Maquetación 1 Ibérica 36 (2018): 171-194 ISSN: 1139-7241 / e-ISSN: 2340-2784 Abstract Citations, i.e. references to previous research, serve scholars to construct their own authority and support their new knowledge claims. However, scholars from different linguacultural backgrounds may use citations differently to achieve these purposes. This study compares the use of reporting verbs in articles written in English by scholars from Anglophone contexts and EAL scholars, in order to identify divergent usage and examine to what extent diverse discursive patterns containing these verbs are used and accepted in international publications. Results show that, although the overall frequency of reporting verbs is similar in both corpora, there were differences in the types of verbs (i.e. research verbs, discourse verbs and mental verbs) most frequently used, in the variety of verbs, and in the frequency of some specific verbs. Further, the usage of reporting verbs in the texts written by the EAL scholars appears to be determined by their knowledge of and conformance to generic conventions but also by their bilingual literacy. Keywords: reporting verbs, citation, variation, academic writing, EAL. Resumen V ar ia c i ón e n p rá c ti c a s d e e s c ri tu ra a c a dé m ic a : U n an ál is i s d e l os v e rb os d e a tr ib uc i ón en pu bl i c ac io ne s e n in g l é s Las referencias a estudios previos permiten a los autores de textos académicos construir su propia autoridad y apoyar sus conclusiones. Sin embargo, estas referencias pueden ser usadas de forma distinta por académicos de distintas culturas y lenguas para lograr estos objetivos. Este artículo compara el uso de verbos de atribución en artículos de investigación de lingüística escritos en inglés por autores anglófonos y españoles, para identificar usos diferentes y examinar Variation in academic writing practices: The case of reporting verbs in English- medium research articles María-José Luzón University of Zaragoza (Spain) mjluzon@unizar.es 171 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 171 Ibérica 36 (2018): 171-194 MARíA-JoSé LUzón hasta qué punto se aceptan prácticas discursivas variadas en las publicaciones internacionales. El análisis muestra que, aunque la frecuencia de verbos de atribución es similar en ambos corpus, hay diferencias en el tipo de verbos (verbos de investigación, de discurso o mentales) que se usan más frecuentemente, en la variedad de verbos, y en la frecuencia de algunos verbos. Los resultados sugieren que el uso de verbos de atribución en el discurso de los lingüistas españoles está determinado por su conocimiento de las convenciones genéricas, pero también por su literacidad bilingüe. Palabras clave: verbos de atribución, citas, variación, escritura académica Inglés como Lengua Adicional. 1. Introduction English has become the dominant language of communication in international academic contexts and scientific publication, with scholars using EAL (English as an Additional Language) for publication outnumbering first-language English scholars (Lillis & Curry, 2010; Pérez- Llantada, 2012). The high linguistic and cultural diversity of scholars publishing in English has aroused interest in the publishing and writing practices of EAL scholars (Lillis & Curry, 2010; Hanauer & Englander, 2013; Kuteeva & Mauranen, 2014; Cargill & Burgess, 2017). These studies have shown that the rhetorical choices of EAL scholars often arise from their discursive tradition and concluded that the imposition of Anglophone rhetorical conventions pose considerable challenges and difficulties for these scholars when attempting to publish in English (e.g., they may have to revise and resubmit their manuscripts because they diverge from the accepted norms of research reporting; research writing is particularly time consuming for them) (see Curry & Lillis, 2004; Belcher, 2007; Uzuner, 2008; Lillis & Curry, 2010). These concerns about the disadvantages of EAL scholars have led to calls for the acceptance of variation and multicultural rhetorical strategies in academic written texts (Belcher, 2007; Mauranen et al., 2010; Canagarajah, 2013; Lillis & Tuck, 2016). The consideration of practices that diverge from those deployed by Anglophone scholars from the perspective of “language variation”, rather than “language deficit”, is shared by three related approaches to academic writing: academic (bi)literacies, translingualism, and English as a Lingua Franca (ELF). The academic literacies approach (e.g., Lillis & Curry, 2010; Lillis & Tuck, 2016) sees academic writing as situated social practice, where 172 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 172 aspects such as identity, institutional context, power relations or ideology play an important role, and argues for the acceptance of diverse semiotic/ linguistic practices for academic knowledge-making. The translingual approach (e.g., Horner, 2011; Canagarajah, 2013) views the integration of (bilingual) language practices not as a problem to solve, but as a resource to produce meaning and achieve communicative effectiveness. ELF researchers also claim that, since most readers and writers of research published in English are non-native, there is no justification for imposing the “linguistic standards of a national community” (Mauranen et al., 2010: 184). Jenkins’ (2015) most recent conceptualization of ELF incorporates perspectives of the translingual approach: in ELF interactions the other languages that the participants know are also present, even if they are not used, giving way to what Jenkins calls “language leakage” (2015: 75). This echoes a basic concept in translingualism: “communication transcends individual languages”, that is, languages “mutually influence each other” (Canagarajah, 2013: 6). Rather than being detrimental, this influence offers possibilities to construct meaning and convey one’s voice. These three approaches emphasize that bi/multilinguals draw on their communicative repertoires for meaning construction and stress the need to accept these multiple discursive practices. Recent research on academic writing actually reflects a trend towards an increasing acceptance of non-standard linguistic forms in some high- prestige international journals (Rozycki & Johnson, 2013), which intimates that in order to get a paper accepted for publication grammatical accuracy may be less relevant than other factors like genre knowledge or awareness of the rhetorical self (Hyland, 2016; Tribble, 2017). Research has also revealed the acceptability of non-conventional hybrid rhetorical structures in research paper abstracts (Lorés-Sanz, 2016) and the use of recurring academic phraseology in L2 published papers that deviates from the use of academic phraseology by writers from L1 English backgrounds (Pérez-Llantada, 2014), which suggests that EAL scholars may be reshaping rhetorical patterns in innovative ways (Lorés-Sanz, 2016) and that Anglophone norms are merging with “culture-specific linguistic features” (Pérez-Llantada, 2014: 192). However, research on variation in academic texts and on the contribution of EAL authors to the evolution of English as used for international publishing is still scarce and there is, therefore, a need for more studies on the discursive practices of EAL scholars seeking publication. An integral part of academic discourse that can shed light on EAL scholars’ writing practices is citation, i.e., the inclusion in the text of a reference to VARIATIon In ACAdEMIC wRITIng PRACTICES Ibérica 36 (2018): 171-194 173 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 173 another source. Citation is generally acknowledged as a key feature for knowledge construction (and persuasion) in academic writing (Hyland, 1999; Charles, 2006). References to previous research enable writers to construct their own authority by showing that they are knowledgeable members of the community, to position their research within an existing state of disciplinary knowledge and create a niche for this research, and to show their allegiance to specific orientation within a disciplinary community (Hyland, 1999; Charles, 2006; Harwood, 2009). Much research into citation has been devoted to the analysis of the form and/or rhetorical function of citations in texts by a specific group of writers (e.g. novice writers, L2 students) (e.g. Thompson & Tribble, 2001; Charles, 2006; Petrić, 2007; Harwood, 2009; Mansourizadeh & Ahmad, 2011; Petrić & Harwood, 2013; Samraj, 2013). Some studies have focused particularly on reporting verbs, analyzing their tense and voice (e.g. Shaw, 1992), their denotation, i.e. the types of activities they refer to (Thompson & Ye, 1991; Hyland, 2002), or their role as evaluative devices, used to express the cited author’s stance towards the reported information and the writer’s commitment (Thompson & Ye, 1991; Hyland, 1999; Bloch, 2010). Yet despite the extensive body of research on citation, most studies have concentrated on texts by Anglophone writers, both expert and novice (Hyland, 1999; Charles, 2006) or by non-native student writers (Petrić, 2007). Although less attention has been paid to the citation practices of non- Anglophone scholars, research has revealed the influence of the scholars’ discursive