Reader Engagement in English and Persian Applied Linguistics Articles

There is an increasing interest in the way academic writers establish the presence of their readers over the past few years. Establishing the presence of readers or what Kroll (1984, P.181) calls imagining “a second voice” is accomplished when a writer refers explicitly to their readers using explicit linguistic resources (reader engagement markers). Although there are some cross-disciplinary studies and only one cross-cultural study (Hinkel, 2002) which has investigated how writers in different disciplines/cultures acknowledge the presence of their readers, no contrastive study has ever been reported to have examined how academic writers from Persian and English writing cultures address their readers in their texts.Drawing on 60 applied linguistics articles (20 English articles written by native English applied linguists, 20 English articles written by native Persian applied linguists and 20 Persian articles written by native Persian applied linguists), this study aimed at seeing how native Persian and English writers engage their readers in their articles. Hyland’s (2005a) interactional model of stance and engagement was used as an analytical framework to identify the type and frequency of reader engagement markers in these three groups of articles. The result of the analysis showed significant differences in the way native Persian and English represent their readers. Also, considerable differences were observed in categorical distribution of reader engagement markers.


Acknowledging the presence of readers or what
) calls imagining a "second voice" in a written text has begun to receive attention after a period during which it had been under-researched.To accomplish a successful writer-reader interaction, a writer should take account of readers' personal traits, their background knowledge, processing constraints, recognition of face needs of readers (Myers, 1989) and their social, cultural backgrounds that might affect communicative exchange which a written discourse is aimed for.In fact, as advocated by Thompson and Thetela (1995, p. 103), a major feature of interaction is "to project a reader-in-the-text with whom the reader is invited to identify or converge".Unless the readers' interests, needs, knowledge and anticipations are taken into consideration in constructing a text, interaction fails.In Duzsak's terms (1994, p.292-293), anticipations in discourse are believed to "stem from a number of sources and underlie various aspects of discourse processing, such as appropriacy of topics (content schemata) and of rhetorical form (formal schemata), the general knowledge and cultural suppositions (systems of shared norms, beliefs, values, and stereotypes)".
Thus the onus is on the writer to fully understand the identity of the reader because it gives a "wealth of tacit and explicit knowledge about the form of the discourse and the way the subject can be treated" (Park, 1986, p. 483).Acknowledging the importance of audience analysis, Kroll (1984, p.174) has suggested three perspectives on audience: rhetorical perspective which sees the act of writing as "persuasive in intent" while informational perspective views the act of writing as a "process of conveying information".But for social perspective, writing for readers is "like all human communication, a fundamentally social activity, entailing processes of inferring the thoughts and feelings of the other persons involved in an act of communication" (Kroll, 1984, p. 179).This means that the structure of a written genre is informationally, rhetorically, and stylistically constrained by the awareness of audience.Thus one can argue that each genre views its audience differently as how to present information to them.Hyland (2010, p. 118) differentiates between research articles and popular science arguing that research articles are written "for a professional audience with a high degree of specialised expertise.Information is presented with considerable exactness, foregrounding procedures and using technical jargon, nominalisations, precise measurements, cautious inferences from data, and acronyms" whereas popular science "is produced for audiences without a professional need for information about science but who want to keep abreast of developments".
Most of the studies have been devoted to how academic writers project themselves into their texts by commenting on the possible accuracy or credibility of a claim, conveying an attitude towards both propositions and readers both across disciplines (Vassileva, 1998;Tang and John, 1999;Hyland, 1999;2001a;2002a;Ivanic and Camps, 2001;Biber, 2006;Starfield and Ravelli, 2006) and across cultures (Vassileva, 2001;Dahl, 2004;Martinez, 2005;Mur Dueñas, 2007;Shelden, 2009).The ways writers relate to the reader and engage with them have been relatively neglected in the literature.
