Lithuanian academic discourse revisited: Features and patterns of scienti ﬁ c communication

: Over the past several decades, there has been an increasing interest in academic discourse investigations with a speci ﬁ c focus on disciplinary, cultural, and generic aspects of academic text con - struction. Studies of Spanish, Italian, Greek, Portuguese, French, German, and Russian ( inter alia ) aca - demic discourse have revealed not only the universal features characteristic of many writing cultures, but also unique rhetorical features, typical of only some of them. In this article, we focus on academic discourse investigations in scienti ﬁ c texts written in one of the lesser - studied languages, Lithuanian. We review here studies of Lithuanian scienti ﬁ c discourse and what those studies reveal about the speci ﬁ c features of Lithuanian academic discourse. The ﬁ rst part of the article identi ﬁ es how rhetorical structures di ﬀ er between English and Lithuanian academic texts, whereas the second part presents a comprehensive over - view of the functional groups of linguistic units, such as metadiscourse markers, stance and engagement markers, and conceptual metaphors, and how these are di ﬀ erent or similar to English academic discourse. The overview of the empirical studies shows the current state of knowledge about Lithuanian academic discourse and highlights some of the items that distinguish Lithuanian writing. Furthermore, we highlight the importance of identifying academic writing traditions of other languages to acknowledge diversity.


Introduction
Over the past several decades, research about academic text construction of languages other than English has received more attention through discourse investigation focusing on disciplinary, cultural, and genrespecific differences (Schmied 2011, Chitez et al. 2018). Many studies have convincingly argued for the existence of "disciplinary culture" within which specific fields of science and individual disciplines follow their distinct traditions of argument weaving, persuasion, and the expression of author stance (inter alios, Berkenkotter and Huckin 1995, Fløttum et al. 2006, Hyland and Bondi 2006, Hyland and Guinda 2012. Genre-based studies have revealed how specific academic genres are crafted, how they evolved or declined over time, and what specific features characterize specific genres, such as research articles, abstracts, textbooks, or academic book reviews, and compared them across disciplines. However, discipline and genre are not the only factors that have attracted the attention of academic discourse investigators over the years. The cultural background of the authors and its effect on the academic text have been the focus of structures differ between English and Lithuanian academic texts (research article introductions, book reviews, etc.), and the second part presents a comprehensive overview of the functional groups of linguistic units, such as metadiscourse markers (Hyland 2005a), stance and engagement markers (Hyland 2005b) and conceptual metaphors, and how these are different or similar to English academic discourse. The studies selected in the review are the research texts, written by doctoral students and experts, which are focused on the above-mentioned aspects. As such, these studies reveal the ways in which the argumentation of Lithuanian academic discourse is shaped both on the macro-and on the micro-structure level, in accordance with the established model and framework used in studies of English academic discourse. They have been searched for in the database "Lituanistika" which includes high-quality research publications about Lithuanian studies.¹ Due to the limited size of the article, we could not overview some other interesting aspects that are relevant for academic discourse investigations, such as the varied scope of the data in the studies reviewed as well as the methodologies employed. Though very relevant, these issues fell outside the focus of the present article and hold potential for future examination.
Although this review is by no means exhaustive, it shows a number of tendencies in the choice of rhetorical structures and metadiscourse elements in Lithuanian academic discourse that diverge from patterns and functional distribution of markers in English academic discourse and reflect a national writing tradition.

Rhetorical structure 2.1.1 Rhetorical structure and what it tells us about features specific to Lithuanian
The rhetorical structure of a text pertains to its overall composition, focusing on individual sections (i.e. introduction, body, and conclusions) and is considered part of the analysis of the macrostructure of an academic text. Different academic genres (i.e. research articles and abstracts) reveal divergent structural patterns or moves (Hyland 2006, 27), which make up an essential component of authorial argumentation. As noted by Livnat (2012, 26), the structure of a research article has its own goal and function which is to present a convincing argument and persuade the readers of the credibility of that argument.
The development and arrangement of an argument may be more conventional, following a certain model, or more creative, adjusting to the needs of a particular discipline or cultural tradition. Within academic genre analysis, much attention has been devoted to the examination of the rhetorical structure of research article introductions by drawing on Swales ' (1990) CARS model, which includes three main moves: establishing a research territory (Move 1), establishing a niche (Move 2), and occupying the niche (Move 3). Each move consists of a number of steps.² For instance, by establishing a research territory (Move 1), the author may claim centrality, i.e. importance of research (1); by establishing a niche (Move 2), the scholar may indicate a gap in the current research (2); and by occupying the niche (Move 3), the structure of the article may be outlined (3), as shown in the following examples: (1) The role of housing wealth in underpinning household spending and borrowing decisions has been extensively discussed, particularly in the United Kingdom (UK) and the United States (USA). In the UK, housing is a particularly important component of the household's wealth portfolio.
[…] (Move 1) (Šinkūnienė 2014, 55) (2) A writer who has been nominated for the Man Booker prize four times, whose novels are popular set reading in British secondary schools, and who has achieved critical acclaim and publishing success throughout the English-speaking world might be expected to be the object of a significant academic industry, yet there have been just three critical overviews of McEwan's novels in the last 10 years. (Move 2) (Šinkūnienė 2014, 74) (3) The article is structured as follows. The next section presents a model of individual employee crime that suggests several potential influences on this activity. Then Section II describes our data and methods before Section III presents and discusses the results. Section IV concludes.
