Iberica 13 Ibérica 41 (2021): 131-154 ISSN: 1139-7241 / e-ISSN: 2340-2784 Abstract The recent rise in the number of English-medium journals in non-Anglophone countries stems from the hegemony of English in global scholarly communication, and language requirements outlined by such journals to authors offer a valuable yet often overlooked source of information in terms of underlying values and assumptions concerning knowledge production and circulation. Drawing on the notion of “semiperiphery” to refer to Brazil’s current standing within global scientific publishing, this exploratory study focuses on author guidelines of Brazilian English-medium journals indexed in the SciELO database to map language requirements established for manuscript submission and to assess, from a discursive perspective, whether they reinforce or disprove dominant ideologies regarding academic publishing, knowledge construction and dissemination, and the status of English as the academic lingua franca. The research data included 98 journal guidelines sections from seven subject areas, and results suggest that journals across the disciplinary spectrum endorse the primacy of (native speaker) English in knowledge dissemination, which calls for a critique of scientific monolingualism on the part of research policymakers, in Brazil and elsewhere. Keywords: Author guidelines; English-medium journals; Brazil; SciELO; Semiperiphery. Resumen Políticas linguísticas na semiperiferia: uma análise de diretrizes para autores em periódicos brasileiros que publicam em inglês Language policies from the semiperiphery: an analysis of author guidelines in Brazilian English-medium journals Gisele Dionísio da Silva Centre for English, Translation, and Anglo-Portuguese Studies, NOVA University of Lisbon (Portugal) giseledionisio@gmail.com 131 Ibérica 41 (2021): 131-154 GiSELE DiONíSiO DA SiLvA O aumento recente no número de periódicos que publicam exclusivamente em inglês em países não anglófonos alinha-se com a hegemonia do inglês no âmbito da comunicação acadêmica global, e as exigências de linguagem feitas por tais periódicos aos autores surgem como fonte de dados ao mesmo tempo valiosa e pouco explorada no que tange aos valores e às concepções que subjazem à produção e circulação de conhecimento. Pautado no conceito de “semiperiferia” para definir a atual posição do Brasil no contexto internacional de edição científica, este estudo exploratório analisa a seção “diretrizes para autores” de periódicos científicos brasileiros indexados no SciELO que publicam somente em inglês. Objetiva-se, com isso, mapear as exigências de linguagem atribuídas ao processo de submissão de manuscritos e avaliar, sob uma perspectiva discursiva, se estas reforçam ou contradizem ideologias dominantes concernentes à edição acadêmica, à construção e disseminação de conhecimento e ao estatuto do inglês como língua franca da ciência. Os dados incluem diretrizes de 98 periódicos vinculados a sete grandes áreas do conhecimento, e os resultados indicam uma tendência transdisciplinar de endossar a primazia do inglês (nativo) na circulação de conhecimento, tendência essa que demanda uma crítica do monolinguismo científico por parte de agentes decisórios e formuladores de políticas de incentivo à pesquisa, tanto no Brasil quanto em outros países. Palabras clave: Diretrizes para autores; Periódicos em inglês; Brasil; SciELO; Semiperiferia. 1. Introduction English as the lingua franca of global communication has taken over academic practices in the last few decades. Growing pressure to publish in this language now affects many scholarly requirements, such as those for assessing the quality of researchers’ work, granting tenure, and funding projects. The effects of this institutionalised pressure have had such an impact on scholars’ scientific output all over the world that English for Research Publication Purposes (ERPP) is now a thriving research field (Canagarajah, 2002; Curry & Lillis, 2004; Lillis & Curry, 2010; Lillis et al., 2010; Bennett, 2014a & 2014b). Researchers from non-Anglophone contexts, non-native English speakers (NNES) who produce scholarly work in their local language(s) and have an increasing need to disseminate their findings to a larger global audience (motivated both by personal satisfaction and professional constraints), often face considerable difficulties in getting their work published in so-called “international” journals. These, despite the apparent diversity evoked by the 132 qualifier “international”, usually amount to English-medium publications controlled by a restricted group of large commercial publishers such as Elsevier, Wiley-Blackwell, Springer, and Taylor & francis, which currently spearhead the highly profitable global market of scholarly publishing (these publishers accounted for over 50% of all papers published in 2013, according to a study by Larivière, haustein & mongeon, 2015). Therefore, according to Lillis and Curry (2010, p. 6), “together ‘English’ and ‘international’ constitute an important indexical cluster used to signify ‘high quality’”. The quality of such publications is measured by powerful indexes such as the Science Citation index (SCi), created in the 1960s by the institute for Scientific information (iSi)1 to rank the world’s most important journals. To achieve this, the iSi devised the impact factor (if), a calculation based on a two-year interval which involves dividing the number of times articles from a given journal were cited by the number of articles that are citable. The if sets in motion a cyclical trend in that it “both determines the inclusion of particular journals in indexes [and] helps contribute to the higher status of a journal” (Lillis & Curry, 2010, p. 15). Even though the if still constitutes a major (if not central) bibliometric criterion of journal quality worldwide (Lillis et al., 2010; Packer & meneghini, 2015), it has been subjected to widespread criticism, including the claims that it is insufficient for measuring journal quality and that it does not realistically reflect research being carried out in developing countries (Caramelli & Rocha e Silva, 2010). Packer and meneghini (2015, p. 20, my translation) endorse such a critique: “The general belief or perception that the impact factor is correlated with the quality of research published by a journal and, in turn, with the quality of the journal itself affects and seriously restricts the development of journals published at national level”. Be that as it may, journals continually strive to have high ifs and receive positive assessment from national governmental bodies as well as from citation indexes such as those of the Web of Science (WoS). These indexes follow a specific set of criteria to evaluate journals, namely basic publishing standards, editorial content, international focus, and citation analysis (Testa, 2016). Publishing standards cover aspects such as peer review, acknowledgements, ethical practices, publishing format, timeliness, international editorial conventions and, most importantly, full-text English. Testa (2016) underscores the importance of English for the international outreach of academic publishing: LANGuAGE POLiCiES fROm ThE SEmiPERiPhERy: AN ANALySiS Of AuThOR GuiDELiNES iN BRAziLiAN ENGLiSh-mEDium jOuRNALS Ibérica 41 (2021): 131-154 133 English is the universal language of science. for this reason our focus is on journals that publish full text in English or, at the very least, bibliographic information in English. There are many journals covered in Web of Science Core Collection that publish articles with bibliographic