There are some studies (Tapper, 1994;Webber, 1994;Swales et. al, 1998;Hyland 2001b;2002b;2000c;2005a;2005b;Fortanet, 2004) which have addressed the presence of readers across disciplines in academic discourse.Not only have these studies contributed to the loss of traditional tag of academic writing "as an objective, faceless and impersonal form of discourse and come to be seen as a persuasive endeavor involving interaction between writers and readers" (Hyland, 2005a, p. 173) but also they have been influential in popularising the importance of viewing readers as active participants in constructing and disseminating knowledge.These studies have indicated that there are considerable variations in the employment of reader engagement markers across different disciplines.
Applied linguistics as one branch of social sciences has been chosen as the field of the study since it has been proved to exhibit higher use of interpersonal markers between the reader and the writer.Duszak (1997, p. 11) suggests, contrary to writing in the hard sciences, writing in the humanities and the social sciences is likely to be permeated by "language-and-culture-bound discoursal preferences".Applied linguistics as one example of soft sciences, as Swales et al. (1998, p. 103) speculate, "may require rather more specific forms of reader-text management".Also as Yakhontava (2006, p. 163) argues "it is possible to assume that in linguistics which, like other disciplines in the humanities, is more subjective and consequently more sensitive to national contexts, textual patterns may show greater variability, yielding the considerable divergences".
In spite of the fact that "there are universal characteristics within any discourse (genre) which imposes uniformity on members of discourse communities no matter what language they happen to use with some tolerance for individual stylistic variation" (Widdowson, 1980, p. 61), Mauranen (1993) argues that "writing cultures into which any academic writer was socialized when first learning to write, usually his or her national culture, may constrain her or her choice of rhetorical strategies" (p.5).The purpose of the present study is to determine whether genre-driven (discourse community) or culture-driven (national writing conventions) features condition the use of these strategies.That is why Atkinson (2004, p. 227) calls for a "better conceptualisation of culture in any contrastive study" or intercultural rhetoric(to use Conner 's term, 2004).Of course, it should be mentioned that the appropriate and competent use of these rhetorical choices help the writer project themselves into their texts, convey an attitude towards propositions, and keep their readers engaged throughout the argument.This reader engagement has to be exploited and reflected in one way or another in the way the text is written.One way by means of which this is successfully accomplished is "interactional metadiscourse" (Thompson, 2001, p. 59) in general and "reader engagement markers" (Hyland, 2005a, p. 154) in particular.
Despite the importance of representing readers in academic discourse, there is only one cross-cultural study (Hinkel, 2002) which has investigated the presence of readers in texts written by native and non-native English writers.She found considerable discrepancies in the use of reader engagement markers (REMs) among her study groups.Drawing on Hyland's (2005a) model of stance and engagement, this paper examines the type and frequency of REMs in three groups of applied linguisticsarticles (20 English articles written by native English writers, 20 English articles written by native Persian writers and 20 Persian articles written by native Persian writers).
Research question 1: 1. Is there any significant difference between the type and amount of reader engagement markers employed by native writers of English and native writers of Persian (both writing in English) in their research articles about applied linguistics?
Research question 2: 2. Is there any significant difference between the type and amount of reader engagement markers employed by native writers of English writing in English and native writers of Persian writing in Persian in their research articles about applied linguistics?Research question 3: 3. Is there any significant difference between the type and amount of reader engagement markers employed by native writers of Persian writing in English and native writers of Persian writing in Persian in their research articles about applied linguistics?

Corpora
The corpus was built from three groups of articles (20 English articles written by native English applied linguists, 20 English articles written by native Persian applied linguists and 20 Persian articles written by native Persian applied linguists).All English articles written by native English writers were published in leading international journals (i.e.Applied Linguistics, English for Specific Purposes, Journal of Second Language Writing, Journal English for Academic Purposes, TESOL Quarterly, and System) whereas English articles written by native Persian writers were published in both international journals as well as peer reviewed journals in Iran.Finally, it is necessary to note that all target Persian articles were published in Persian journals in Iran.The publication period was from 2001 and 2010.

Procedures and design
All articles were selected on the basis of the three criteria set by Nwogu (1997) namely, representivity, reputation, and accessibility.In terms of representivity, the articles published in these journals were fairly representative of the genre (research articles) in content and field (applied linguistics).But in regard to reputation, there was no doubt that the selected English journals were all popular journals.Accessibility was the ease with which the articles in the corpus could be obtained.