[…] (Move 3) (Šinkūnienė 2014, 61) The CARS model has been tested in a number of cross-linguistic studies examining how introductory sections of research studies or PhD theses are constructed in English and other languages, such as Portuguese (Hirano 2009), Spanish (Mur-Dueñas 2010, Sheldon 2011, and Norwegian (Solli and Ødemark 2019), in different disciplines (applied linguistics, business management, computing, etc.). These studies show a number of divergences from the CARS model in the realisation of individual moves and steps, which highlights different national strategies and standards of shaping argumentation on the macrostructure level. For instance, Brazilian scholars writing in Portuguese (Hirano 2009) as well as Spanish researchers (Mur-Dueñas 2010) tend not to indicate a gap (Move 2) in their studies, which may be explained by the specific cultural context, i.e. avoiding criticism of local researchers or facing less competition while addressing the national audience. National academic discourses share some features with English academic discourse, as shown by a contrastive study of English and Spanish research article introductions (Sheldon 2011) and by Norwegian introductory sections of research articles (Solli and Ødemark 2019); however, they seem to represent their own standards, diverging from the conventions of the Anglo-American tradition of academic writing.
The following subsection will present a number of rhetorical choices typical of introductions and acknowledgements attested in Lithuanian research articles and discuss the rhetorical structure and evaluation of book reviews. The studies reviewed will highlight how various rhetorical moves and steps in different academic genres in Lithuanian diverge from the rhetorical patterns established in English academic discourse and how these divergences reflect national features.

Rhetorical structure studies in Lithuanian academic discourse
Some of the Lithuanian academic discourse studies have investigated the structural features of research articles and academic book reviews from the rhetorical moves and steps perspective (the CARS model by Swales 1990Swales , 2004. The analysis of research article introductions in economics, medicine, and linguistics in Lithuanian and English (Ryvitytė 2003) revealed that even though Lithuanian scholars use centrality claims to emphasise the importance of their research area, they do this somewhat weaker than their English-speaking colleagues. The lowest occurrence of centrality claims is observed in Lithuanian linguistic research articles and is explained by the author as the result of a lower competition in getting the research published. Another interesting feature of Lithuanian introductions was the absence of continuing a tradition as a way to open the niche (only one introduction in the medical field used this way to open a niche). The author observes a strong tendency, especially in the field of economics, to open the niche by indicating a gap "often referring to research-external factors justifying the transition to a new tradition" (Ryvitytė 2003, 99). Finally, indicating the research article structure was also not popular in the introductions of Lithuanian researchers (only one such case in one economics article).
Another study that looked at research article introductions using the CARS model focused on articles written by British and Lithuanian scholars in economics, sociology, linguistics, and literature (Šinkūnienė 2014). The first part of the study looked at the introductions of the texts in social sciences. The centrality claim was present in most of the analysed introductions and was constructed in very similar ways in both languages. Just like in Ryvitytė's research (2003), the tendency in Lithuanian economics and sociology articles was to create a niche by indicating a gap rather than by continuing a tradition. A popular motivation for the research to be conducted was the claim that similar studies were done in other countries but not in Lithuania. This was different in the English language articles which were split evenly between the two ways of establishing a niche (continuing a tradition and indicating a gap). Similar to Ryvitytė's research (2003), Lithuanian economists did not provide the structure of their article in their introductions. Conversely, all English language articles displayed this rhetorical step. The tendency to end the introduction with the outline of the article was noticed in economics articles written in some other cultures such as Norwegian and French (Fløttum et al. 2006, 166). Šinkūnienė (2014, 62) explains this lack of the structural plan of the study by the length of Lithuanian articles which tended to be much shorter than the English ones. However, Lithuanian scholars, especially in economics, were keen on providing details on methods and theoretical framework in the introduction, whereas this was not so frequent in the research article introductions written in English.
One of the interesting rhetorical steps of Move 3, identified in the updated CARS model (Swales 2004), was emphasising the value of the present research. This step was present in over half of the articles in English but occurred in far fewer Lithuanian articles which could point to the fact that Lithuanian researchers in social sciences are not so promotional in showing the value of their research (Šinkūnienė 2014).
The second part of the study focused on the rhetorical structure of research article introductions in the humanities. The moves and steps were realised in similar ways in the texts of both Lithuanian and British scholars. The centrality claim was present in nearly half of the studies both in Lithuanian and in English. Perhaps, the major difference could be observed in the fact that Lithuanian linguists explicitly indicated the niche to a smaller extent than their English-speaking colleagues. Drawing on the interviews with informants, the author explains this tendency by the local nature of the audience which perhaps needs less proof that the study at hand is necessary. The lack of an explicit niche for the study could also be related to a lower competition to publish the research (Šinkūnienė 2014, 71).