tradition and social context on these practices. Some studies have analyzed the influence of the geographical location of scholars and the medium of publication (national vs. international) on the types of sources cited (Lillis & Curry, 2010; Lillis et al., 2010) and have revealed the pressure experienced by non-Anglophone writers to cite English-medium texts. other studies report cross-linguistic differences in citation practices (e.g., Mur-dueñas 2009; Hu & wang, 2014). Mur-dueñas (2009) reports differences in citation frequency and citation types between business management RAs written in English and in Spanish, and Hu and wang (2014) found differences in the type of citations used in Chinese- and English-medium journals of applied linguistics and medicine. Focusing on English-medium publications, research has also shown that the L1 background of non-Anglophone scholars influences their citation practices when writing in English (e.g., okamura, 2008; Rowley-Jolivet & Carter, 2014; dontcheva-navratilova, 2016; Hryniuk, 2016). okamura (2008) examined MARíA-JoSé LUzón Ibérica 36 (2018): 171-194174 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 174 citation forms in scientific articles written in English by Anglophone and by Japanese writers and found that the latter tended to use a single form of integral citation, while in the writing of the Anglophone writers there was more variety of forms. Rowley-Jolivet and Carter (2014) compared uncorrected manuscripts written in English by expert French researchers and published RAs by Anglophone researchers and found a much higher number of “reporting-that” clauses and a larger range of verbs in “reporting-that” structures in Anglophone writing than in drafts written in English by French scholars. These authors suggest that the reason for these differences could be that other types of reporting structures are used in French academic discourse and this influences the French researchers’ writing in academic English. Likewise, dontcheva-navratilova (2016) states that differences in intended readership and literacy traditions may account for the differences in the function of citations in research articles published in English by Anglophone scholars (targeted at an international readership) and by Czech linguists (targeted at a more local readership). To contribute to a better understanding of EAL scholars’ citation practices, the present article focuses on the use of reporting verbs when citing others’ work. despite the role of reporting verbs as evaluative devices to construct authorial identity (Thompson & Ye, 1991; Hyland, 2002), little research has addressed how EAL scholars use these verbs, although this scarce research suggests that the authors’ linguistic backgrounds may affect their use (Rowley-Jolivet & Carter, 2014). The aim of this article is to analyze and compare the use of reporting verbs in linguistics English-medium research articles written by Anglophone scholars and EAL scholars, in order to ascertain whether variations occur and to examine to what extent international publishing is accepting hybrid discursive practices regarding these verbs. More specifically, I will analyze whether the texts written by the two groups of writers display different citation practices as regards the frequency of use of reporting verbs, the semantics of these verbs, and the citation patterns where these verbs occur. 2. Corpus and Method 2.1. Corpus For the study, a corpus of 60 RAs in linguistics published between 2000 and 2009 was analyzed. 30 RAs were written in English by scholars from VARIATIon In ACAdEMIC wRITIng PRACTICES Ibérica 36 (2018): 171-194 175 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 175 Anglophone contexts (EngAL sub-corpus) (totaling 241,533 words) and 30 RAs were written in English by EAL scholars (Spanish scholars)1 (SPEngAL sub-corpus) (totaling 237,982 words). The corpus used is part of a larger corpus: the Spanish–English Research Article Corpus (SERAC). The SERAC corpus contains 24 articles for the EngAL sub-corpus and 24 articles for the SPEngAL corpus. I added six articles to each of the sub- corpora to have more data for the analysis of variation2. The criteria used to select RAs was the affiliation of the authors to Anglophone institutions (for the EngAL corpus) or to Spanish institutions (for the SPEngAL corpus). The authors’ first and last names provided additional confirmation of their status as Anglophone or EAL authors. For each of the sub-corpora 10 articles were taken from each of three international high-impact journals3 (Journal of Pragmatics, English for Specific Purposes and Lingua). Using the same journals for the compilation of the two sub-corpora guarantees likely audience, similar publication impact, and similar publication guidelines and editorial gatekeeping. 2.2. Method The first step of the analysis consisted in the manual identification of the reporting verbs (henceforth RVs) in the corpus. Through careful reading of the two sub-corpora, a list of all the verb lemmas that functioned as RVs in each of them was compiled. Then, the number of occurrences of each RV was calculated. For this purpose, WordSmith tools (Scott, 2008) was used to generate concordances for the verbs. Every concordance line was then examined manually, to count only occurrences of verbs that functioned as RVs in context. I use the term “reporting verb” to refer to those occurrences of verbs that in a particular rhetorical context are used to report on claims or ideas by other authors. Therefore, suggest is considered a reporting verb in (1a), but not in (1b). (1) a. welty (1989) suggests that (…) it is not always the most suitable one for students who are not familiar with the discussion method (SPEngAL) b. The analysis of collocational frameworks in the medical paper suggests that some frameworks are central to the phraseology of this genre. (SPEngAL) Frequencies of RVs in both corpora were compared and the academic section of the CoCA (Corpus of Contemporary American English) was MARíA-JoSé LUzón Ibérica 36 (2018): 171-194176 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 176 used to check whether RVs present in the SPEngAL but absent in the EngAL occurred in a larger English corpus. I then analyzed whether Spanish and Anglophone scholars used the different types of RVs with the same frequency. For the categorization of RVs, Hyland’s (2002) model was used. drawing on Thompson and Ye (1991), Hyland (2002) distinguished between “research acts”, which contain verbs which describe experimental activities or findings (e.g., calculate, discover, find), “cognition acts”, which contain verbs which represent the researcher’s mental processes (e.g., believe, assume), and “discourse acts”, which involve verbal expressions (e.g., discuss, describe). At this stage, the 30 most frequent RVs in each sub-corpus were selected, classified according to the acts to which they refer (research verbs, discourse verbs and mental verbs) and the frequency of occurrence of each type of verbs in the two sub-corpora was compared. The next step involved gathering all the occurrences of each RV and coding them according to the following criteria: type of citation and citation pattern. This coding helps to reveal differences in the cited author’s prominence and visibility in the two sub-corpora. The coding form to register the information for each RV can be found in Appendix A. (i) The type of citation where the verb occurs. I drew on Swales’ (1986) distinction between integral citation (i.e., the name of the author4 is included in the sentence, e.g., “Swales (1991) claims that…”), which foregrounds the cited author, and non-integral citation (i.e., the name of the author is provided in brackets), which places emphasis on the content. However, following Fløttum et al. (2006) and Hewings et al. (2010), I distinguished integral citation, non-integral citation and direct quotation. As Hewings et al. (2010) point out, although the occurrences of direct quotation could be classified as integral or non-integral citation, keeping “direct quotation” as a separate category can provide useful information on differences regarding author prominence in both corpora. (ii) The pattern where the verb occurs. For this analysis I drew partly on Thompson and Tribble’s (2001: 95-96) classification of functions of citation. They divided integral and non-integral citation into sub-types, taking into account both the function and the form of the citation. I adapted Thompson and Tribble’s (2001) classification, to focus mainly on form, that is, on the grammatical VARIATIon In ACAdEMIC wRITIng PRACTICES Ibérica 36 (2018): 171-194 177 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 177 choices made by writers when using RVs. In this case the occurrences of direct quotation have been subsumed in the integral/non-integral categories because there were only 13 occurrences of non-integral quotation and considering them separately would make the classification unnecessarily complex. I distinguished the following categories: 1. Patterns in integral citation (including integral quotation). 