Initially, all texts were stored electronically for automatic search purpose.Then, the results of the automatic search were manually cross-checked by the researchers.Since the type and occurrence of reader engagement markers categories were extremely varied and multifunctional, a context-sensitive analysis of each marker had to be carried out before it was finally counted.To invoke Swales' ( 2004) differentiation of theory-based and data-based articles, only data-based articles were included in the corpus to ensure the corpus comparability.
All quotations, linguistic examples, footnotes, bibliographies, tables, and figures were excluded.Also, the titles of all articles and the abstract sections were not checked for reader engagement markers.That was done on the base of Gillaerts and Van de Velde's (2010) argument asserting that abstract sections are not appropriate sections to look for reader engagement markers because they are quite short.Hyland's (2005a) interactional model of stance and engagement was selected as an analytical framework to identify the type and frequency of REMs.It consists of two dimensions: one is stance which includes features which refer to the ways writers project themselves into their texts and convey their judgments, attitudes, and commitments.The other, which is the topic of our study, is engagement which refers to the ways writers relate to their readers and establish their presence in their texts.The type and frequency of REM were counted in context to make sure that they were all instances of REMs.Since the total number of all words in each three groups of articles was not identical, REMs were calculated per every 1000 words (as was the case in Mauranen, 1993;Hyland, 2005a;Dafouze-Milne, 2008).Then, independent Samples T-Tests were employed to find out whether there were statistically significant differences among three groups of articles in the type and frequency of REMs.

Overall distribution of REMs in all three groups of applied linguistics research articles
In order to see whether there were differences among three groups of articles (English articles written by native English applied linguists, English articles written by native Persian applied linguists, and Persian articles written by native Persian applied linguists) in the use of REMs, we calculated the frequency of REMs per every 1000 words.Table 1 shows the frequency of REMs per every 1000 words, their mean values and significance levels.As Table 1 shows, the frequency of REMs per every 1000 words in G1 is 10.825 followed by G2 and G3 using 7.110 and 5.350 per every 1000 words, respectively.This means that G1 used REMs approximately twice as many as G3.The result of Independent Samples T-Test confirmed that the difference between G1 and G3, on the one hand, and G1 and G2, on the other hand, was statistically significant (p-value < 0.05).But the difference between G2 and G3 was not significant (p-value > 0.05).Based on their mean values, the divergence in the use of REMs among three groups of applied linguistics articles is also indicated in Figure 1.
One explanation for this divergence in the use of REMs could be linked to the way different writing cultures view their readers.This study is in line with those of Kaplan (1966), Clyne (1987), Hinds (1987), Crisemore, et al. (1993), Mauranen (1993), Duszak (1994), Valero-Garcѐs (1996) Vassileva (2001), Dahl (2004) Shelden (2009).They indicated that discoursal patterns and practices were inextricably interconnected with and conditioned by disparate intellectual styles and cultural traditions.For example, Duszak (1994) argued that styles of intellectual debating, scientific prose, and attitudes to knowledge and academia were interconnected with cultural values, norms, and beliefs.In other words, the generic constraints on academic prose reflect the cultural habits of the writer's academic community.Another explanation lies in the fact that there is "no or little formal L1 composition instruction" (Marandi, 2003, p. 24) in Iran at high schools.This can make writing scholarly articles a frightening task not only in their native language but also in English.Thus composition instruction courses seem inevitable both in L1 and L2 especially in the use of metadiscourse in general and reader engagement markers in particular for native Persian writers writing both in English and Persian.However, she found some similarities in categorical distribution of metadiscoursal elements among native and non-native English thesis writers.This, in turn, invokes the importance of genre-based approaches to the teaching of writing in ESP classes.