The introduction of research articles is not the only genre analysed in Lithuanian academic discourse in terms of rhetorical structure. Another genre which was analysed using the adapted CARS model by Swales (1990) was academic book reviews (Ryvitytė 2004). The author analysed Lithuanian and English academic book reviews in linguistics. The study showed that the rhetorical structure of the book reviews in both languages was similar, but there were slightly different strategies used by Lithuanian and English reviewers. The main cultural difference was that the Lithuanian reviewers seemed to be more restricted in expressing their evaluation. Ryvitytė (2004, 116) explains this drawing on the fact that Lithuanian linguistic community is much smaller than the international one, and therefore, "the possibility of direct contact is much greater," thus possibly restricting the reviewers from a more elaborate evaluation.
Finally, one more specific structural part of academic texts that has received some attention in Lithuanian studies of academic discourse is acknowledgements (Šinkūnienė and Dudzinskaitė 2018). The study was based on acknowledgement texts in English and Lithuanian representing two genres within the field of humanities: scientific books written by British and Lithuanian researchers in their native languages, and doctoral dissertations written by Lithuanian doctoral students in Lithuanian. Hyland's (2004) model of the generic structure of acknowledgments ( Table 1) was used for the analysis. The study indicated that the British scholars seem to place more importance on acknowledgements as their texts were lengthier and included thanks for far more individuals and institutions than the acknowledgements of the Lithuanian scholars. This was especially obvious in the case of Lithuanian doctoral dissertations, which rarely included acknowledgements. The authors explain this tendency by the tradition to acknowledge individuals or institutions orally immediately after the defence. The generic structure analysis revealed some differences with doctoral students tending to be more grateful for moral support than established scholars. Acknowledgements in scientific books in both cultures seem to be constructed in a similar manner to the exception of the popularity of the dedication step. The British scholars more frequently dedicated their published books to somebody, typically a close person. The general conclusion that the authors of the article make is that "even though there were some differences between genres and cultures, the most frequent moves of acknowledgments and their formulations seem to be universal despite the cultural background of writers or the genre, and include thanking for academic assistance, resources and moral support" (Šinkūnienė and Dudzinskaitė 2018, 84).
To sum up, rhetorical structures in research article introductions, book reviews, and acknowledgements in Lithuanian across different fields and disciplines reveal a number of divergences in realising particular moves and steps attested in English academic discourse, following Swales' and Hyland's models of analysis. Among the divergent features, we find examples of the absence of continuing a tradition while opening a niche in article introductions (e.g. economics and sociology), lack of explicit indication of a niche (e.g. linguistics), lack of outlining the structure of an article (e.g. economics) or highlighting the value of research. Acknowledgements in Lithuanian scholars' books or PhD theses are shorter and contain fewer references to institutions and individuals than those of British scholars.

Functional groups 2.2.1 Functional groups of linguistic units and what they tell us about features specific to Lithuanian
Features of academic discourse emerge not only from the analysis of the macrostructure of scientific texts but also from a thorough examination of their microstructure, i.e. grammatical markers (e.g. modal verbs, personal pronouns, and directives) and lexical devices (e.g. evaluation/attitudinal markers, hedges, and boosters) that contribute to the development of effective authorial argumentation at the paragraph and sentence level. These units have been widely investigated in the areas of metadiscourse, stance, engagement, modality, and evidentiality. Although they have been addressed in diverse frameworks presupposing different theoretical and practical implications, the core functions of these markers are concerned with the author's presence in the text and a rapport with the audience.
One of the widely applied frameworks in cross-linguistic studies is Hyland's (2005a) model of metadiscourse, which entails a range of markers aimed at structuring the text (e.g. transitions, frame markers, endophoric markers, evidentials, and code glosses) and markers which establish authorial voice or a relationship with the addressee (e.g. hedges, boosters, attitude markers, and self-mentions). Both types of metadiscourse devices will be discussed in greater detail in Section 2.2.2. This framework has been quite effective in disclosing some features of national academic discourse as well. For instance, a study of business management research articles produced by American-based scholars and by Spanish writers has shown that Spanish writers tend to use fewer metadiscourse markers than their American peers (Mur-Dueñas 2011, 3074-5), which reflects different styles of writing and writer-reader relationship created in academic discourse of the two cultures.
Similarly, cross-linguistic studies exploring specific metadiscourse categories, such as metatextual markers (e.g. in this section, here, above, and below) in English, Norwegian, and French (Fløttum et al. 2006, 64), reformulation markers and citations in English and Spanish (Cuenca 2003, Mur-Dueñas 2009, Murillo 2012, as well as hedges in Bulgarian and English (Vassileva 2001) and in French-English-Norwegian (Vold 2006) have illustrated how these metadiscourse elements contribute to the characteristics of national academic discourses. For instance, the dominance of metatextual markers in English articles, in comparison to their Norwegian and French counterparts, may point to more reader-oriented English academic discourse, which signposts various discourse relations (Fløttum et al. 2006, 64). The lower frequencies of reformulation markers in Spanish business management research articles, compared to the English articles, reflect the national target audience the Spanish texts address (Murillo 2012). The higher frequencies of hedges in English academic discourse, as compared to Bulgarian, French, or Spanish academic texts (Vassileva 2001, Vold 2006, Mur-Dueñas 2011, suggest more prominent dialogicity of the English texts that address the international audience. The salience of authorial presence in academic discourse has been revealed in numerous studies devoted to the use of personal pronouns across languages. As illustrated by the empirical cross-linguistic evidence, the first-person singular pronoun is avoided in Spanish (Mur-Dueñas 2007), Italian (Molino 2010), German (Vassileva 1998, Sanderson 2008, Bulgarian, French, and Russian academic discourses (Vassileva 1998). Researchers prefer using the first-person plural pronoun we and impersonal or passive constructions (Mur-Dueñas and Šinkūnienė 2016). In German, this tendency is explained by the culturally specific 'I-taboo' (Sanderson 2008, 71), and in French, it is considered as le moi haïssable 'hateful me'. The comparison of the use of personal pronouns in English, Norwegian, and French articles in linguistics, economics, and medicine confirms that the French academic texts contain the lowest number of first-person pronouns (Fløttum et al. 2006, 63).