information in English and full text in another language. however, it is clear that the journals most important to the international research community are publishing full text in English. This is especially true in the natural sciences. There are notable exceptions to this rule in the arts & humanities and in social sciences topics. (...) Nonetheless, full text English is highly desirable, especially if the journal intends to serve an international community of researchers. The dominance of English within science is, therefore, a given to those responsible for establishing criteria to assess researchers’ work and productivity. Thus, to conform to such linguistic and editorial standards and to enhance their visibility within the scholarly community, journals from non-Anglophone countries have increasingly resorted to English as their sole medium of communication or, at best, as the preferred language alongside national or local ones. The incidence of such “English-medium journals” (Lillis & Curry, 2010; Lillis et al., 2010; Lillis, 2012; Petrić, 2014) – a phenomenon viewed by Petrić (2014) as relatively under-researched – now affects the whole spectrum of disciplinary areas, after an initial predominance in the hard sciences. in the Brazilian academic scenario, publishing in English is also on the rise (Packer, 2016). According to a report produced by Clarivate Analytics (2017) for the Coordination for the improvement of higher Education Personnel [Coordenação de Aperfeiçoamento Pessoal de Nível Superior – CAPES], a Brazilian federal government agency which oversees research funding, Brazil ranked 13th in the world as regards the number of peer-reviewed papers published between 2011 and 2016, and during the same period 87 Brazilian journals were included in the WoS indexes. The report does not make any explicit reference to English publications, but from Testa’s quotation above it can be inferred that these journals offer, if not full texts in English, at least bibliometric information in this language. however, it does acknowledge that the inclusion of Brazilian journals “shows the Web of Science editors’ recognition of the quality and international interest in Brazilian research” (Clarivate Analytics, 2017, p. 11). Therefore, complying to international standards – or to those established by the WoS – involves tackling the language issue directly. Brazilian journal editors have sought to do this by establishing normative and often non- GiSELE DiONíSiO DA SiLvA Ibérica 41 (2021): 131-154134 negotiable guidelines for authors wishing to publish research articles and other texts (e.g. book reviews, review articles, case reports, and short communications). The “author guidelines” section, known also as “instructions for authors” or “submission guidelines”, comes forth as a valuable yet often overlooked source of information regarding editors’ expectations on the linguistic, structural, and bibliographic aspects of manuscripts. As the point of entry of scholars’ access to publishing, its potential for discourse analysis and for research on academic writing has yet to be fully explored; indicative of this is the fact that the literature review has so far produced only three studies which focus specifically on author guidelines in journals related to various fields of knowledge (for Economics, see henshall, 2012, 2018; for Applied Linguistics, see viégas, 2016). This article examines author guidelines of Brazilian English-medium journals from across the disciplinary spectrum, with a view to mapping out requirements established by editors for the submission of manuscripts and, ultimately, to assessing their stance related to dominant ideologies regarding academic publishing, knowledge construction and dissemination, and the status of English as the academic lingua franca. Given that journal editors are major gatekeepers of scholarly publishing and have the power to establish which bodies of knowledge circulate and how, learning more about journals’ language policies is crucial to raising scholarly awareness of the way textual constraints may affect the wider context of the production and distribution of knowledge – even if policy and practice may (and often do) not overlap (henshall, 2018, p. 27). The remainder of this paper provides an overview of the theoretical framework adopted for this study, which draws on the notions of “semiperiphery” in academic writing, “discourse communities”, and “literacy brokers”; summarises the research data set and methodology; presents and discusses the main findings of the data analysis; and offers concluding remarks on the topic at hand. 2. The semiperiphery in academic writing Current debates on the politics and practices of academic writing often refer to scientific communities in terms of either the centre (or core) or the periphery, a terminology first proposed by world-systems theory in the 1970s (Wallerstein, 1984). According to Bennett (2014a, p. 157), although the LANGuAGE POLiCiES fROm ThE SEmiPERiPhERy: AN ANALySiS Of AuThOR GuiDELiNES iN BRAziLiAN ENGLiSh-mEDium jOuRNALS Ibérica 41 (2021): 131-154 135 theory was developed “to take account of long-term shifts in patterns of power and culture that could not adequately be explained on the level of the nation-state”, under a primarily economic focus, its major points and terminology have since been applied to various social, political, cultural, and educational discussions, among them that of academic writing. Since there appears to be a direct relationship between national wealth and academic and scientific achievement (Bennett, 2014a), scholars have widely resorted to dividing the world into centre – or, broadly speaking, “Western” – and peripheral countries. The former are contexts of intellectual excellence and considerable investment in education and research, while the latter tend to have low research output and insufficient material resources. however, such a division is continually subject to instability, given the fluid nature of both core and periphery – it is possible, hence, to find peripheries within the centre and centres within the periphery (Canagarajah, 2002, p. 42). The hegemonic role of English among the world’s languages, which Canagarajah (2002, p. 40) sees as motivated by “nonlinguistic reasons”, that is to say, by political and economic dominance, has a twofold contribution: it spreads centre values to the periphery and gives centre institutions access to peripheral ones, thus serving as their role models. hence, what scholars in peripheral countries actually “import” from the dominant academic cultures is not only linguistic and rhetorical norms, but also epistemological paradigms of knowledge construction. This further consolidates Anglophone research values and practices, in Canagarajah’s (2002, p. 43) view: (...) publishing according to the conventions and terms set by the center academic communities influences in no small way the representation of periphery knowledge. The publishing requirements, epistemological paradigms, and communicative conventions established by the center shape the knowledge that gets constructed through (...) journals. Such knowledge serves the interests of the center more than the periphery. The journals thereby disseminate partisan knowledge globally. Situated in-between core and periphery, both geographically and economically, are countries that combine characteristics of both groups – these have been assembled under the label of semiperiphery, proposed by Wallerstein (1984) within the world-systems theory and subject to multiple applications, including