1.1. Human subject (cited author) + RV. A noun group referring to the cited author functions as the subject of the RV or as the agent in a passive construction. (2) a. Biber et al. (1999) show that these items convey a great deal of meaning (SPEngAL) b. … communicative tasks, defined by Skehan (1998: 268) as activities … (EngAL) 1.2. non-human subject (research noun+ cited author) + RV. The subject of the RV is a noun group whose head is a research noun (e.g., findings) and where the cited author is typically a modifier. (3) More recent work by Aijmer (2002) emphasizes that…. (EngAL) 2. Patterns in non-integral citation (including non-integral quotation). 2.1. general inanimate reference (discourse/research noun with a general reference) + RV (active /passive). The discourse noun is used to refer to several sources and the citation in brackets provides examples of these sources. (4) Some notable studies (Aguado de Cea, 1993; Belda Medina, 2003; gallench Pérez & Posteguillo gómez, 2001) have provided different approaches…(SPEngAL) 2.2. general animate reference (plural reference to cited authors) + RV. A plural animated noun (writers, researchers) is the head of a noun phrase that functions as the subject of the RV. (5) other authors mention the possibility of zero plural (…) (Lorenzo Criado, 1994, p. 200; Seco, 1972, p. 139). (SPEngAL) MARíA-JoSé LUzón Ibérica 36 (2018): 171-194178 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 178 2.3. RV in the passive voice or introductory ‘it’ as subject of RV in the passive voice (e.g., it has been reported that). (6) Research into teacher beliefs has been critiqued for eliciting (…) (Munby 1984; woods 1996) (EngAL) These patterns reflect different degrees of authors’ visibility as participants in the reporting process. while assigning the cited author the subject position (pattern 1.1.) makes him/her more prominent and visible, pattern 2.3. deflects attention from the author. 3. Results 3.1. Frequency of reporting verbs The number of occurrences of RVs in the two sub-corpora was very similar: 648 (272.28 per 100,000 words) in the SPEngAL corpus and 631 (261.24 per 100,000 words) in the EngAL corpus. This similarity indicates that Spanish expert writers in linguistics are fully aware of the function of RVs in their discipline. despite this similarity in the frequency of RVs, there is a striking difference regarding lexical variety: a much higher number of different verbs are used in the SPEngAL corpus (129 vs. 88). A comparison of the verbs that are used by Spanish and Anglophone scholars may help to explain this difference (see Table 2). As can be seen, 65 verbs occur in both corpora, a high number of verbs (64) occur only in the SPEngAL corpus, and a much smaller number of verbs occur only in the EngAL corpus. VARIATIon In ACAdEMIC wRITIng PRACTICES Ibérica 36 (2018): 171-194 179 VARIATION IN ACADEMIC WRITING PRACTICES Ibérica 36 (2018): …-… b. … communicative tasks, defined by Skehan (1998: 268) as activities … (ENGAL) 1.2. Non-human subject (research noun+ cited author) + RV. The subject of the RV is a noun group whose head is a research noun (e.g., findings) and where the cited author is typically a modifier. (3) More recent work by Aijmer (2002) emphasizes that…. (ENGAL) 2. Patterns in non-integral citation (including non-integral quotation). 2.1. General inanimate reference (discourse/research noun with a general reference) + RV (active /passive). The discourse noun is used to refer to several sources and the citation in brackets provides examples of these sources. (4) Some notable studies (Aguado de Cea, 1993; Belda Medina, 2003; Gallench Pérez & Posteguillo Gómez, 2001) have provided different approaches…(SPENGAL) 2.2. General animate reference (plural reference to cited authors) + RV. A plural animated noun (writers, researchers) is the head of a noun phrase that function as the subject of the RV. (5) Other authors mention the possibility of zero plural (…) (Lorenzo Criado, 1994, p. 200; Seco, 1972, p. 139). (SPENGAL) 2.3. RV in the passive voice or introductory ‘it’ as subject of RV in passive voice (e.g., it has been reported that). (6) Research into teacher beliefs has been critiqued for eliciting (…) (Munby 1984; Woods 1996) (ENGAL) These patterns reflect different degrees of authors’ visibility as participants in the reporting process. While assigning the cited author the subject position (pattern 1.1.) makes him/her more prominent and visible, pattern 2.3. deflects attention from the author. 3. Results 3.1. Frequency of reporting verbs Occurrences Per 100,000 words Number of different verbs used SPENGAL 648 272.28 129 ENGAL 631 261.24 88 Table 1. Frequency of RVs in the SPENGAL and ENGAL corpora. 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 179 Some of the verbs that occur in both corpora are part of phrases where the reporting expression is actually the whole phrase (e.g., draw a conclusion, provide evidence). Interestingly, although these verbs are present in both corpora, they do not always occur in the same combinations. For instance, the combinations draw+ proposal or provide+ approach occur in the SPEngAL corpus, but not in the EngAL corpus. Since the absence of a specific combination in any of the corpora may be due to the size of these corpora, I searched for draw+ proposal and provide+ approach in the academic section of the CoCA, to check whether they occurred in a larger English corpus. The searches yielded no occurrence of draw+ proposal, and a very low frequency of provide+ approach (1.28 per million). The SPEngAL corpus also displayed “multi-word verb+ noun” combinations which occurred only once (build up a typology, play down the role of, flesh out a framework) (e.g. 7) and were absent both in the EngAL corpus and in the academic section of the CoCA corpus. Although these combinations might be considered “incorrect” with regard to EnL (English as a native Language) standards, they are intelligible and have been accepted as such by participants in the reviewing and editorial process. The occurrence of these unconventional combinations is in agreement with Pérez-Llantada’s (2014) finding that formulaicity in L2 expert writing is not completely native-like. MARíA-JoSé LUzón Ibérica 36 (2018): 171-194180 MARÍA-JOSÉ LUZÓN Ibérica 36 (2018): …-… The number of occurrences of RVs in the two sub-corpora was very similar: 648 (272.28 per 100,000 words) in the SPENGAL corpus and 631 (261.24 per 100,000 words) in the ENGAL corpus. This similarity indicates that Spanish expert writers in Linguistics are fully aware of the function of RVs in their discipline. Despite this similarity in the frequency of RVs, there is a striking difference regarding lexical variety: a much higher number of different verbs are used in the SPENGAL corpus (129 vs. 88). A comparison of the verbs that are used by Spanish and Anglophone scholars may help to explain this difference (see Table 2). As can be seen, 65 verbs occur in both corpora, a high number of verbs (64) occur only in the SPENGAL corpus, and a much smaller number of verbs occur only in the ENGAL corpus. 1. RVs occurring in both corpora (65) 2. RVs occurring only in the SPENGAL corpus (64) 3. RVs occurring only in the ENGAL corpus (23) accept, address, advocate, analyze, apply, argue, assume, call, characterize, cite, claim, classify, comment, compare, conclude, consider, contend, criticize, deal with, define, demonstrate, describe, develop, discuss, distinguish, document, draw, emphasize, establish, examine, explain, find, focus, give, highlight, hypothesize, identify, include, indicate, introduce, investigate, maintain, make, note, notice, offer, observe, perform, point out, present, propose, prove, provide, reveal, put it, refer, report, say, show, state, stress, study, suggest, use, write acknowledge, add, adhere, agree, allude, approach, argue against, attest, attribute, be aware thati, be concerned, believe, build up, broaden, capture, carry out, center, challenge, claim for, contribute, declare, defend, estimate, differentiate, divide, do (research)ii, echo (suggestions), endorse, expound, flesh out (a framework), follow, hold (a view), illustrate, incorporate, insist, lay (emphasis on), link, list, mention, obtain, pay attention to, play down, point to, postulate, propound, put forward, qualify, reach (a conclusion), recognize, recommend, refine, reject, remark, reproduce, resort to, see as, signal, suffer, sustain, talk about, think, underline, understand, warn account for, assert, assign, confirm, constrain, dismiss, draw on, employ, formulate, interpret, instruct, outline, posit, produce (an analysis), question, raise (a point), review, specify, support, take (a position), term, test, view i Within RVs I also consider collocations of the type BE+ adjective. ii The object of the verb is provided to show how these verbs are used as a RV in the corpus. Table 2. Comparison of RVs used in the SPENGAL and the ENGAL corpus. Some of the verbs that occur in both corpora are part of phrases where the reporting expression is actually the whole phrase (e.g., draw a conclusion, provide evidence). Interestingly, although these verbs are present in both corpora, they do not always occur in the same combinations. For instance, the combinations draw + proposal or provide+ approach occur in the SPENGAL corpus, but not in the ENGAL corpus. Since the absence of a