Markers examined within the fields of modality, especially epistemic modality, i.e. modal verbs (e.g. may, might, must, would, etc.), adjectives (e.g. possible, probable), and adverbials (e.g. perhaps, certainly, possibly, etc.), cast light on the author's commitment to the information conveyed. Markers of evidentiality, i.e. adverbials (clearly, obviously, evidently) and complement-taking predicates ([CTPs] e.g. see, seem, say), express the author's sources of information for the proposition. Modal and evidential markers have been thoroughly studied in English (Biber 2006, Hidalgo-Downing 2017, French Wirth 2007, Grossmann andTutin 2010), and Spanish academic discourses (Oliver del Olmo 2009, Alonso-Almeida and Adams 2012) in terms of their functions and distributional properties. Modal and evidential markers may function as hedges or boosters and add to the reliability of the information provided. They may also explicitly or implicitly indicate degrees of subjectivity or intersubjectivity and inevitably contribute to the construction of authorial voice and reader engagement strategies.
Unlike rhetorical structures in research articles and book reviews, linguistic units representing various functional groups have received much more attention in Lithuanian academic discourse, which may signal an overall preoccupation and greater concern with the grammatical, lexical, and discursive choices in Lithuanian scientific texts.

Studies of functional groups of linguistic units in Lithuanian academic discourse
The usage patterns of functional groups of linguistic units in Lithuanian are typically contrasted with the patterns characteristic of English academic discourse. Most of the studies in this area are based on relatively small size self-compiled corpora; however, the large synchronic corpus of academic language CorALit (Corpus Academicum Lithuanicum, about 9 million words of Lithuanian academic discourse texts in humanities, social sciences, technology, biomedicine, and physics) is also employed.
One of the popular markers of investigation are hedges, i.e. "devices such as possible, might and perhaps, which indicate the writer's decision to recognize alternative voices and viewpoints and so withhold complete commitment to a proposition" (Hyland 2005a, 52). A pilot contrastive study of hedges in Lithuanian and English research articles in medicine and linguistics (Šinkūnienė 2008) revealed that English-speaking researchers use significantly more hedges than Lithuanian researchers in both disciplines. A more extensive study of hedging in Lithuanian and English research articles in medicine and linguistics (Šinkūnienė (2011)) confirmed the same tendency for Lithuanian researchers to hedge less than their English-speaking colleagues in both disciplines. Even though the types of the most frequent linguistic items used as hedges were different in both languages (English preferred lexical verbs and modal verbs; Lithuanianadverbs and adverbials), the pragmatic functions of hedges were similar irrespective of the languages. The tendency for Lithuanian scholars to hedge less than their English-speaking colleagues was further confirmed by a sample study of Šinkūnienė (2014) of research articles in four analysed disciplines (economics, sociology, literature, and linguistics) in Lithuanian and English.
With regard to the studies of hedging based on large corpora containing texts from many disciplines, Šinkūnienė (2012) offers a detailed investigation of adverbial hedges in five different science fields of Lithuanian discourse based on the 9 million word Corpus of Academic Lithuanian. The study showed that there is a marked difference in the frequency of the use of hedges, with most of them occurring in the soft sciences (humanities and social sciences). The hard science texts (in physics, biomedical field, and technology) displayed fewer hedges. Overall, however, it is noted that "Lithuanian researchers are not very keen to soften their observations with adverbials marking tentativeness and possibility" (Šinkūnienė 2012, 162).
A cross-linguistic glimpse into the rhetorical strategy opposite to hedging, i.e. boosting or rhetorical means to strengthen the proposition (by using such markers as no doubt, surely, certainly, of course, etc.), shows that both English and Lithuanian researchers seem to boost to a similar extent (Šinkūnienė 2014). However, these results are very tentative as they are based on a small sample study of 40 research articles in economics, sociology, literature, and linguistics in Lithuanian and English.