academic writing (e.g. collection of essays in Bennett, 2014b; Solovova, Santos & veríssimo, 2018). in terms of academic culture, GiSELE DiONíSiO DA SiLvA Ibérica 41 (2021): 131-154136 semiperipheral countries play a dual role in the sense that they emulate and reinforce values and practices from the centre while, at the same time, welcoming new elements into the system as a result of their contact with the periphery. in mediating knowledge channels between the centre and the periphery, the semiperiphery strives to establish a balance within its own academic milieu – on the one hand, “by emulating the centre and professing a commitment to its standards” (Bennett, 2014a, p. 168) and, on the other, “by giving a platform to alternative practices and non-standard forms of knowledge” (2014a, p. 169). from a linguistic standpoint, this process of assimilation from both “globality” and “locality” (Salager-meyer, 2015) may generate phenomena such as diglossia and language change (Bennett & muresan, 2016, p. 97), as well as further boost and legitimise the dominance of English as the lingua franca of scholarly output. As regards the scope of this study, the notion of semiperiphery seems applicable to the case of Brazil. While the country’s academic community, in some disciplines more than in others, increasingly orbits around practices and values from the Anglophone centre (what Lillis & Curry, 2010 name its centripetal pull), e.g. by the growing number of English-medium journals (Packer, 2016; Alcadipani, 2017), a closer collaboration with the WoS, and the establishment of partnerships with “world leaders in research, like Germany and the uSA” (Clarivate Analytics, 2017, p. 13), it also shows signs of resisting Anglophone dominance. Examples of this centrifugal force (Lillis, 2012, p. 716) include 1) an effective open-access (OA) policy for journals listed in bibliographic database SciELO (Scientific Electronic Library Online), created in Brazil in 1998 (addressed in more detail in the next section); 2) continuing publication in Portuguese in most journals from the humanities2 and 3) a growing “regional collaboration within Latin America” (Clarivate Analytics, 2017, p. 14). A segment of the Brazilian academic community has, in fact, promoted a movement of resistance to the hegemony of English and to the often unreflective endorsement of the Anglophone publish or perish logic; such a movement is voiced by Alcadipani (2017, p. 409, my translation), who strongly urges against upholding an “epistemic colonialism” that assigns Brazilian scholarly output a subaltern role on the world stage: if national English-medium journals or those now resorting to English are rarely cited by the “global” audience, seldom integrate rankings of “global” journals, and are rarely ever cited outside Brazil, there is little reason why a researcher who has written an article in English that he/she considers to be LANGuAGE POLiCiES fROm ThE SEmiPERiPhERy: AN ANALySiS Of AuThOR GuiDELiNES iN BRAziLiAN ENGLiSh-mEDium jOuRNALS Ibérica 41 (2021): 131-154 137 “good” should submit it to an English-medium Brazilian journal. Publishing in such journals will not help his/her scholarly output to become “global” or the author to participate in a “global” academic debate. Alcadipani concludes that “publishing in Portuguese” (2017, p. 410) may constitute a means of resisting the centripetal pull of Anglophone academic practices. it must be pointed out, however, that the author directs his criticism at national journals that publish in English, rather than at scholars who seek to publish in English in international journals. The increase in English-medium journals being published in Brazil is viewed in a very different light by Packer (2016), co-founder of SciELO, who regards it as “one of the most outstanding developments” in Brazil’s process of research dissemination over the past few years, in addition to being one of three of SciELO’s stages of internationalisation (the other two being minimum quotas for foreign researchers in editorial boards and for foreign authors in journals’ overall content). These radically opposing views are indicative not only of the contentiousness surrounding the issue of publishing (solely) in English, but also of the ambivalent – hence, semiperipheral – status of Brazil within global scholarly practices, in general, and publishing trends, in particular. 3. Discourse communities in academic publishing in recent decades, the term “community” has been associated with qualifiers like “discourse”, “practice”, and “speech” in attempts to describe and explain how a given social group both construes and is construed by language (Canagarajah, 2002). Given that this study seeks to highlight discourses connected with the Brazilian publishing scene, as well as the way specific communicative goals are conveyed through textual instructions to authors (instructions that amount to an academic textual genre in its own right), it draws on the concept of “discourse community”, defined by Swales (1990, p. 24) as “a group of people who link up in order to pursue objectives that are prior to those of socialization and solidarity”. inspired by the earlier and highly influential concept of “speech community” from sociolinguistics, Swales sought to understand the internal dynamics of written scholarly communication (as opposed to the oral dimension of the parent concept) by contending that genres are “communicative vehicles for the achievement of goals” (1990, p. 46) which underscore and define all GiSELE DiONíSiO DA SiLvA Ibérica 41 (2021): 131-154138 operations of a discourse community. Endorsement of genres is but one of six characteristics proposed by Swales to identify a group of individuals as a discourse community, in addition to it subscribing to a set of common public goals; having mechanisms of intercommunication and effective participation among its members; acquiring specialised terminology; and establishing a threshold level between novices and experts (Swales, 1990, pp. 24-27). for the purposes of this study, the mechanisms of intercommunication and the possession of genres and technical terminology are particularly useful to map out the ways in which the Brazilian academic community establishes communicative norms via journals’ author guidelines section. Academic publishing, as a specific scholarly practice, is a domain that oversees a series of conflicts and negotiations between members of different (sub-)communities. To examine the author guidelines section of Brazilian English-medium journals and to shed light on the underlying set of values and ideologies that dictate what is to be published and how, this study also endorses the distinction between communicative and social conventions proposed by Canagarajah (2002). The former is further divided into “textual” and “publishing” conventions, with textual conventions concerning “matters of language, style, tone, and structure that characterize academic texts” (Canagarajah, 2002, p. 6), and publishing conventions encompassing aspects such as “the protocol for submitting papers, revisions, and proofs; the nature of the interaction between authors and editorial committees [and] bibliographical and documentation conventions” (2002, p. 6). Social conventions are, in turn, “the rituals, regulations, and relationships governing the interaction of members of the academic community” (2002, p. 6). Such a distinction is useful to the extent that it relates textual norms and constraints to the larger social practices at work within academic writing. in addition, this study draws on Lillis and Curry’s (2010) useful classification of academic communities that multilingual researchers are writing for. The authors envision seven potential communities which orbit around “disciplinary, geolinguistic and applied versus theoretical [dimensions]” (Lillis & Curry, 2010, p. 42): 1) national academic community in the local national language; 2) national applied community in the local national language; 3) national academic community in English medium; 4) “international” academic community in the local national language; 5) intranational academic community in English medium; 6) other national academic community in national languages; 7) “international” academic LANGuAGE POLiCiES fROm ThE SEmiPERiPhERy: AN ANALySiS Of AuThOR GuiDELiNES iN BRAziLiAN ENGLiSh-mEDium jOuRNALS Ibérica 41 (2021): 131-154 139 community in English. The scope of this study immediately relates to the third type of community but, as will be seen in the discussion of results, some of the journal guidelines in the data set are clearly aimed at achieving the status of the seventh type of community. 