specific combination in any of the corpora may be due to the size of these corpora, I searched for draw+ proposal and provide+ approach in the academic section of the COCA, to check whether they occurred in a larger English corpus. The 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 180 (7) Although early research played down the role of the LI in second language learning (dulay, Burt, & Krashen, 1982)... (SPEngAL) Regarding columns 2 and 3 in Table 2, what is interesting is not the presence of some verbs in only one of the corpora, since this may be due to the size of the corpora, but the high number of verbs that only occur in the SPEngAL corpus. These include verbs of Latin origin, which have a frequently used Spanish cognate (e.g., adhere, allude, attribute, center, contribute, declare, defend, expound, mention, propound, recognize, sustain). Their high frequency in the SPEngAL corpus could be the result of “language leakage” (Jenkins, 2015) from the EAL scholars’ L1. L1 influence could also explain the occurrence in the SPEngAL corpus of specific verbs that do not occur in the EngAL corpus, such as signal (see example 8), and that could be considered “false cognates”. (8) Janney and Arndt (1992: 22) signal “the lack of agreement among investigators about how politeness should be defined” (SPEngAL) The two occurrences of signal could be due to the fact that the Spanish verb señalar is the equivalent for note/ point out. The meaning that the author intends to convey (signal = point out) could be different from the meaning that would be construed by monolingual native English readers (signal = to make a sound or movement that tells someone something), but not necessarily different from the meaning construed by other multilingual readers, who may be familiar with cognates of señalar (meaning point out) in other European languages (e.g. signaler in French, segnalare in Italian). This raises the question of the fluidity of lexical meaning in intercultural communication. It should not be taken for granted that English users from different linguistic backgrounds share identical meanings of a lexical item (Xu & dinh, 2013), but this semantic divergence does not necessarily lead to misunderstanding, since in intercultural communication lexical meanings are co-constructed and negotiated in context. The audience of articles in international journals is composed of readers from different multilingual backgrounds, who will draw on their linguistic repertoires (including their L1 and other languages they may know) to construct meaning. In example (8) the use of signal has been accepted as valid in the reviewing process, which suggests that certain semantic divergences might be overlooked both by authors and readers, as long as there is no major communication breakdown. VARIATIon In ACAdEMIC wRITIng PRACTICES Ibérica 36 (2018): 171-194 181 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 181 A comparison of the frequency of the RVs in column 2 in the different sections of the CoCA reveals that some of them are less common in written academic English than in other more informal registers, e.g. talk about (760 occurrences per million in the spoken section vs. 51.70 in the academic section), play down (3.58 occurrences per million in the newspaper section vs. 0.83 in the academic section). This is also the case with some verbs which occur more frequently in the SPEngAL than in the EngAL corpus, e.g. say (5,199 occurrences per million in the spoken section vs. 946 in the academic section), put it (95 occurrences per million in the spoken section vs. 36 in the academic section). This suggests that SPEngAL authors may be less sensitive to register variation in English than EngAL authors. This is in agreement with previous research revealing that the academic texts of L2 writers display some elements more typical of spoken discourse (Hinkel, 2003) and with Cohen’s (2001) recount of his own difficulty in discriminating between formal and informal register when using academic Hebrew. This result is also related to the finding that the development of L2 writing involves a move from the features of spoken English to those typical of formal writing (Shaw & Liu, 1998). 3.2. Types of RVs in the SPENGAL and ENGAL sub-corpora Table 3 shows the top RVs in the two corpora (31 in the SPEngAL corpus, 30 in the EngAL corpus) with their frequency of occurrence. The first figure indicates the number of instances of the RV and the second figure expresses the frequency of each RV as a percentage of all RVs in each sub- corpus. MARíA-JoSé LUzón Ibérica 36 (2018): 171-194182 VARIATION IN ACADEMIC WRITING PRACTICES Ibérica 36 (2018): …-… SPENGAL ENGAL Verb total % (of 648) Verb total % (of 631) point out 45 6.94 find 74 11.72 propose 36 5.55 propose 54 8.55 argue 34 5.24 note 48 7.6 claim 29 4.47 argue 43 6.8 show 25 3.85 suggest 39 6.18 consider 25 3.85 show 38 6.02 find 23 3.54 discuss 18 2.85 suggest 23 3.54 identify 17 2.69 state 22 3.4 assume 17 2.69 note 21 3.24 analyze 16 2.53 define 14 2.16 examine 14 2.21 explain 14 2.16 consider 13 2.06 analyze 13 2.00 claim 13 2.06 focus 13 2.00 describe 13 2.06 describe 12 1.85 study 12 1.9 distinguish 11 1.7 point out 12 1.9 Table 3. Top RVs in the two corpora. Drawing on Hyland’s (2002) distinction, the verbs in Table 3 were classified into research verbs (i.e., verbs referring to procedures/experimental activities or used to state findings), discourse verbs (i.e., verbs involving verbal expression) and cognitive verbs (i.e., verbs referring to mental processes). Table 4 below shows the classification of the top RVs used in the two corpora and the percentage of each type. 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 182 drawing on Hyland’s (2002) distinction, the verbs in Table 3 were classified into research verbs (i.e., verbs referring to procedures/experimental activities or used to state findings), discourse verbs (i.e., verbs involving verbal expression) and cognitive verbs (i.e., verbs referring to mental processes). Table 4 below shows the classification of the top RVs used in the two corpora and the percentage of each type. VARIATIon In ACAdEMIC wRITIng PRACTICES Ibérica 36 (2018): 171-194 183 VARIATION IN ACADEMIC WRITING PRACTICES Ibérica 36 (2018): …-… distinguish 11 1.7 point out 12 1.9 mention 10 1.54 present 11 1.74 maintain 9 1.38 call 9 1.42 accept 8 1.23 compare 9 1.42 call 8 1.23 report 9 1.42 stress 8 1.23 cite 8 1.26 conclude 7 1.08 distinguish 8 1.36 examine 7 1.08 investigate 8 1.36 observe 7 1.08 use 8 1.36 point to 7 1.08 provide 7 1.1 provide 7 1.08 develop 7 1.1 put it 7 1.08 document 7 1.1 comment 6 0.92 observe 6 0.95 compare 6 0.92 focus 6 0.95 identify 6 0.92 refer 6 0.95 study 6 0.92 Total 470 Total 550 Table 3. Top RVs in the two corpora. Drawing on Hyland’s (2002) distinction, the verbs in Table 3 were classified into research verbs (i.e., verbs referring to procedures/experimental activities or used to state findings), discourse verbs (i.e., verbs involving verbal expression) and cognitive verbs (i.e., verbs referring to mental processes). Table 4 below shows the classification of the top RVs used in the two corpora and the percentage of each type. MARÍA-JOSÉ LUZÓN Ibérica 36 (2018): …-… Verb group SPENGAL ENGAL Verbs % of total Verbs % of total research show, find, analyze, focus, observe, examine, study (7 verbs/ 95 occurrences) 20.21 find, show, analyze, examine, study, investigate, use, develop, observe, focus (10 verbs/ 189 occurrences) 34.36 discourse point out, propose, argue, claim, suggest, state, note, define, explain, describe, distinguish, mention, maintain, accept, stress, call, compare, conclude, put it, point to, provide, comment, identify (23 verbs/ 350 occurrences) 74.46 propose, note, argue, suggest, discuss, identify, claim, describe, point out, present, call, report, compare, cite, distinguish, provide, document, refer (18 verbs/ 331 occurrences) 60.18 cognitive consider (1 verb/ 25 occurrences) 5.31 assume, consider (2 verbs/ 30 occurrences) 5.45 470 100 550 100 Table 4. Types of RVs in the SPENGAL and the ENGAL sub-corpora. The percentage of cognitive RVs is very low in both corpora, consider and assume being the only cognitive verbs among the top RVs. However, there is a striking difference in the frequency of use of research and discourse verbs in the two corpora. Although there is a clear preference for discourse verbs over research verbs in the two corpora, scholars in an Anglophone context used a much higher number of research verbs than Spanish scholars. Table 3 shows the high frequency in the ENGAL corpus of two research verbs used to state findings, find, show (which together comprise 17.74% of the RVs in the ENGAL corpus, compared to 7.39% in the SPENGAL corpus). By contrast, the most frequent RV in the SPENGAL corpus is the discourse verb point out (6.94% of the RVs in the SPENGAL corpus vs. 1.9% of the RVs in the ENGAL corpus). Hyland and Tse (2005: 60) suggest that academic writers use RVs strategically to focus on a specific type of support for their own claims: “By framing evaluations with either a research, cognitive, or discoursal focus, writers can also implicitly signal whether they intend their judgements to be understood as grounded in