Condit, 1996). By providing a list of linguistic representations of different categories of engagement resources, this study helps EFL writers express their own voice by putting forward argumentative propositions.

One objection leveled against English writing classes is concerned with the lack of teaching critical evaluative skills (Qian & Krugly-Smolska, 2008; Xi, 2016). The findings of this study can help teachers of EAP (English for academic purposes) courses to show their students how to evaluate published authoritative voices in LR sections of their articles.

Despite the useful insights the current study can have, like other studies, this study suffers from some limitations as well. First of all, this study just investigated engagement as one domain of evaluation, and all the domains of AT have not been taken into consideration. To consider all other resources such as attitude and graduation, further studies are needed to be conducted. Moreover, this study examined engagement in just LR section of articles, which is sometimes merged with introduction section. Further studies can address evaluation or each of its resources in other sections of articles.