One more functional group of linguistic units employed to either express stance or engage the reader is personal pronouns aš 'I' and mes 'we' and their forms. Personal pronouns are generally one of the most researched linguistic items in academic discourse in various languages. Unlike in English, in the Lithuanian language, the pronoun function can be realised not only by the personal pronoun form but also by a verbal inflection, so both (4) and (5) are possible: (4) Šiandien tur-ime 'have-PRS.1PL' rašytinės lietuvių kalbos didelės apimties tekstyną […]. 'Today we have a large corpus of written Lithuanian.' (5) Mes 'we.nom' visi žin-ome 'know-PRS.1PL', kad jis pralaimėjo prezidento rinkimus. 'We all know that he lost the presidential elections.' (Šinkūnienė 2018, 63-4) The studies of personal pronouns in Lithuanian academic discourse show that Lithuanian scholars use far less personal pronouns than their English native-speaking colleagues (Šinkūnienė 2010, 2014, 2018) irrespective of the disciplinary field. This tendency was not very surprising keeping in mind that personal pronouns in academic texts are not very frequently used in many other cultures and languages in comparison to English. A review of 22 empirical studies focusing on personal pronoun use in academic texts in 13 languages of the world shows that only Norwegian scholars tend to use personal pronouns to a similar extent as the English native-speaking academics (Mur-Dueñas and Šinkūnienė 2016). In those cases when they choose to employ a personal pronoun, Lithuanian scholars in the humanities (linguistics and literature) tend to opt for the inclusive mes 'we', but avoid aš 'I'. In the analysed disciplines of social sciences (economics and sociology), Lithuanian scholars prefer the exclusive mes 'we' to explicitly refer to themselves as authors especially describing the application of the methodology of the study or the results obtained. The inclusive mes 'we' in social sciences is frequently used to direct the attention of the reader to the study, especially in combination with the perception verb matyti 'see' (Šinkūnienė 2014).
It is interesting to note that there is a tendency for Lithuanian linguists to use personal pronouns to a greater extent when they write in English. A study of research articles written by Lithuanian linguists in Lithuanian and then the same authors in English shows that in their Lithuanian language texts, scholars resort exclusively to we and its forms, whereas when they write in English, they also tend to use the personal pronoun I. This could be "the result of the influence of the Anglo-American academic writing tradition, and reflect an attempt of the scholars to adapt their writing style to the conventions of the language in which they compose the text" (Šinkūnienė 2018, 77). However, even when they use the personal pronoun I in their English language articles, "Lithuanian linguists do not tend to argue for their position in the texts, but rather assume a much safer role of the writer, who guides the reader through the text and the argument" (ibid.). Conversely, British linguists appear in the arguer role more frequently than their Lithuanian colleagues (ibid.).
Finally, the study of Lithuanian personal pronouns across five science fields in the 9 million word Corpus of Academic Lithuanian (CorALit) (Linkevičienė and Šinkūnienė 2012) confirms that the frequency of personal pronoun mes 'we' and especially aš 'I' seems to be low in Lithuanian academic discourse. It also confirms the trend for the researchers in the hard science fields to employ personal pronouns to a much smaller extent than researchers in the soft fields. The inclusive mes 'we' is favoured more by scholars in the humanities and social sciences, whereas the exclusive mes 'we' is more popular in biomedical, physical, and technological fields.
Another group of functional markers that received some attention in studies of Lithuanian academic discourse are reformulation markers, which belong to the interactive category of metadiscourse markers in Hyland's (2005a) model and are part of the so-called code glosses employed in the scientific text to signal explanation, rewording, elaboration, or illustration (Hyland 2007). Typical examples of reformulation markers in English are i.e., that is, that is to say, in other words, to put it another way/differently/simply. In Lithuanian, the most common reformulation markers are very similar and include tai yra/t.y. 'that is/i.e.', kitaip tariant/sakant 'to say differently/to put it another way', kitais žodžiais tariant 'in other words'.
Šinkūnienė (2014) shows that Lithuanian writers use more reformulation markers in their research articles in economics, linguistics, literature, and sociology than their British colleagues in the same disciplines. The author suggests that this could be the case due to Lithuanian scholars' wish to facilitate the reader's comprehension of the text by providing a substantial number of explanations, reformulations, and periphrasis. Another study which was based on much larger corpora (CorALit for Lithuanian and Corpus of Contemporary American English for English empirical data) looked at the use of reformulation markers in the fields of humanities, technological sciences, and (bio)medical sciences. The results of this much larger study confirmed the same trend, i.e. reformulation markers were more frequently employed in Lithuanian academic texts than in English. The author concludes that these results could suggest different rhetorical principles typical of the writing style of the two cultures (Šinkūnienė 2019).
The use of endophoric markers has also been investigated in Lithuanian academic discourse (Šinkūnienė 2014). Endophoric markers "refer to information in other parts of the text" (Hyland 2005a, 49) and include such linguistic items as in Section 3, as mentioned above, as will be mentioned below with similar linguistic expressions in Lithuanian (for example, kaip jau minėta 'as was mentioned', kaip bus matyti toliau 'as will be seen later', žr. kitame skyriuje 'see next chapter'). These markers are frequent in Anglo-American academic discourse where generally reviewing and previewing are considered to be helpful guiding strategies facilitating the creation of coherent and reader-friendly text. The study of Lithuanian endophoric markers showed that Lithuanian researchers, especially in literature and sociology, are not very eager to guide their readers. However, it is interesting to note that their British colleagues also show similar patterns, so it could be the result of disciplinary conventions rather than cultural influence. The arguments in those two fields are frequently constructed as a one-directional flow, with rare reviewing and previewing. The major difference in the use of endophoric markers can be observed in the articles on linguistics and economics: here, the Lithuanian scholars use far less endophoric markers than the British scholars. Though it may be tempting to suggest some cultural differences here or the impact of the teaching academic writing traditions (Dahl 2004), the author notes that Lithuanian research articles in those two disciplines are twice shorter than the British articles; hence, the difference in the use of these markers could be the result of the less urgent need to tell the reader what has been already said or what will be said (Šinkūnienė 2014).