research practices, interpretive practices, or reporting practices”. Hyland (2009) found a clear preference for research act verbs in the hard disciplines and for discourse act verbs in the soft disciplines and attributed it to the differences in the way knowledge is constructed, with the hard disciplines emphasizing the importance of experimental results to support claims and the 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 183 The percentage of cognitive RVs is very low in both corpora, consider and assume being the only cognitive verbs among the top RVs. However, there is a striking difference in the frequency of use of research and discourse verbs in the two corpora. Although there is a clear preference for discourse verbs over research verbs in the two corpora, scholars in an Anglophone context used a much higher number of research verbs than Spanish scholars. Table 3 shows the high frequency in the EngAL corpus of two research verbs used to state findings, find, show (which together comprise 17.74% of the RVs in the EngAL corpus, compared to 7.39% in the SPEngAL corpus). By contrast, the most frequent RV in the SPEngAL corpus is the discourse verb point out (6.94% of the RVs in the SPEngAL corpus vs. 1.9% of the RVs in the EngAL corpus). Hyland and Tse (2005: 60) suggest that academic writers use RVs strategically to focus on a specific type of support for their own claims: “By framing evaluations with either a research, cognitive, or discoursal focus, writers can also implicitly signal whether they intend their judgements to be understood as grounded in research practices, interpretive practices, or reporting practices”. Hyland (2009) found a clear preference for research act verbs in the hard disciplines and for discourse act verbs in the soft disciplines and attributed it to the differences in the way knowledge is constructed, with the hard disciplines emphasizing the importance of experimental results to support claims and the humanities relying on the strength of arguments and on the reference to the words of previous researchers. As can be seen, both in the SPEngAL and the EngAL corpus discourse act verbs are more frequent than research act verbs, which shows the importance of argument and interpretation in linguistics. However, the higher frequency of discourse act verbs in the SPEngAL corpus when compared with the EngAL corpus could suggest that, while both Spanish and Anglophone writers prefer to ground their claims on the cited authors’ reporting of and interpretation of their own research, Spanish writers are less willing to emphasize the research results of cited authors- as in example (9)- instead of their authority- see example 10. (9) Laufer (1989) found that learners who knew 95 per cent of the words in text were more likely to be successful readers (EngAL) (10) gains (1999) claims that the informality of e-mail messages is not retained in business messages (SPEngAL) MARíA-JoSé LUzón Ibérica 36 (2018): 171-194184 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 184 3.3. Types of citations of which reporting verbs are part Table 5 gives the frequency of RVs in the three different types of citation considered in the research. It should be pointed out that I did not count the occurrences of the different types of citation in the corpora, but the frequency of these types with a RV. A chi-square test was carried out in order to determine the significance of the differences in the two corpora. The distribution of citation types with RVs was found to be significantly different between the two corpora (χ²= 50.92, df = 2, p < 0.0001). Table 5 shows a striking difference in the use of RVs in the two corpora: the percentage of RVs+ quotation is much higher in the SPEngAL corpus than in the EngAL corpus (18.51% vs. 5.7%). This suggests that Spanish writers resort to quotation much more frequently than Anglophone writers. A possible explanation is that, since writers use direct quotations to give other authors’ voices the floor (Fløttum et al., 2006), Spanish scholars tend to use more quotations to provide a more authoritative support to their claim, as illustrated in example (11). (11) First, we have provided further evidence that discourse is pre-patterned (…). As Sinclair puts it (Sinclair, 1991: 108): “by far the majority of text is made of the occurrence of common words in common patterns”. (SPEngAL) It is also interesting to compare which RVs are used in quotations: VARIATIon In ACAdEMIC wRITIng PRACTICES Ibérica 36 (2018): 171-194 185 VARIATION IN ACADEMIC WRITING PRACTICES Ibérica 36 (2018): …-… humanities relying on the strength of arguments and on the reference to the words of previous researchers. As can be seen, both in the SPENGAL and the ENGAL corpus discourse act verbs are more frequent than research act verbs, which shows the importance of argument and interpretation in Linguistics. However, the higher frequency of discourse act verbs in the SPENGAL corpus when compared with the ENGAL corpus could suggest that, while both Spanish and Anglophone writers prefer to ground their claims on the cited authors’ reporting of and interpretation of their own research, Spanish writers are less willing to emphasize the research results of cited authors- as in example (9)- instead of their authority- see example 10. (9) Laufer (1989) found that learners who knew 95 per cent of the words in text were more likely to be successful readers (ENGAL) (10) Gains (1999) claims that the informality of e-mail messages is not retained in business messages (SPENGAL) Laufer (1989) found that learners who knew 95 per cent of the words in text were more likely to be successful readers and had better comprehension scores. 3.3. Types of citations of which reporting verbs are part Type of citation SPENGAL ENGAL total % total % Integral citation (excluding quotations) 415 64.04 487 77.17 Non-integral citation (excluding quotations) 113 17.43 108 17.11 Quotation 120 18.51 36 5.7 Total 648 100 631 100 Table 5. Occurrences of reporting verbs in integral citations, non-integral citations and quotations. Table 5 gives the frequency of RVs in the three different types of citation considered in the research. It should be pointed out that I did not count the occurrences of the different types of citation in the corpora, but the frequency of these types with a RV. A chi-square test was carried out in order to determine the significance of the differences in the two corpora. The distribution of citation types with RVs was found to be significantly different between the two corpora ("#= 50.92, df = 2, p < 0.0001). Table 5 shows a striking difference in the use of RVs in the two corpora: the percentage of RVs+ quotation is much higher in the SPENGAL corpus than in the ENGAL corpus (18.51% vs. 5.7%). This suggests that Spanish writers resort to quotation much more frequently than Anglophone writers. A possible explanation is that, since writers use direct quotations to give other authors’ voices the floor (Fløttum et al., 2006), Spanish scholars tend to use more quotations to provide a more authoritative support to their claim, as illustrated in example (11). MARÍA-JOSÉ LUZÓN Ibérica 36 (2018): …-… (11) First, we have provided further evidence that discourse is pre-patterned (…). As Sinclair puts it (Sinclair, 1991: 108): “by far the majority of text is made of the occurrence of common words in common patterns”. (SPENGAL) It is also interesting to compare which RVs are used in quotations: SPENGAL (120) ENGAL (36) state (13), define (10), point out (10), call (8), explain (8), argue (7), put it (6), describe (5), maintain (5), claim (4), conclude (4), note (4), observe (4), say (4), refer (3), stress (3), consider (3), mention (2), signal (2), suggest (2), talk (2), remark (2), write (2), add (1), advocate (1), comment (1), find (1), prove (1), show (1), underline (1) note (10), call (5), describe (4), observe (2), argue (2), claim (2), refer (2), state (2), conclude (1), consider (1), define (1), propose (1), put it (1), say (1), suggest (1) Table 6. Verbs occurring in quotations in the SPENGAL corpus and the ENGAL corpus5. As can be seen, the most frequent RVs in quotation are different in the two corpora. Point out and note are both factive verbs (Hyland, 2002) used to show that the writer accepts the truth of what the original author said, but while ENGAL authors prefer note, SPENGAL authors prefer point out. Note occurs in 28.57% of the quotations in the ENGAL corpus and only in 3.3% of the quotations in the SPENGAL corpus. Similarly, point out, which occurs in 8.3% of the quotations in the SPENGAL corpus, is absent in the ENGAL corpus. Some verbs are closely associated to quotation in the SPENGAL corpus: 85% of the occurrences (6 out of 7) of put it and 59% of the occurrences of state are used to introduce a quotation. These are non-factive verbs, always used by Spanish writers in the corpus to introduce a quotation which supports the writer’s statement (see examples 11 and 12). (12) A close examination of the Introduction unit of the abstracts revealed that this section was the most complex unit (…). Swales himself draws our attention to the complexity of research article introductions when he states that: Introductions are known to be troublesome [...].