Categorical distribution of REMs in all three groups of applied linguistics research articles
As regards the five categories of REMs, 'personal asides' stood out as the most frequently element with the 'directives' being the second noticeable feature in all three groups of articles.As demonstrated in Table 2, G1 used 4.472 with G2 and G3 using 3.058 and 0.957 'personal asides' per every 1000 words, respectively.The result of Independent Samples T-Test showed that the difference between G1 and G2 was not significant (p-value > 0.05).But the difference between G1 and G3, on the one hand, and G2 and G3, on the other, was statistically significant (p-value < 0.05).It appears that English articles (written by native English and Persian writers) were much more eager to address their readers directly "by briefly interrupting the argument to offer a comment on what has been said" (Hyland, 2005a, p. 183).Employing this strategy, native English and Persian writers while writing in English not only obviate any processing difficulties on the part of the reader by simply annotating but also make an interpersonal dialogue with the reader as the following example from the corpora show: (see Appendix for the subcategories of all five categories of English) 1.While at least some of the assumptions have thus been challenged, and others likely will as the field evolves, emergentism (at least one that is informed by the chaos complexity theory perspective on it that I hold) does so wholesale, rather than piecemeal.(Larsen-Freeman, 2006) 2. In addition, seen from a genre-analytic perspective, SFL not only provides a detailed description of the rhetorical functions and linguistic structures of English (cf.Halliday, 1985Halliday, a & b, 1994)), but goes further and relates the contextual dimensions of register/genre to the semantic and grammatical organization of language itself.(Ansary and Babaii, 2004) 3. Persian Example Translitrated (DaryekmotaleaeAliakbaribatahhiyeporseshnamienazaredastandarkaranamouzaeshzabandabirestanhayeiranrajooyag hardideast(Nemooneporseshnamedarzamimeamadeast).

Persian Example Translated
In a study, Aliakbari (1381) has investigated the perspectives of those involved in teaching of language at high schools in Iran through a questionnaire (one sample of the questionnaire has appeared in the appendix).(Pishghadam and Moafian, 2007) The incidence of 'reader pronouns' shows divergence among three groups of articles in the sense that the highest occurrence of this category belongs to G1 (i.e., 1.649) followed by G3 and G2 using 1.234 and 0.353 'reader pronouns' per every 1000 words, respectively.The result of Independent Samples T-Test showed that there was no significant difference between G1 and G3 (p-value > 0.05).This means that the two groups displayed the same pattern of use in this category.However, there was a significant difference between G1 and G2 and G2 and G3 (p-value < 0.05).Both G1 and G3 employed more instances of this category than G2 did.This suggests that two writing cultures though different in employment of some REMs follow identical patterns of use regarding explicitly addressing their readers.This explicit representation of the reader is reflected in the following examples from the corpora: 4. If we look at inter-judge agreement (Table 2) there is considerable variation.(Alderson, 2007) 5. We notice that all the three learner levels find the implicit argumentative texts more difficult than the other text types and versions.(Alavi and Abdollahzadeh, 2008) (Pasmitavanghoftkemanabaiaddonbalebehtarinrahkarhabashimbalkebaiadrahkarharabatavojjohbe tafavoyhayefardiafradvabaftezabanimoredbarrasigharardahim) Persian Example Translated (Thus one can argue that we should not look for the best procedures.But we should study the procedures in accordance with learners' individual differences and their linguistic contexts).(Pishgadam, 2008) The use of 'we', as one of the subcategories of this type of REMs, implies that both the writer and the reader are involved in the argument and the writer takes account of the reader's background knowledge and his or her ability to follow the argument.In fact, the use of inclusive 'we, as Kou (1999, p. 133) argues, "shortens the distance between readers and writers and stresses the solidarity between readers and writers".
The occurrence of 'appeals to shared knowledge' in G1 per every 1000 words was 1.408 with G2 and G3 having 1.230 and 1.291 per every 1000 words, respectively.No significant difference among three groups of articles was observed (p-value > 0.05).It appears that in the two writing conventions writers, as Myers (1989, p. 8) argues, "show their solidarity with the community more subtly by exhibiting responses that assume shared knowledge and desires".The following examples from the corpora clarify the point: 7. It can be suggested, of course, that the sentences in question might have the same effect upon native listeners.(Fox, 2004) 8.The notion of readability is, of course, a controversial issue since researchers have found that linguistic complexity may not be detrimental to comprehension.(Mehrpour and Riazi, 2004) 9. Persian Example Transliterated (Albattenazariate Oxford eghbalebishtariroo be roo shod ast (SILL) vaazporseshnameoobesiarestefadeshodeast).