4. Literacy brokers Academic publishing involves not only those who produce research and later report their findings to the scientific community, i.e. the actual authors of texts, but a number of others (individually or collectively) who intervene in varying forms and degrees on the shaping of manuscripts, both in terms of form and content. These interventions are often given broad and straightforward labels, such as proofreading, reviewing, editing, and translating, but their actual nature and effects on academic writing have as yet been little researched (Lillis & Curry, 2010). Given that publishing, whether done for academic or commercial purposes, is a highly complex and multilayered process that involves collaboration (and often conflict and negotiation) between a number of professionals, the role they play cannot be underestimated. hence, this study draws on the notion of “literacy brokering” proposed by Lillis and Curry (2010, p. 88), a meta-category that encompasses “all the different kinds of direct intervention by different people, other than named authors, in the production of texts”. The use of the term “brokering” stems from the need to acknowledge “that intervention in academic texts is not a neutral enterprise, but rather involves participants of unequal status and power” (2010, p. 88). Literacy brokers include two main categories: “language” and “academic” brokers. Language brokers are asked to intervene on texts due to their expertise in language issues – these include, on the one hand, professional brokers such as translators, proofreaders, and language specialists in general, and, on the other, informal brokers such as friends or family members. Academic brokers are, in turn, experts who work in universities or research facilities and have general, disciplinary or sub-disciplinary knowledge regarding the text they are called upon to correct or review. 5. Research data and methodology The research data set comprises author guidelines from all 98 Brazilian English-medium journals indexed by SciELO, a digital library for OA GiSELE DiONíSiO DA SiLvA Ibérica 41 (2021): 131-154140 journals that was created in Brazil in 1998 (for the history of SciELO, see Packer et al., 2014). Nowadays it houses journals from 16 countries from Latin America and Europe, as well as from South Africa, and its extensive list of contributors helps to put Brazil among the top-ranking countries publishing OA journals worldwide – according to Lillis and Curry (2010), the country had the world’s second highest number of OA journals in 2009. The decision to investigate language policies in Brazilian journals included in this particular research database was grounded on this study’s contention that Brazil enjoys a semiperipheral standing as regards its scientific output – a standing which places it both at the borderlines of centres of research excellence and at the core of a research network encompassing scholars from Latin America and other parts of the world. SciELO, a platform initially created in Brazil and still (proportionately) spearheaded by Brazilian academics3, despite its multinational composition, comes forth as a suitable source of data for this purpose. The 98 journals that publish only in English account for one third of all Brazilian journals currently indexed by SciELO (296); those that publish only in Portuguese or in Portuguese and other languages (e.g. Spanish) were excluded from the data set. Of the journals, 32.65% presented author guidelines in English only, whereas 67.35% exhibited instructions in both Portuguese and English. in addition, 60.20% of journals required authors to submit manuscripts with titles, abstracts, and descriptors in English only, whereas 39.79% asked authors to submit these in two or even three languages (English and Portuguese or, in some cases, English, Portuguese, and Spanish). The journals surveyed cover seven subject areas, according to the classification put forward by SciELO, itself based on the classification of broad disciplinary areas devised by Brazil’s National Council for Scientific and Technological Development [Conselho Nacional de Desenvolvimento Científico e Tecnológico – CNPq], the governmental body responsible for overseeing research and development. Subject areas are as follows: Agrarian Sciences, Applied Social Sciences, Biological Sciences, Engineering, Exact/Earth Sciences, health Sciences, and human Sciences. An eighth subject area listed in the index, that of Linguistics, Literature and Arts, does not have Brazilian English-medium journals indexed in SciELO and was, therefore, excluded from the study. journals’ distribution according to subject area is shown in figure 1. LANGuAGE POLiCiES fROm ThE SEmiPERiPhERy: AN ANALySiS Of AuThOR GuiDELiNES iN BRAziLiAN ENGLiSh-mEDium jOuRNALS Ibérica 41 (2021): 131-154 141 As regards sample size, the share of English-medium journals in relation to the total number of journals per area is as follows: Agrarian Sciences – 71.8%; Biological Sciences – 59.3%; Engineering – 45%; Exact/Earth Sciences – 40%; health Sciences – 35.1%; human Sciences – 7.8%; Applied Social Sciences – 2.4%. Such distribution shows a predominance of English- medium outlets within the natural and hard sciences (a contention also made by Petrić, 2014, p. 194 for English-medium journals in Serbia). Ten journals (10.2%) were listed in more than one subject area due to their multidisciplinary nature, but in this study were assigned only to the first field of knowledge listed in SciELO. 