Another group of functional items that have received some attention in Lithuanian academic discourse is citations (Petrauskaitė andŠinkūnienė 2015, Šinkūnienė 2017). The article by Petrauskaitė and Šinkūnienė (2015) focuses on PhD dissertations written by Lithuanian early career researchers in sociology and cultural studies. The results of the study are based on a number of quantitative and qualitative citation parameters (citation density, year and type of the publication, syntactic integration of citation, citation clusters, and the distribution of citations in different sections of the dissertation) and show clear disciplinary differences in the ways citations are employed. The sociological work relies more on research articles and prefers the non-integral citing mode with less detailed referencing to the sources (just the year of the publication indicated). The cultural studies dissertation, on the other hand, shows a preference for books and the integral citation mode dominating which allows to shift emphasis to the cited author rather than to the information (Petrauskaitė and Šinkūnienė 2015).
The influence of the disciplinary tradition is also noted in a contrastive study of citation patterns in literature and linguistic research articles written by Lithuanian and British scholars in their native languages (Šinkūnienė 2017). The study focused on frequency, syntactic integration patterns, and types of citations. The author concludes that "no striking cultural differences were observed in the analysed texts, which might point towards discipline as a decisive factor in citational trends in literature and linguistics" (Šinkūnienė 2017, 267).
Evaluation has also received some attention in Lithuanian academic discourse studies. A small study of the evaluation of citations in scientific books within the field of humanities reveals that Lithuanian authors do not express their evaluation concerning the cited author or his/her text (Damošius 2007). The expression of the evaluation was also studied in 60 linguistic academic book reviews in English and Lithuanian published in academic journals in 2008-2014 (Ryvitytė 2015). The author found that Lithuanian scholars were more cautious in expressing their criticism than their English-speaking colleagues. Another interesting finding was that English reviewers expressed their overall positive evaluation of the book early in the review, while Lithuanian reviewers "displayed greater reservation" (ibid).
Another functional group of units investigated in Lithuanian academic discourse are evidential markers, expressed by a variety of structural devices, such as non-agreeing adjectives deriving from the semantic domain of perception (akivaizdu 'evident', aišku 'clear') or comparison (panašu 'like, likely') (Ruskan 2012), non-agreeing participles (manoma '(is) thought to', žinoma '(is) known to, obviously', teigiama '(is) claimed to'), perception-based verbs (matyti 'see', atrodyti 'seem'), nouns related to knowledge and truth (žinia 'knowledge', tiesa 'truth') (Usonienė 2013(Usonienė , 2015(Usonienė , 2016 as well as perception-based adverbs (akivaizdžiai 'evidently', aiškiai 'clearly') (Ruskan 2015a). A number of studies have compared the distribution, functional, and formal features of evidential devices in fiction and academic discourse (based on data from the Corpus of Academic Lithuanian). By indicating the source of information for the proposition (inference or report), evidential markers in academic discourse contribute significantly to authorial argumentation and grounding of subjective statements. Occurring in impersonal patterns (e.g. it is evident/clear/likely/thought/known that), they may background the authorial voice and indicate the shared status of information, i.e. common knowledge with the addressee. The structural diversity of evidential markers in Lithuanian allows for various configurations of authorial argumentation (persuasion, distance, cautiousness, etc.).
Evidential markers in Lithuanian academic discourse have also been investigated from a cross-linguistic and cross-disciplinary perspective. Ruskan (2015b) compared two types of grammaticalising evidential expressions in Lithuanian and English academic discourse in the soft and hard sciences, namely parentheticals deriving from participle-based CTPs in Lithuanian (e.g. sakoma '(is) said to', manoma '(is) thought to') and constructions of passivized verbs of communication and cognition taking the infinitival complement (e.g. be said to, be thought to) in English. Both the passive constructions with the infinitival complement in English and the parenthetical participle-based CTP clauses in Lithuanian function as reportives which express hearsay, mindsay, proof, and general knowledge. Despite some functional similarities, the evidential markers showed distributional differences in individual science areas. In English, the passive constructions were more frequent in the hard sciences than the soft sciences because they primarily express an objective source of information related to other scholars' opinions, beliefs, considerations, and assumptions or convey reports of general knowledge. In Lithuanian, the parenthetical CTP clauses were more frequent in the soft sciences because they may also function as markers of interaction and engagement with the reader that are hardly attested in the hard sciences. Thus, the evidential parenthetical expressions in Lithuanian academic texts play an important role in the construal of dialogicity and writer-reader relationship in the text.