Persian Example Translated
(Of course, Oxford's theories have been much welcome (SILL) and used a lot.(Borzabadi andNejati, 2008) Hyland (2005a, p. 184) believes that these engagement elements seek to "position readers within apparently naturalized boundaries of disciplinary understandings".Writers' reliance on pre-agreed upon knowledge obviates any processing difficulties and misunderstandings on the part of the reader because all writers' arguments are in line with the reader's line of thinking and reasoning.The findings of the present study regarding the use of the category went against Hinkel's (2002) study calming that non-native English writers drew on more appeals to shared knowledge compared with their native English counterparts.This study revealed that the extent to which all three groups of writers appealed to common knowledge shared by the writer and the reader did not significantly differ.'Questions' are the "strategy of dialogic involvement par excellence, inviting engagement and bringing the interlocutor into an arena where they can be led to the writer's viewpoint" (Hyland, 2002c, p. 185).This is among different rhetorical strategies by means of which writers can acknowledge the presence of their readers by maneuvering readers into accepting the writer's viewpoint or follow a particular line of argument.The importance of 'questions' is also acknowledged by Swales (1994) where he argues that by using questions writers establish a niche in research article introductions as these examples from the corpora illustrate: 10.To put it another way: If top-down evidence conflicts with bottom-up, which of the two is the second language learner most likely to trust?(Field, 2004) 11.Given that their ages, motivations and so on are the same, why are there such differences?(Fahim and Pishghadam, 2007) Persian Example Translitrated (Barayemesaltaklifedarsiestandaradvaporseshnamemontabaghba an chist?)Persian Example Translated (For example, what are standard assignment and the questionnaire which is matched with it?)(Borzabadi and Nejati, 2008) The results showed that the highest frequency of occurrences of 'questions' was for G3 which used 0.623 followed by G1 and G2 using 0.529 and 0.396 per every 1000 words, respectively.But no significant difference among three groups of articles (p-value > 0.05) was found.Like the category of 'appeals to shared knowledge', the use and frequency of this category ran against Hinkel's (2002) study.She found that 'questions' were more frequent in texts written by non-native English writers.Also, the result of this study challenged Swales and Feak's (1994) and Chang and Swales' (1999) claims that the use of 'questions' in academic writing is inappropriate.The use of 'questions' seems to make it possible for these writers (at least in this study) to invite and bring the reader into the discourse, arouse their curiosity and attention, and keep them engaged to follow the argument in the way preferred by the writer as it is evident in Webber's (1994, p. 266) statement where he points out that "questions create anticipation, arouse interest, challenge the reader into thinking about the topic of the text, and have a direct appeal in bringing the second person into a kind of dialogue with the writer, which other rhetorical devices do not have to the same extent".It is thus possible to suppose that the use of the 'questions' shows discipline than cultural specificity.That is to say, the use of 'questions' seems to be a common rhetorical strategy by means of which writers engage their readers no matter what language background they belong to.