5.1. Analytical parameters Parameters for analysing author guidelines sections converged, to a certain extent, with those proposed by viégas (2016) and henshall (2012, 2018) to categorise journals’ language requirements. viégas’ (2016, p. 65, my translation) parameters are “language”, “language variety (for spelling)”, and “third-party revision”, whereas henshall’s (2018, pp. 29-30) are variety of English, grammatical correctness, and readership constraints. in view of the nature of this study’s data set, a decision was made to adapt and add analytical categories. firstly, since all 98 journals analysed publish only in English, viégas’ first criterion was not applicable. Secondly, in line with this study’s terminological orientation, “third-party revision” was renamed “literacy brokers”, a broader category which includes other forms of language mediation such as translation and copy-editing. Therefore, four major parameters inductively emerged from the data analysis: English GiSELE DiONíSiO DA SiLvA Ibérica 41 (2021): 131-154142 Journals’ distribution according to subject area is shown in Figure 1. Figure 1. Distribution of Brazilian English-medium journals indexed by SciELO, according to subject area. A language variety, language style, literacy brokers, and publishing conventions. The first parameter, English variety, refers to journal guidelines’ indication of varieties of English to be used by NNES authors when submitting their texts. in the data set, specifications included American English, British English, American or British English, and English as a general term. journals that failed to specify a variety of English but conditioned the publication of manuscripts to language revision effected by editing companies based in English-speaking countries (such as the united States) were pooled under the corresponding varieties. The second parameter, language style, concerns both micro-level aspects related to accuracy in grammar and vocabulary, as well as macro-level, rhetorical aspects (in line with Canagarajah’s communicative textual conventions). The third parameter, literacy brokers, pertains to the language professionals involved with editing, copy-editing, proofreading or translating texts prior to publication by journals. Lastly, the fourth parameter, publishing conventions, covers journal guidelines’ specifications of citation and reference styles – which, in some cases, may imply editors’ preference for a particular language variety (a point already made by henshall, 2018, p. 29) – as well as of texts’ general structure and layout. The following section presents and discusses the main results of the data analysis. The data comprised textual information regarding instructions for authors made available on journals’ websites in SciELO and, occasionally, on external websites (for journals published by editorial conglomerates such as Elsevier and Springer). Explicit references to parameters 1 (English variety) and 3 (literacy brokers) were tallied under the categories that emerged from the data set: for parameter 1, “American English”, “British English”, “American or British English”, and “English”; for parameter 3, “Proofreading by a NES (native speaker of English)”, “Proofreading by a NES who is also an expert in the field”, “Professional revision/editing”, and “Translation”. Textual information pertaining to parameters 2 (language style) and 4 (publishing conventions) were pooled under the respective categories. Direct quotations from journals’ author guidelines are indicated by the initial letter of their respective subject area (Agrarian Sciences – A; Applied Social Sciences – AS; Biological Sciences – B; Engineering – E; Exact/Earth Sciences – E/E; health Sciences – h; human Sciences – hS) followed by a number in brackets. Quotations from guidelines originally presented in Portuguese on journals’ websites have been translated into English for the purposes of this article. LANGuAGE POLiCiES fROm ThE SEmiPERiPhERy: AN ANALySiS Of AuThOR GuiDELiNES iN BRAziLiAN ENGLiSh-mEDium jOuRNALS Ibérica 41 (2021): 131-154 143 6. Results 6.1. English language variety The analysis of journals’ language requirements shows that a vast majority of them (69.38%) did not specify language variety, opting instead for English as an umbrella term. This finding may be explained by a lack of perception on the part of NNES (or, indeed, of those of any language) regarding specific language varieties, which leads them to focus on the function of communication rather than on linguistic and cultural identity (henshall, 2018, p. 31). Across the disciplinary sample surveyed in this study, the health Sciences yielded the largest share (39.71%), followed by Agrarian Sciences (26.47%) and by Biological Sciences and Engineering (11.76% each), which suggests a general lack of awareness of (or of concern for) language variation in these areas of knowledge. journals that endorsed the use of American English (or simply American spelling), both explicitly and implicitly (i.e. through the recommendation of language services provided by uS-based companies), accounted for 23.47% of the total. While such a preference is not surprising, given the considerable influence of American culture among Brazilians in general4, it reveals Brazilian editors’ continuing endorsement of national standards of English amid growing support for the English as a Lingua franca (ELf) movement (jenkins, Baker & Dewey, 2018). four journals (4.08%) pointed out that “either North American or British English is admitted, but not a mix of both registers” (hS5), hence either expressing concern with language standardisation or emulating international journals. yet another journal acknowledged that, “[a]lthough we encourage submissions from around the globe, we require that manuscripts be submitted in American English” (B6), an almost apologetic statement specifically addressed to NES authors from countries other than the united States and, to a lesser extent, to NNES authors who use other varieties. As previously mentioned, the effective resonance of American academic practices in Brazilian English-medium journals can also be felt implicitly, through the common endorsement of uS-based language and editing companies. Seventeen of the 23 journals (73.91%) pooled under the American English category gave such endorsement, mostly to the same company, as will be addressed in the subsection on literacy brokers. Lastly, 3.06% of journals did not provide any language requirements, a finding that signals an imbalance between language policy and actual practice GiSELE DiONíSiO DA SiLvA Ibérica 41 (2021): 131-154144 (henshall, 2018) – since all texts published by them are, invariably, in English – and, ultimately, a taken-for-granted attitude on the part of editors regarding English as the sole conduit for disseminating scholarly knowledge. 6.2. Language style journals generally employed broad and vague language style requirements, e.g. “clear and concise” (h8), “good scientific English” (B3), “correct scientific English” (h26 – my translation), “quality version” (A2, A3, A14, A23 – my translation), “acceptable English usage” (h18), “[a]ppropriate use of the English language” (h18). No further benchmarks are given to establish what counts as “good”, “correct” or “acceptable” written English, and further ethnographic research with the gatekeepers involved would be required to assess whether there are disciplinary differences at play in such requirements in the Brazilian context (a methodological choice adopted by hynninen and Kuteeva, 2017 in their analysis of research writing standards in finland and Sweden). moreover, the most repeated qualification for manuscripts was “conciseness”, with the noun and its adjective “concise” having been mentioned by 4.08% of journals each; its Portuguese equivalents, “conciso”/“concisa”, were mentioned by 15.15% of journals which presented author guidelines in this language. in addition, some journals made specific reference to micro-level aspects of grammar, vocabulary, and rhetoric. for instance, one publication requested that authors “[o]pt for short sentences in direct and active form” (h6 – my translation), in line with Anglophone journals’ current preference for active over passive voice5. Another journal alerted authors not to use “clichés, slang, and colloquial words or expressions such as ‘in the present study’” (B11) – the latter a literal translation of an expression commonly used by Portuguese speakers in academic writing, which indicates that editors largely envision this to be the main