(SPENGAL) In the SPENGAL corpus state usually occurs in the patterns: “As stated by+ author”, “As author(s)+ state(s)”, “Accordingly, author(s)+ state(s) that”. The use of state in the SPENGAL corpus is similar to that in the learner corpus analyzed by Bloch (2010). He found that in expert writing state was sometimes used to present a claim with which the writer did not agree, while in the learner corpus state tended to be used to report the cited author’s claims as facts. This is interesting because the evaluative potential of words seems to be easily available to Anglophone authors, but it may not be so evident for EAL authors. Therefore, the evaluative meaning of RVs in academic texts by EAL writers might not be exactly the same as the meaning construed by Anglophone speakers. Table 7 shows the frequency of RVs in the different citation patterns. Interestingly, the relative frequency of each pattern was not significantly 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 185 As can be seen, the most frequent RVs in quotation are different in the two corpora. Point out and note are both factive verbs (Hyland, 2002) used to show that the writer accepts the truth of what the original author said, but while EngAL authors prefer note, SPEngAL authors prefer point out. Note occurs in 28.57% of the quotations in the EngAL corpus and only in 3.3% of the quotations in the SPEngAL corpus. Similarly, point out, which occurs in 8.3% of the quotations in the SPEngAL corpus, is absent in the EngAL corpus. Some verbs are closely associated to quotation in the SPEngAL corpus: 85% of the occurrences (6 out of 7) of put it and 59% of the occurrences of state are used to introduce a quotation. These are non-factive verbs, always used by Spanish writers in the corpus to introduce a quotation which supports the writer’s statement (see examples 11 and 12). (12) A close examination of the Introduction unit of the abstracts revealed that this section was the most complex unit (…). Swales himself draws our attention to the complexity of research article introductions when he states that: Introductions are known to be troublesome [...].(SPEngAL) In the SPEngAL corpus state usually occurs in the patterns: “As stated by+ author”, “As author(s)+ state(s)”, “Accordingly, author(s)+ state(s) that”. The use of state in the SPEngAL corpus is similar to that in the learner corpus analyzed by Bloch (2010). He found that in expert writing state was sometimes used to present a claim with which the writer did not agree, while in the learner corpus state tended to be used to report the cited author’s claims as facts. This is interesting because the evaluative potential of words seems to be easily available to Anglophone authors, but it may not be so evident for EAL authors. Therefore, the evaluative meaning of RVs in academic texts by EAL writers might not be exactly the same as the meaning construed by Anglophone speakers. Table 7 shows the frequency of RVs in the different citation patterns. Interestingly, the relative frequency of each pattern was not significantly different between the two corpora (χ²= 3.33 df= 4, p > 0.5), which suggests that both groups of expert writers in linguistics share genre knowledge concerning the function of the different citation patterns in RAs in their discipline. As can be seen, in both corpora the most frequent pattern where RVs occur is the one where the author is most visible, pattern 1.1.: “Human MARíA-JoSé LUzón Ibérica 36 (2018): 171-194186 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 186 subject (cited author)+ RV”, and the other patterns are used with a similar frequency. 4. Discussion and conclusions The analysis of the two corpora of English-medium RAs in linguistics, one written by Anglophone scholars and the other by Spanish scholars, shows variation regarding the use of RVs in academic writing. The present findings thus indicate that the discourse of English-medium RAs written by Spanish expert writers displays many linguistic features and rhetorical practices found in texts written by Anglophone scholars, but also examples of language use that diverge from Anglophone style and norms. Both corpora display a similar number of occurrences of RVs and a preference for discourse RVs to construct knowledge. This may be attributed to the fact that both groups of scholars are expert users of academic English who make an effective use of RVs to achieve the goals of their disciplinary cultures. However, although the corpus analyzed is not large enough to make confident generalizations, the study has revealed some differences in the way these two groups of writers used RVs in English- medium research articles, which allows for some tentative conclusions. The analysis of the corpora showed that Spanish scholars used a higher percentage of discourse act RVs than Anglophone scholars and a higher percentage of RVs in quotations, thus giving the cited author the floor more VARIATIon In ACAdEMIC wRITIng PRACTICES Ibérica 36 (2018): 171-194 187 VARIATION IN ACADEMIC WRITING PRACTICES Ibérica 36 (2018): …-… different between the two corpora ("#= 3.33 df= 4, p > 0.5), which suggests that both groups of expert writers in Linguistics share genre knowledge concerning the function of the different citation patterns in RAs in their discipline. As can be seen, in both corpora the most frequent pattern where RVs occur is the one where the author is most visible, pattern 1.1.: “Human subject (cited author)+ RV”, and the other patterns are used with a similar frequency. SPENGAL (648 occurrences) ENGAL (631 occurrences) Patterns total % of RVs total % of RVs Integral structures 1.1. Human subject+ RV (Biber et al. show that) 518 79.93 506 80.19 1.2. Non-human subject+ RV (recent work by Aijmer emphasizes that) 9 1.38 11 1.74 Non-integral structures 2.1. General inanimate reference+ RV (Some studies have provided) 51 7.87 58 9.19 2.2. General animate reference+ RV (Some writers emphasize that) 23 3.54 13 2.06 2.3. Introductory it or RV in passive voice (Research into … has been critiqued for) 47 7.25 43 6.81 Total 648 100 631 100 Table 7. Citation patterns of which RVs are part. 4. Discussion and conclusions The analysis of the two corpora of English-medium RAs in Linguistics, one written by Anglophone scholars and the other by Spanish scholars, shows variation regarding the use of RVs in academic writing. The present findings thus indicate that the discourse of English-medium RAs written by Spanish expert writers displays many linguistic features and rhetorical practices found in texts written by Anglophone scholars, but also examples of language use that diverge from Anglophone style and norms. Both corpora display a similar number of occurrences of RVs and a preference for discourse RVs to construct knowledge. This may be attributed to the fact that both groups of scholars are expert users of academic English who make an effective use of RVs to achieve the goals of their disciplinary cultures. However, although the corpus analyzed is not large enough to make confident generalizations, the study has revealed some differences in the way these two groups of writers used RVs in English-medium research papers, which allows for some tentative conclusions. The analysis of the corpora showed that Spanish 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 187 often. According to the findings, the group of EAL scholars investigated seems to have a clearer preference for citation practices that enable them to support their own claims by recourse to other scholars’ ideas, assertions and interpretation. Thus, the Spanish authors in this study could be exploiting RVs to make their knowledge claims acceptable by the editors and reviewers of high impact international journals and thus secure publication in these journals. The study has also revealed that, although the overall frequency of RVs was similar in both corpora, the texts written by Spanish scholars displayed a higher number of different verbs, some of them with very few occurrences. Spanish scholars utilized non-routine verb+noun combinations and a high number of verbs from a Latin origin, both true and false cognates. Consistent with previous research (Pérez-Llantada, 2014; Lorés- Sanz, 2016) this variability indicates that the EAL writers activated their L1 literacy resources to meet their communicative needs when writing in English. The comparison of the two corpora also reveals slight differences in the use of some RVs: some verbs are more strongly associated with specific patterns or specific functions in one corpus than in the other (e.g. the use of point out, state and put it to introduce quotations in the SPEngAL corpus). The present findings suggest that, although the two groups of writers share most generic conventions regarding RVs, the usage of these verbs in the writing of Spanish scholars is influenced by their multilingual literacy repertoire, which supports previous research showing the “hybrid” practices (“largely, but not completely, native-like”) of EAL scholars (Ädel & Erman, 2012; Pérez-Llantada, 2014: 84). The analysis of the data also suggests emerging features in academic writing by EAL scholars (or, at least, in the academic writing by Spanish scholars) (e.g. verbs