Although a number of studies dealing with various functional groups have looked at Lithuanian academic discourse from a cross-linguistic perspective with English, they have allowed for the identification of some important features specific to Lithuanian academic discourse. As shown above, some functional groups or individual linguistic markers may be quite prominent and widespread, whereas other types of markers are not dominant. The salient features of Lithuanian academic discourse on the microstructure level point to less hedging, infrequent use of personal pronouns (especially aš 'I'), frequent use of reformulation markers, cautious evaluation in book reviews, and an important role of evidential markers in expressing authorial argumentation and degrees of (inter)subjectivity. It seems that like with structural choices attested in research articles and book reviews, the repertoire of strategies to foreground or background the author's argumentation on the micro-level of analysis varies from explicit conventional means of expression to more implicit realisations of author positioning and addressee guidance in discourse, which again may reflect different cultural traditions of argumentation and academic writing.

Conceptual groups 2.3.1 Conceptual groups of linguistic units and what they tell us about features specific to Lithuanian
Alongside various linguistic units reviewed in the previous section (metadiscourse markers, engagement and stance markers, and evidential and modal markers), conceptual metaphors⁴ have been another area of investigation in academic discourse. A special role of metaphors in academic discourse can be signalled by the higher percentage of "metaphor-related words" in comparison to other registers, such as news, fiction or conversation (Herrmann 2013, 137). The study of metaphors in academic discourse may reveal culture or language-specific conceptualisation of abstract notions and ideas and show how metaphors shape authorial argumentation and add to the construction of a scientific text. Like various metadiscourse markers, metaphors may cast light on the expression of authorial voice and subjectivity and thus can be considered as means of expressing evaluation (Martin and White 2005). Apart from the evaluation (6), they may specify the reference of an entity (7), create textual cohesion (8) or provide illustrative explanation (Herrmann 2013, 299), as in the following examples:  (Herrmann 2013, 217) In (6), the metaphorical noun springboard evaluates positively the essays in the book discussed. In (7), the noun attacks used metaphorically refers to an abstract concept of conflict, and hence, this metaphor performs a referential or ideational function. In (8), the noun attacks shows a connection with the previous utterance and thus performs a textual function.
The concepts behind the metaphorical constructions may significantly add to the understanding of the features of Lithuanian academic discourse since metaphors reveal the way humans conceptualise reality, i.e. what source domains they choose to portray more abstract target domains. The following subsection reviews several studies exploring metaphorical expressions in Lithuanian academic texts.
 4 In the studies reviewed in this article, "metaphor is understood in terms of one domain mapped onto another (A mapped onto B), e.g. the expression to arrive at certain criteria is a metaphorical expression of the metaphor REASONING IS A JOURNEY" (Šeškauskienė 2008, 274), whereas metonymy pertains to mapping/s "within one domain, e.g. the expression the data suggest refer to the researcher suggesting that the data could be interpreted in one or another way" (ibid. 274).

Studies of conceptual groups of linguistic units in Lithuanian academic discourse
Most studies into metaphor in Lithuanian academic texts have been conducted from a cross-linguistic perspective. Šeškauskienė (2008) has explored metonymical and metaphorical expressions in English and Lithuanian linguistic texts collected from the journal of Applied Linguistics (ca. 20,000 words) and Kalbotyra (ca. 18,000 words) by applying the Conceptual Metaphor Theory (Lakoff andJohnson 1980/2003). In both languages, the prevailing metaphorical concept was LANGUAGE (ANALYSIS)/REASONING AS A BUILDING/STRUC-TURE (Šeškauskienė 2008, 274); however, metaphorical expressions were slightly more frequent in English than in Lithuanian. Metonymical expressions realised through the combinability pattern inanimate subject + active verb (e.g. the paper suggests/discusses) were also more frequent in English linguistic texts and did not occur in a number of Lithuanian texts (ibid. 274-5).
Another cross-linguistic study of the pattern inanimate subject + active verb (Šeškauskienė 2010), based on a larger set of data (the Lithuanian corpus of linguistic texts comprising 51,828 words), did not show any marked distributional differences of the pattern in English and Lithuanian academic texts, thus suggesting the similarity of the pattern in the two languages. Differences were observed in the types of dominant inanimate subjects and verbs in the two languages, showing more language-specific choices while presenting or explaining research. As claimed by Šeškauskienė (2010, 95), the pattern inanimate subject + active verb in Lithuanian entails "a variety of contexts: explanatory, presentation and change-of-state-orlocation, while English articles give preference to the explanatory context." In English linguistic articles, this pattern reflects a more general picture of conducting research, whereas in Lithuanian texts, it highlights specific procedures of carrying out research (ibid. 95). The study showed that in terms of metaphorical expression, the pattern inanimate subject + active verb in both Lithuanian and English linguistic texts can be conceptualised as RESEARCH IS A PERSON and RESEARCH IS A JOURNEY. Despite some overall similarities in the realisation of dominant metaphors, the study found a number of differences implying varying national traditions or conventions in academic writing as far as metaphoricity is concerned. For instance, the realisation of the metaphor RESEARCH IS A JOURNEY in English linguistic articles entails "more details about moving towards a destination and obstacles," whereas in Lithuanian linguistic texts, the focus is on "going itself" (ibid. 95). Among the noteworthy differences is the occurrence of language-specific metaphors. As illustrated by Šeškauskienė (2010, 94), the conceptual metaphor attested only in Lithuanian linguistic articles is LANGUAGE/RESEARCH IS PERFORMANCE (9): (9) Jungiant atskiras teksto dalis į globalų tekstą ypač svarbūs yra junginiai, atliekantys nuorodos vaidmenį.