target audience in terms of authorship and expect authors to “tone down” their Portuguese textual rhetoric (Bennett & muresan, 2016). yet another requested that authors “[n]ever use abbreviations that spell common English words, such as fuN, PiN, SCORE, and SuN” (h13), a curious admonition which, while motivated by a fear of having research articles appear too informal, lacks pragmatic validity. English numerical conventions were also stressed by some of the journals – for instance, 3.06% requested that authors should express decimal values by employing “dots instead of commas” (E/E4). Such prescriptive recommendations corroborate hynninen and Kuteeva’s (2017, p. 63) contention that L2 users LANGuAGE POLiCiES fROm ThE SEmiPERiPhERy: AN ANALySiS Of AuThOR GuiDELiNES iN BRAziLiAN ENGLiSh-mEDium jOuRNALS Ibérica 41 (2021): 131-154 145 of English acting as gatekeepers may have conservative views towards standard English norms and thus reject variation. however, these recommendations may also be interpreted as didactic advice on the part of editors that seek to foster efficient writing practices in English by NNES authors; this view would help assign an educational role to semiperipheral English-medium journals, aimed at assisting “local authors develop skills needed for publishing in English” (Petrić, 2014, p. 201). Lastly, 7.14% of journals specifically asked authors to provide objective and concise titles for their articles and, above all, to avoid expressions like “Effects of ”, “influence of ”, “Study on”, “Considerations on”, etc., deemed as “irrelevant” (A24 – my translation) and “unnecessary” (E7 – my translation). This need to avoid words and expressions regarded as too broad for a scientific report corroborates Lillis and Curry’s (2010) claim that highly specialised journals, such as those concentrated in the Anglophone world, tend to be very restricted in scope, “where tiny yet significant contributions to specific knowledge areas can be noted and evaluated” (2010, p. 128). from this perspective, pinpoint contributions to scientific knowledge require an equally precise language in order to be effectively conveyed. This kind of norm, subsumed by a perception of English as the default language of science, often contrasts with other rhetorical traditions such as that of Portuguese and other Romance languages, which favours verbal expansiveness and elaboration (Bennett & muresan, 2016, pp. 101-114). This rhetorical contrast raises a number of problems for the translation of academic texts from Portuguese into English (Bennett, 2011). 6.3. Literacy brokers The analysis shows that literacy brokers operating in the Brazilian academic publishing scene may be individual or collective, as well as relate to both language and content expertise. in the data set, 20.40% of journals stated that manuscripts should be proofread by a NES, whether one of the authors or not. Of this sub-group, 45% of author guidelines stressed the twofold need for native speakers with scientific expertise, usually a colleague; in line with the terminology proposed by Lillis and Curry (2010), he/she would amount to an academic broker, but in the present case combines both language and academic brokering. moreover, 37.75% of journals expressed the need for manuscripts to be submitted to professional revision or editing6, whereas 22.44% GiSELE DiONíSiO DA SiLvA Ibérica 41 (2021): 131-154146 recommended manuscripts to be translated. These requirements entailed, in most cases, two others: 1) that authors must present a certificate validating the revision or translation of manuscripts (32.20% of the 59 journals which endorsed revision and/or translation), and 2) that such a certificate be issued by language editing companies suggested by journals themselves (33.89%). The data show that certain companies are endorsed by various journals from different subject areas, which suggests that journal editors tend to recommend brokers whose work they are familiar with to colleagues in charge of other journals. One company in particular stood out among the rest, having been endorsed by 32.20% of journals: American journal Experts (AjE), a uS-based enterprise that offers proofreading, translation, and editing services to academics worldwide. Given that its in-house proofreaders and translators are, according to its own website7, NES and PhD holders from highly specialised fields, the company combines language and academic brokering and, as a result, exerts considerable influence over both the form and content of English-medium research articles. The company’s outreach helps to reinforce the legitimacy and hegemony of American English not only in Brazil, but also throughout Latin America. As for translation, another major brokering activity in these journals’ manuscript submission process, it operates mostly under the following conditions: authors may submit manuscripts in Portuguese (or, occasionally, in Spanish) but, in case these are accepted for publication, they must provide the English translation at their own expense by hiring professional translators or translation agencies. One journal even suggested that authors should “seek financial aid for translation in their own institutions, especially in postgraduate departments” (h7 – my translation), which confirms the widespread notion that high-quality translation services tend to be costly (Lillis & Curry, 2010, p. 112). Of the journals which explicitly referred to translation, however, 18.18% do not outsource translation, but consider it as an in-house task to be performed by their own editorial team; by requiring authors to provide Portuguese titles, abstracts, and descriptors, such journals offer to translate these into English for researchers with no knowledge of Portuguese: “A Portuguese language title, abstract and keywords will also be published. if you are fluent in Portuguese, please provide this as indicated below. if not, we will translate your original text into Portuguese” (h23). Other references to translation concerned a macro-linguistic level in terms of literality – “Provide a concise (...) and accurate Portuguese or Spanish translation of the English abstract” (B11); “Abstract: up to 250 words, it LANGuAGE POLiCiES fROm ThE SEmiPERiPhERy: AN ANALySiS Of AuThOR GuiDELiNES iN BRAziLiAN ENGLiSh-mEDium jOuRNALS Ibérica 41 (2021): 131-154 147 must be a faithful translation of the Portuguese text” (A21 – my translation) – as well as a micro-linguistic level, as regards specific lexical categories: “Names of sponsor institutions should be listed in their original spelling and not translated to English” (B11); “Geographic descriptors must NOT be translated” (B11). Across the disciplinary sample included in this study, the human Sciences had the largest proportionate share of journals requesting manuscripts’ revision or translation or both (71.42%) at the hands of literacy brokers, followed by Agrarian Sciences (64.28%) and Biological Sciences (56.25%). This finding suggests that viewing revision or proofreading as a means to achieve language correctness is not an exclusive concern of scholars from the humanities, encompassing also those active in other subject areas (as observed by hynninen and Kuteeva, 2017 with regard to historians and computer scientists). Three journals, one from each of these subject areas, are published by international leaders in the field (two by Springer and one by Elsevier), and their instructions to authors hence follow these publishers’ standards. The large majority of these journals, though, are published by university or independent presses in Brazil, which leaves open to contention the role of publishers in establishing journals’ language policies – a role highlighted by henshall (2018, p. 32) as being central to policies on grammatical correctness. This issue lies outside the scope of this article, but merits future investigation. 