of spoken register, frequent use of cognates), which, as a result of the increasing number of EAL researchers publishing in international journals, could, as Rozycki and Johnson (2013) point out, contribute to gradually altering the features of academic written communication. An interesting finding is that the Spanish writers in this study seem to be relatively less register-sensitive than Anglophone writers and they may also overlook subtle semantic divergences when using “false” cognates or when using some RVs for evaluation. It can thus be concluded that a key to effective communication at the receptive level is relative tolerance towards register and code variations, for instance, accepting usage that might be slightly less formal than expected or accepting words whose meaning is slightly different from the one that would be construed by a native speaker. MARíA-JoSé LUzón Ibérica 36 (2018): 171-194188 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 188 As convincingly argued by Belcher, there is a need for a “new reader responsibility”, based on flexibility and contextual adaptation: readers of international publications need to be flexible and “willing, when needed, to accept responsibility for constructing meaning from what may seem less- than-immediately-transparent text” (Belcher, 2014: 65). As Horner concludes, EAL (multilingual) writers have to employ attitudes like “tolerance for language variation” and “strategies of accommodation and negotiation” (Horner, 2011: 302). The findings of this study are consistent with previous research showing the presence of non-standard forms, innovative rhetorical patterns, and culture- specific features in peer-reviewed academic writing (Rozycki & Johnson, 2013; Pérez-Llantada, 2014; Lorés-Sanz, 2016) and support Hyland’s (2016) call for a more inclusive view of academic publishing, not based on L1 vs. L2 dichotomies. The study has thus sought to contribute to providing information on variations that are successfully used by EAL authors to achieve their purposes and shows the need to conceptualize academic English as English-medium discourse characterized by variability, fluidity and hybridity, with room for the diverse rhetorical styles and textual choices of expert scholars (including Anglophone scholars) with different L1 backgrounds, who use various linguistic/ semiotic resources to report research. The results of this study have implications both for academic literacy instruction and for publication reviewing and gate-keeping. one important implication of this study is that the teaching of academic English should be approached from a “language variation” perspective, acknowledging the discourse practices of scholars from varied linguistic backgrounds, as also postulated by wingate (2015) for students. Aligning with Curry and Lillis (2004) and Belcher (2007), the results of this study highlight the need for academic writing courses for EAL scholars, which raise their awareness of how expert writers (both Anglophone and non-Anglophone) make rhetorical choices intended to develop their authorial identity and get their claims accepted. As Heng Hartse and Kubota (2014) argue, an “error-based” approach, which focuses on identifying non-standard forms in the writing of L2 scholars and helping them to avoid these “errors”, is not adequate, because, as this study has shown, some forms that are not preferred by Anglophone scholars are accepted in actual practice. Materials to be used in those courses and criteria to assess acceptable performance should, therefore, not be based on corpus observation (data driven Learning) of VARIATIon In ACAdEMIC wRITIng PRACTICES Ibérica 36 (2018): 171-194 189 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 189 what Anglophone scholars do, but on the study of the processes and texts of expert scholars from different linguistic backgrounds. These materials would help to promote the acceptance of varying norms, lexico-grammatical forms that might deviate from the Anglophone norms, and creative uses of language intended to achieve communicative effectiveness. As Hyland (2016) also contends, the assumption that conformity to native-speaker standards affects paper acceptance may discourage EAL authors from revising their manuscripts. Therefore, materials exposing scholars who write in English to a diversity of discursive practices may reassure them when submitting their contributions to English-medium journals. This is in line with a critical pragmatic approach to the teaching of academic writing, which acknowledges the need to teach dominant discourse conventions, but also encourages students to question these conventions (Harwood & Hadley, 2004). In terms of publication reviewing, this study supports the recommendation of other scholars (Belcher, 2007; Heng Hartse & Kubota, 2014) that reviewers, editors, and other literacy brokers should accept linguistic and discursive choices that diverge from Anglophone conventions but are effective and should not assess submissions against Anglophone norms but against the criterion of international intelligibility. A sensible practice when considering the acceptability of forms in academic writing by EAL users is the “variation-oriented” perspective proposed by Heng Hartse & Kubota (2014: 82): the acceptance of difference and variation while maintaining high standards of intelligibility, communicative effectiveness, and linguistic accuracy. one limitation of this study is that it has involved EAL scholars in a single discipline and in a single geopolitical context. Therefore, future research is necessary on the use of RVs and function of citations in the English- medium writing of EAL scholars in other disciplines and locations. This research would contribute to identifying possible features which, even if infrequent in the writing of Anglophone authors, are preferred by multilingual scholars from different L1 backgrounds. A study of the use of RVs and citation patterns by Anglophone and multilingual scholars in different disciplines could also reveal differences in how the practices of multilingual scholars diverge from those of Anglophone scholars across disciplines, which would help to understand the rationale for the citation practices of scholars from different linguistic backgrounds and in different social and cultural contexts. Finally, in line with current ethnographic studies MARíA-JoSé LUzón Ibérica 36 (2018): 171-194190 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 190 on academic writing and literacy practices (Lillis & Curry, 2010), this study could be complemented with interview-based research where some of the authors in the corpora could provide information about their citation practices and explain the motivations for their preferences regarding specific RVs or citation patterns, thus furthering our understanding of the situatedness of citation practices in English-medium texts. Acknowledgements This research was supported by the Spanish Ministry of Economy and Competitiveness (project code: FFI2012-37346; and project code: FFI2015- 68638-R) and by the Regional government of Aragón (project code: H21). Article history: Received 9 November 2017 Received in revised form 2 February 2018 Accepted 2 February 2018 References VARIATIon In ACAdEMIC wRITIng PRACTICES Ibérica 36 (2018): 171-194 191 Ädel, A. & B. Erman (2012). “Recurrent word combinations in academic writing by native and non-native speakers of English: A lexical bundles approach”. English for Specific Purposes 31,2: 81- 92. Belcher, D. (2007). “Seeking acceptance in an English-only research world”. Journal of Second Language Writing 16: 1-22. Belcher, D. (2014). “What we need and don’t need intercultural rhetoric for: A retrospective and prospective look at an evolving research area”. Journal of Second Language Writing, 25: 59-67. Bloch, J. (2010). “A concordance-based study of the use of reporting verbs as rhetorical devices in academic papers”. Journal of Writing Research 2: 219-244. Canagarajah, A.S. (2013). Translingual Practice: Global Englishes and Cosmopolitan Relations. New York: Routledge. Cargill, M. & S. Burgess (eds.) (2017). Publishing Research in English as an Additional Language: Practices, Pathways and Potentials. Adelaide: University of Adelaide Press. Charles, M. (2006). “Phraseological patterns in reporting clauses used in citation: A corpus-based study of theses in two disciplines”. English for Specific Purposes 25: 310-331. Cohen, A.D. (2001). “From L1 to L2: The confessions of a sometimes frustrated multiliterate” in D. Belcher & U. Connor (eds.), Reflections on Multiliterate Lives, 79-95. Clevedon: Multilingual Matters. Curry, J. M. & T. Lillis (2004). “Multilingual scholars and the imperative to publish in English: Negotiating interests, demands, and rewards”. TESOL Quarterly 38: 663-688. Dontcheva-Navratilova, O. (2016). “Rhetorical functions of citations in linguistics research articles: A contrastive (English-Czech) study”. Discourse and Interaction 9/2: 51-74. Fløttum, K., T. Dahl & T. Kinn (2006). Academic Voices Across Languages and Disciplines. Amsterdam: Benjamins. Hanauer, D.I. & K. Englander (2013). Scientific Writing in a Second Language. Anderson, SC: Parlor Press. Harwood, N. (2009). “An