'Directives' are defined as utterances which instruct the reader to perform an action or to see things in a way determined by the writer.Although 'directives' are among those rhetorical strategies that prerequisite for a successful write-reader interaction, these devices can act as "bald-on records" (Brown and Levinson, 1987, p. 67) threat to face due to the fact that any academic writing presupposes appropriate audience relationships.Thus the use of 'directives' can imply some imposition to follow the argument in the way preferred by the writer as the following examples show: 12.Some of these publications also include suggestions for writing RPT reports, such as providing specific constructive criticism of the faculty member reviewed (seeAmerican Council on Education et al., 2000 andHiggerson, 1999for illustrations of preferredand non-preferred evaluative comments).(Hyon, 2008) 13.It should be noted that since the design of the study was an Ex-post-facto design andthe aim was NOT to investigate the effectiveness of strategy instruction in FD/FI subjects, the participants didn't receive any explicit instruction in using metacognitive and cognitive strategies in the course of this study.(Ghonsoloi and Eghtesadi, 2006) 14.Persian Example Transliterated (Barayemoghaiesemotaghaiierhayeghunaghunbesiarmohammastke be zarayebestandaradshodetavajohkonimvana be zarayebgheire standard shode) Persian Example Translated (It is very important to pay attention to standardized quotients not to non-standardized quotients).(Jamshidian and Khomeijani, (2009) In this category G1 used 2.769 'directives' followed by G2 and G3 using 2.074 and 1.245 per every 1000 words, respectively.The result of Independent Samples T-Test yielded no significant difference between G1 and G2 (p-value > 0.05).But the difference between G1 and G3, on the one hand, and G2 and G3, on the other, was significant.In other words, G1 and G2 made approximately identical uses of 'directives' in their articles.This ran against Tapper's (1994) finding that half of the occurrences of her subject's use of 'directives' in English appeared to be influenced by her subjects' culturally-bound understandings of the nature of 'directives'.Figure 2 indicates the divergence in the use of these categories based on their mean values.As indicated, 'personal asides' are the most frequent REMs (with the highest mean) with 'directives', 'appeals to shared knowledge', 'reader pronouns' and 'questions ranking second third and fourth, and fifth, respectively.

Conclusion
Despite a relative uniformity of academic papers imposed by the requirements of the genre, the result of the present study revealed significant differences in the use of REMs.That is to say, native English writers used REMs twice as many as Persian writers writing in Persian per every 1000 words.Thus based on the result of the present study, one can argue that in English writing culture writers supposedly imagine a "second voice" (Kroll, 1984, p. 181) or a "reader-in-the-text" (Thompson and Thetela's, 1995, p. 103) -that is immanent in the text and with whom the writers makes a dialogic conversation throughout the argument.With this in mind, English writers build on linguistic resources to make their readers' presence noticed, take into account their processing capabilities, and adjust their argument in case any possible objections arise.It gives the impression that English writers view their readers as real and active players rather than passive observers of the argument.In contrast, native Persian writers tend to have adopted the role of "a solitary writer focusing on the propositional content of the text more than on his/her reader and on the formal structure of the paper" (Valero-Garcѐs, 1996, p. 291).All this suggests that English writing culture tends to be reader-oriented culture compared with Persian writing culture.
Based on the findings related to categorical distribution of REMs in three groups of applied linguistics articles we could conclude that the variation in the use of these REMs might have roots in cultural particularities.In simple words, the categorical use of REMs was influenced by the writers' national writing culture conventions.However, attributing all the discrepancies to cultural differences has been warned against by Spack (1997).Cultural differences, he argues, cannot solely account for the disparities observed in a contrastive study.But we have used it as an umbrella term to "embrace rather diverse interpretations stemming from a wide range of possible influences, varying, for example, from sociopolitical factors to purely intralinguistic peculiarities, which are assumed to interplay and leave their joint imprints upon the texts created within different linguistic and cultural environments" Yakhontava (2006, p. 154).
As a limitation of the study, it is to be noted that the role of gender was considered to be neutral.As Tse and Hyland (2008) argue, the issue of gender can be influential in the construction of successful writer-reader interaction.Thus it warrants a separate study to investigate the effect of gender on the "process of drawing-in readers" (Grabe and Kaplan's term, 2000, p. 197) in written discourse.Another limitation was examining applied linguistics articles as a representative of soft disciplines in which higher uses of interpersonal markers are observed.Future research is needed to investigate the frequency and distribution of REMs in various fields across soft/hard-discipline continuum.
The general finding showed that in comparison with both groups of native Persian writers, native English writers used far more REMs to represent readers in their texts.It suggests that native English writers have accomplished writer-reader interaction more successfully than their native Persian counterparts.Finally, it is to be noted that the findings of the present study could be pedagogically valuable for the large groups of ESL/EFL writers who entertain the idea of publishing scholarly articles in English.(1) N (2) P Figu

Table 1 .
The Independent Samples T-Test to compare distribution of REMs in three groups of articles