6.4. Publishing conventions Publishing conventions in the data set varied extensively among journals, especially with regard to citation and reference styles, but there was some common ground concerning the need to avoid or altogether prevent the listing of references in languages other than English (a practice highlighted by Lillis et al., 2010). in addition to one journal specifically requesting non- inclusion of references to works originally written and published in Portuguese, three others instructed authors to list references in English whenever possible: in case references in other languages were considered necessary, one of them requested that “the title should be translated and presented in English, between brackets” (A6), whereas another made the following concession: “up to four references written in languages other than English will be allowed without the need for justification” (A28 – my translation). Restricting authors’ referencing scope to English-medium publications is clearly tributary to editors’ wish to increase journals’ if and GiSELE DiONíSiO DA SiLvA Ibérica 41 (2021): 131-154148 hence boost their international standing, but the (not so) long-term implications of silencing research produced in local languages, notwithstanding its quality, must be addressed by academia and governmental bodies. A further indication of the hegemonic status of English in these publications may be given by their visual layout. Three journals requested that titles, abstracts, and descriptors in English must come before their equivalents in other languages. One of them stated: “The title in Portuguese or Spanish must be placed below the title in English. The Resumo or Resumen should be placed after the Abstract and Keywords” (h8). Precedence – both literal and metaphorical – of English in academic writing is, once again, motivated by practical bibliometric factors, e.g. faster and more accurate search for research articles and effective citation, but also conforms to an ideology that seeks to legitimise English as the only credible conduit for producing and disseminating knowledge worldwide. 7. Conclusions The status of English as the language of global communication and, as far as academia is concerned, as the medium through which scholars hope to boost the visibility and influence of their research has complex and far- reaching implications; one of the most detrimental is the gradual silencing of knowledge produced and reported in other languages. What the widespread adoption of linguistic and rhetorical norms, as well as of other academic practices from the Anglophone context, actually implies is that it is still up to the centre to settle “what counts as relevant knowledge and who has the right to determine what counts as relevant knowledge” (Lillis & Curry, 2010, p. 161). in some of the Brazilian journals analysed, editors’ wish to increase if and academic visibility has even led them to restrict the number of references written in languages other than English – a clear indication that, albeit for bibliometric purposes, quality scientific output by NNES authors needs to be monolingual, particularly in the natural and health sciences. far from subscribing to a rejection of English-medium publishing (in Brazil and elsewhere) and to a movement of linguistic, rhetorical, and epistemological resistance, this study sought to shed light on a largely unexplored source of data regarding the editorial voice behind academic LANGuAGE POLiCiES fROm ThE SEmiPERiPhERy: AN ANALySiS Of AuThOR GuiDELiNES iN BRAziLiAN ENGLiSh-mEDium jOuRNALS Ibérica 41 (2021): 131-154 149 journals. Since editors are the major gatekeepers of publishing and have the power to establish which bodies of knowledge circulate and how, learning more about journals’ language policies is an important step in raising awareness about the way textual and rhetorical constraints may affect the production and distribution of scientific knowledge. This study’s findings have shown, by recourse to the four major parameters outlined (English variety, language style, literacy brokers, and publishing conventions), that Brazilian English-medium journals follow a diverse array of textual and rhetorical norms, despite some common ground between them. for instance, they tend to refer to English as a homogenous, global language, rather than specifying language varieties – a finding which suggests that Brazilian editors generally favour the promotion of reader fluency more than the association with a given linguistic and cultural tradition. moreover, they tend to view the role of literacy brokers as beneficial to the publication process, though this role is largely expected to be certified, either by language editing companies or by native speakers with expertise in academic writing and/or scientific research. As regards language style and publishing conventions, the considerable diversity in language requirements observed in the data prevents any kind of systematic quantification; this suggests that such instructions to authors, rather than being simply “copied” from older or more prestigious journals, may reflect editors’ personal notions of language standardisation and correctness. in general, there is a clear effort on the part of editors to offer a platform for Brazilian researchers to gain more visibility on a global scale, since many instructions seem to address issues that are familiar to Portuguese speakers (marlow, 2014). Results, therefore, show that these English-medium journals reinforce, through various strategies, traditional notions of academic publishing, knowledge construction and dissemination, and, above all, the status of English as the academic lingua franca. however, there is a seeming paradox in their efforts to achieve international recognition, in view of the high incidence of journal titles (38.77%) that include the nationality adjective “Brazilian”; one journal fully embraces such paradox by being called The international Brazilian journal of urology. This apparent contradiction – coupled with the fact that the majority provide author guidelines in Portuguese and occasionally in Spanish, as well as in English – has the potential to confirm Brazil’s semiperiphery standing within the global scholarly publishing scene. This standing, while retaining a “national” reference through journal titles (even if for purely historical reasons), also GiSELE DiONíSiO DA SiLvA Ibérica 41 (2021): 131-154150 strives to meet international standards for scientific production – exclusive publication in English among them. in summary, despite growing calls for Brazilian scholars to publish in English and, in so doing, to improve their professional assessment and boost career prospects, the fact that English-medium journals account only for one third of SciELO’s total number of journals (with the others still publishing in Portuguese or in Portuguese and English) indicate that the Brazilian academic community in general continues to resist the “centripetal pull” of the Anglophone centre, at least in the short term. As a true semiperiphery location, Brazil continually aims to reach a balance between maintaining its locality and aspiring to universality, between producing (as well as applying) research for the local population while enhancing its competitiveness