interview-based study of the functions of citations in academic writing across two disciplines”. Journal of Pragmatics 41: 497-518. Harwood, N. & G. Hadley (2004). “Demystifying institutional practices: Critical pragmatism and the 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 191 MARíA-JoSé LUzón Ibérica 36 (2018): 171-194192 teaching of academic writing”. Journal of English for Specific Purposes 23: 355-377. Heng Hartse, J. & R. Kubota (2014). “Pluralizing English? Variation in high-stakes academic texts and challenges of copyediting”. Journal of Second Language Writing 24: 71-82. Hewings, A., T. Lillis, & D. Vladimirou (2010). “Who’s citing whose writings? A corpus- based study of citations as interpersonal resource in English medium national and English medium international journals”. Journal of English for Academic Purposes 9: 102-115. Hinkel, E. (2003). “Simplicity without elegance: Features of sentences in L1 and L2 academic texts”. TESOL Quarterly 37: 275-301. Horner, B. (2011). “Writing English as a lingua franca” in A. Archibald, A. Cogo & J. Jenkins (eds.), Latest trends in ELF research, 299-311. Newcastle upon Tyne: Cambridge Scholars Press Hryniuk, K. (2016). “The use of citations in research articles written by Polish and English native-speaker writers” in H. Chodkiewicz, P. Steinbrich & M. Krzemińska-Adamek (eds.), Working with Text and Around Text in Foreign Language Environments. Springer, Cham. Hu, G.W. & G.H. Wang (2014). “Disciplinary and ethnolinguistic influences on citation in research articles”. Journal of English for Academic Purposes 14: 14-28. Hyland, K. (1999). “Academic attribution: Citation and the construction of disciplinary knowledge”. Applied Linguistics 20: 341-367. Hyland, K. (2002). “Activity and evaluation: Reporting practices in academic writing” in J. Flowerdew (ed.), Academic Discourse, 115-130. Harlow, England: Longman. Hyland, K. & P. Tse (2005). “Evaluative that constructions: Signaling stance in research abstracts”. Functions of Language 12: 39-64. Hyland, K. (2009). “Writing in the disciplines”. Taiwan International ESP Journal 1: 5-22. Hyland, K. (2016). “Academic publishing and the myth of linguistic injustice”. Journal of Second Language Writing 31: 58-69. Jenkins, J. (2015). “Repositioning English and multilingualism in English as a Lingua Franca”. Englishes in Practice 2: 49-85. Kuteeva, M. & A. Mauranen (eds.) (2014). Writing for international publication in multilingual contexts. Special issue of Journal of English for Academic Purposes 13. Lillis, T. & M.J. Curry (2010). Academic Writing in a Global Context: The Politics and Practices of Publishing in English. Abingdon, Oxford, UK: Routledge. Lillis, T., A. Hewings, D. Vladimirou, & M.J. Curry (2010). “The geolinguistics of English as an Academic Lingua Franca: Citation practices across English medium national and English medium international journals”. International Journal of Applied Linguistics 20: 111-135. Lillis, T. & J. Tuck (2016). “Academic literacies: A critical lens on writing and reading in the academy” in K. Hyland & P. Shaw (eds.), The Routledge Handbook of English for Academic Purposes, 30- 43. Abingdon: Routledge. Lorés-Sanz, R. (2016). “ELF in the making?: Simplification and hybridity in abstract writing”. Journal of English as a Lingua Franca 5: 53-81. Mansourizadeh K. & U.K. Ahmad (2011) “Citation practices among non-native expert and novice scientific writers”. Journal of English for Academic Purposes 10: 152-161. Mauranen, A. (2012). Exploring ELF. Academic English Shaped by Non-native Speakers. Cambridge University Press, Cambridge. Mauranen, A., N. Hynninen, & E. Ranta (2010). “English as an academic lingua franca: The ELFA project”. English for Specific Purposes 29: 183- 190. Mur-Dueñas, P. (2009). “Citation in business management RAs: A contrastive corpus-based analysis” in E. Suomela-Salmi & F. Dervin (eds.), Cross-cultural and Cross-linguistic Perspectives on Academic Discourse, 49-60. Amsterdam/ Philadelphia: John Benjamins. Okamura, A. (2008). “Citation forms in scientific texts: Similarities and differences in L1 and L2 professional writing”. Nordic Journal of English Studies 7,3: 61-81. Pérez-Llantada, C. (2012). Scientific Discourse and the Rhetoric of Globalization. New York, Continuum. Pérez-Llantada, C. (2014). “Formulaic language in L1 and L2 expert academic writing: Convergent and divergent usage”. Journal of English for Academic Purposes 14: 84-94. Petrić, B. (2007). “Rhetorical functions of citations in high- and low-rated master’s theses”. Journal of English for Academic Purposes 6: 238-253. Petrić, B. & N. Harwood (2013) “Task requirements, task representation, and self- reported citation functions: An exploratory study of a successful L2 student’s writing”. Journal of 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 192 María José Luzón is Senior Lecturer at the University of zaragoza, Spain. She has a Phd in English Philology and has published papers on academic and professional discourse and on language teaching and learning in the field of English for Academic Purposes. Her current research interests include the analysis of online academic genres, especially academic blogs, and the analysis of ELF in academic online and written discourse. NoTES 1 For the sake of brevity, the terms “Anglophone scholars”/ “Anglophone writers”, and “Spanish scholars”/ “Spanish writers” are used herein to refer to these groups. 2 For a more detailed description of the SERAC corpus and of the criteria used in its compilation see Pérez-Llantada (2012). 3 These are Anglophone journals, listed in the Social Science Citation Index, which in the years of compilation had a high impact factor in the discipline (higher than 1). 4 Following Thompson and Ye (1991: 361), the term “author” is used to refer to “the person who is being reported” and “writer” to refer to “the person who is reporting”. 5 Bold has been used to highlight verbs that occur in both corpora. VARIATIon In ACAdEMIC wRITIng PRACTICES Ibérica 36 (2018): 171-194 193 English for Academic Purposes 12: 110-124. Rowley-Jolivet, E. & S. Carter-Thomas (2014). “Citation practices of expert French writers of English: Issues of attribution and stance” in A. Łyda & K. Warchał (eds.), Occupying Niches: Interculturality, Cross-culturality and Aculturality in Academic Research, 17-34. New York: Springer. Rozycki, W. & N. Johnson (2013). “Non-canonical grammar in Best Paper award winners in engineering”. English for Specific Purposes 32,3: 157-169. Samraj, B. (2013). “Form and function of citations in discussion sections of master’s theses and research articles”. Journal of English for Academic Purposes 12: 299-310. Scott, M. (2008). WordSmith Tools version 5. Liverpool: Lexical Analysis Software. Shaw, P. (1992) “Reasons for the correlation of voice, tense and sentence function in reporting verbs.” Applied Linguistics 13,3: 302-319. Seidlhofer, B. (2011). Understanding English as a Lingua Franca. Oxford: Oxford University Press. Shaw, P. & E. Liu (1998). “What develops in the development of second language writing? Applied Linguistics 19: 225-254. Tribble, C. (2017). “ELFA vs. Genre: A new paradigm war in EAP writing instruction?. Journal of English for Academic Purposes 25: 30-44. Thompson, G. & Y. Ye (1991). “Evaluation in the reporting verbs used in academic papers”. Applied Linguistics 12: 365-382. Thompson, P. & C. Tribble (2001). “Looking at citations: Using corpora in English for Academic Purposes”. Language Learning and Technology 5: 91-105. Uzuner, S. (2008). “Multilingual scholars’ participation in core/global academic communities: A literature review”. Journal of English for Academic Purposes 7: 250-263. Wingate, U. (2015). Academic Literacy and Student Diversity: The Case for Inclusive Practice. London: Multilingual Matters. Xu, Z. & T.N. Dinh (2013). “How do ‘WE’ (World Englishes) make sense in ELF communication? Words and their meanings across cultures”. Journal of English as a Lingua Franca 2,2: 365 - 388. 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 193 Appendix MARíA-JoSé LUzón Ibérica 36 (2018): 171-194194 MARÍA-JOSÉ LUZÓN Ibérica 36 (2018): …-… NOTES 1 For the sake of brevity, the terms “Anglophone scholars”/ “Anglophone writers”, and “Spanish scholars”/ “Spanish writers” are used herein to refer to these groups. 2 For a more detailed description of the SERAC corpus and of the criteria used in its compilation see Pérez-Llantada (2012). 3 These are Anglophone journals, listed in the Social Science Citation Index, which in the years of compilation had a high impact factor in the discipline (higher than 1). 4 Following Thompson and Ye (1991: 361), the term “author” is used to refer to “the person who is being reported” and “writer” to refer to “the person who is reporting”. 5 Bold has been used to highlight verbs that occur in both corpora. Appendix VERB: ENGAL corpus (Number of occurrences) SPENGAL corpus (Number of occurrences) Type of citation 1.Integral 2.Non-integral 3.Quotation Syntactic pattern 1.1. (I) Human subject (cited author)+ RV 1.2. (I) Non-human subject+ RV 2.1. (Non-I) General inanimate reference+ RV 2.2. (Non-I) General animate reference+ RV 2.3. (Non-I) RV in passive voice 07 IBERICA 36_Iberica 13 26/10/18 20:23 Página 194