abroad. it is important that all those involved with drafting educational and research policies at national level, while coming up with solutions to boost scholarly productivity, also critically assess the long-term impact of scientific monolingualism on Brazil’s academic culture. Such a debate is naturally extensive to other geolinguistic contexts and discourse communities, and future research on journals’ language policies, whether from an intra-, multi-, or transdisciplinary perspective, should help scholars assess their collective practices and foster more equitable mechanisms of knowledge circulation. Article history: Received 15 September 2019 Received in revised form 12 April 2021 Accepted 05 May 2021 References LANGuAGE POLiCiES fROm ThE SEmiPERiPhERy: AN ANALySiS Of AuThOR GuiDELiNES iN BRAziLiAN ENGLiSh-mEDium jOuRNALS Ibérica 41 (2021): 131-154 151 Alcadipani, R. (2017). Periódicos brasileiros em inglês: a mímica do publish or perish ‘global’. Revista de Administração de Empresas, 57(4), 405–411. Bennett, K. (2011). The scientific revolution and its repercussions on the translation of technical discourse. The Translator, 17(2), 189–210. Bennett, K. (2014a). The ‘butler’ syndrome: Academic culture on the semiperiphery. Revista Canaria de Estudios Ingleses, 69, 155–171. Bennett, K. (Ed.) (2014b). The semiperiphery of academic writing: Discourses, communities and practices. Palgrave Macmillan. Bennett, K. & Muresan, L.-M. (2016). Rhetorical incompatibilities in academic writing: English versus the Romance languages. Synergy, 12(1), 95–119. British Council (2015). English in Brazil: An examination of policy, perceptions and influencing factors. Retrieved 23 August 2019 from https://ei.britishcouncil.org/sites/default/files/latina mericare-search/English%20in%20Brazil.pdf Canagarajah, A. S. (2002). A geopolitics of academic writing. University of Pittsburgh Press. Caramelli, B. & Rocha e Silva, M. (2010). Brazilian medical journals are now facing a big challenge. European Science Editing, 36(2), 38–39. Clarivate Analytics (2017). Research in Brazil. Retrieved 23 August 2019 from https:// www.capes.gov.br/images/stories/download/divers Gisele Dionísio da Silva currently pursues a PhD degree in Translation Studies at the NOvA university of Lisbon, Portugal. her research interests include academic writing, university publishing, specialised translation, and relations between translators and other agents in the publishing industry. She is the author of “O Corvo” no Brasil: a autoria do tradutor (Editora ufG, 2006). GiSELE DiONíSiO DA SiLvA Ibérica 41 (2021): 131-154152 os/17012018-CAPES-InCitesReport-Final.pdf Curry, M. J., & Lillis, T. (2004). Multilingual scholars and the imperative to publish in English: Negotiating interests, demands, and rewards. TESOL Quarterly, 38(4), 663–688. Henshall, A. C. (2018). English language policies in scientific journals: Signs of change in the field of economics. Journal of English for Academic Purposes, 36, 26–36. Henshall, A. C. (2012). English as an international language and language policies in economics journals. Anglo Saxonica, III(4), 131–164. Hynninen, N. & Kuteeva, M. (2017). ‘Good’ and ‘acceptable’ English in L2 research writing: Ideals and realities in history and computer science. Journal of English for Academic Purposes, 30, 53–65. Jenkins, J., W. Baker, & M. Dewey (Eds.) (2018). The Routledge handbook of English as a lingua franca. Routledge. Larivière, V., Haustein, S., & Mongeon, P. (2015). The oligopoly of academic publishers in the digital era. PLoS ONE, 10(6), e0127502. Lillis, T. (2012). Economies of signs in writing for academic publication: The case of English medium ‘national’ journals. Journal of Advanced Composition, 32, 695–722. Lillis, T., & Curry, M. J. (2010). Academic writing in a global context: The politics and practices of publishing in English. Routledge. Lillis, T., Hewings, A., Vladimirou, D., & Curry, M. J. (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(1), 111–135. Marlow, M. A. (2014). Writing scientific articles like a native English speaker: Top ten tips for Portuguese speakers. Clinics, 69(3), 153–157. Packer, A. L. (2016). Cresce a adoção do inglês entre os periódicos SciELO no Brasil. Retrieved 23 August 2019 from https://blog.scielo.org/blog/ 2016/05/10/cresce-a-adocao-do-ingles-entre-osper iodicos-scielo-do-brasil/#.XWPgzi3Ool4 Packer, A. L., & Meneghini, R. (2014). Capítulo 1 –O SciELO aos 15 anos: raison d’être, avanços e desafios para o futuro. In A. L. Packer et al. (Eds.), SciELO – 15 anos de acesso aberto (um estudo analítico sobre acesso aberto e comunicação científica), (pp.15-28). UNESCO. Packer, A. L., N. Cop, A. Luccisano, A. Ramalho, & E. Spinak (Eds.) (2014). SciELO – 15 anos de acesso aberto (um estudo analítico sobre acesso aberto e comunicação científica). UNESCO. Petrić, B. (2014). English-medium journals in Serbia: Editors’ perspectives. In K. Bennett (Ed.), The semiperiphery of academic writing: Discourses, communities and practices (pp.189- 209). Palgrave Macmillan. Salager-Meyer, F. (2015). Peripheral scholarly journals: From locality to globality. Ibérica, 30, 15–36. Solovova, O., Santos J. V., & Veríssimo, J. (2018). Publish in English or perish in Portuguese: Struggles and constraints on the semiperiphery. Publications, 6(25), 1–14. Swales, J. (1990). Genre analysis: English in academic and research settings. Cambridge University Press. Testa, J. (2016). Journal selection process. Clarivate Analytics, 18 July. Retrieved 23 August 2019 from https://clarivate.com/essays/ journalselection-process Viégas, M. R. (2016). O inglês como língua franca e a publicação acadêmica: uma análise de diretrizes para autores de periódicos internacionais [English as a lingua franca and academic publishing: An analysis of author guidelines of international journals]. Master’s dissertation. Federal University of Rio Grande do Sul. Wallerstein, I. (1984). The politics of the world- economy: The states, the movements and the civilizations. Cambridge University Press. Notes 1 The iSi created the Web of Knowledge (nowadays known as Web of Science), an online platform for multidisciplinary scientific citation indexing. The Web is now maintained by Clarivate Analytics. 2 Retrieved 23 August 2019 from http://www.scielo.br/scielo.php?lng=en 3 According to SciELO Analytics, Brazil has a clear lead in the ranking of articles’ affiliating countries within SciELO, accounting for 65.87% of the total number of publications. Retrieved 23 August 2019 from https://analytics.scielo.org/w/publication/article?collection=scl#about-the-chart-5 4 According to the British Council (2015, p. 55): “The youth of Brazil are mostly exposed to the American way of life through popular culture. As a result, many of the younger students will enrol in schools that teach American English and are oriented towards American English proficiency tests, like the TOEfL (currently over half of the English taught in Brazil is American English)”. 5 According to Bennett (2011, p. 194), the “so-called impersonal active” involves using the active voice when the grammatical subject is an object, thing, or material. 6 One journal suggested reading materials on the issue (“Guidance on grammar, punctuation, and scientific writing” – B6), among them an article (marlow, 2014) which gives “tips” for native Portuguese speakers to write like a NES. 7 Retrieved 23 August 2019 from https://www.aje.com/about/our-team LANGuAGE POLiCiES fROm ThE SEmiPERiPhERy: AN ANALySiS Of AuThOR GuiDELiNES iN BRAziLiAN ENGLiSh-mEDium jOuRNALS Ibérica 41 (2021): 131-154 153