Kalb 3 (Šeškauskienė 2010, 94) 'When connecting separate parts of the text into a global text, particularly important are word combinations performing the role of references.' (ibid.) Thus, the absence of particular metaphors or their lower frequencies in a text may suggest language-or culture-specific presentation of concepts.
Continuing the tradition of cross-linguistic research of metaphors, Urbonaitė (2015) has explored fieldspecific metaphors in English and Lithuanian research articles on criminal justice. The data were obtained from five articles (23,670 words) from the Lithuanian journal Teisė ('Law') and five articles (32,191 words) from the Journal of Criminal Justice written in English. The examination of metaphors in legal discourse yielded more similarities than differences since both English and Lithuanian legal discourse contained the dominant OBJECT, HUMAN BEING, and CONTAINER metaphors. The most frequent metaphors were CRIME/CRIMINAL LAW  (Urbonaitė 2015, 25). For instance, punishment could be conceptualised in terms of size and thus illustrate the conceptual metaphor of CRIME/CRIMINAL LAW IS AN OBJECT: (10) bausmė ir jos dydis (punishment and its size ( = measurement)). (Urbonaitė 2015, 18) Urbonaitė (2015, 25) maintains that although both English and Lithuanian legal research articles contain similar law-related metaphors, the metaphors display differences at the level of realisation.
The aforementioned cross-linguistic studies into metaphorical expressions in academic discourse have revealed a number of language/culture-specific and perhaps discipline (linguistics vs legal discourse) specific features. The choice of inanimate subjects and verbs in the pattern inanimate subject + active verb highlights the different ways of presenting research in English and Lithuanian ("picture at large" vs more specific presentation of findings) (Šeškauskienė 2010). In linguistic discourse, language is mostly conceptualised in terms of a building/structure, a measurable entity, an economic activity, a performance/play, or a journey/space/movement (Šeškauskienė 2008, 278), whereas in legal discourse, law-related concepts are viewed in terms of an object, a human being, or a container (Urbonaitė 2015, 25). More inquiry into the metaphoricity of Lithuanian academic discourse could extend the existing picture of national characteristics of Lithuanian academic discourse.

Discussion and concluding remarks
In this article, we have reviewed the scholarly works that help to pin down conventions of Lithuanian academic writing and what these can tell us about the features present in academic writing specific to Lithuanian. The article has aimed to profile the features of the academic discourse created by both PhD students and expert members of the Lithuanian academic community with regard to the rhetorical structure of specific structural parts of academic texts, functional groups of grammatical and lexical markers, and conceptual metaphors.
In sum, the key findings from studies summarised here allow us to conclude that in Lithuanian academic discourse, researchers are not very keen on mitigating their claims, they avoid explicit reference to themselves in the text (though this seems to be changing in texts written by Lithuanian scholars in English), and they tend to use reformulation markers thus facilitating readers' comprehension of texts. However, reviewing and previewing are not so typical of Lithuanian academic texts in some disciplines (e.g. linguistics and economics) in comparison to Anglo-American academic discourse. The wide variety of evidential markers allows Lithuanian scholars to express authorial persuasion, distance, reliability of knowledge, and shared knowledge and to create various reader engagement strategies. Conceptual metaphors attested in Lithuanian linguistic articles reveal a case of language/culture-specific conceptualisation of linguistic research, namely LANGUAGE/RESEARCH IS PERFORMANCE. The collocational patterns of metaphorical expressions in the linguistic articles reflect specific procedures of linguistic research rather than its general picture.
The evidence provided by this review highlights the importance of understanding the national contexts of culture and language. Such understanding provides the grounding for a more proportional outlook about different possible ways how a text convincingly presents an argument, and which textual features enable to persuade the reader of the text's credibility. We may see the Anglo-American tradition, which presents texts through the lens of a well-organised English language of science, as a standard which we can all attest to. However, given the presence of so many more languages in the scientific community, it would be biased to consider the standard of English to be the gold standard which all languages should strive for. This bias would be specifically troubling when it comes to preserving the language of science among the speakers of these other languages. As such, the article has a wider goal of highlighting those features of Lithuanian academic discourse that diverge from English academic discourse and to contribute to the awareness that these differences definitely do not suggest that one tradition is "better" than the other. This wider perspective is important since it helps us to broaden our understanding of diversity that exists in academic discourse and to recognise those various patterns as equal members of the academic community. It is often through comparative studies that we learn more about the writing structures of other languages. One issue which we aim to address in a larger study (Jürine et al. 2021) is to draw comparisons with other languages and cultures in the Baltic States. How well is their academic writing tradition described and understood? If they are well described (as in the case of Lithuanian) what distinguishes these languages and cultures from Lithuanian and others. If they are not well described and understood, why not and what are the predicaments as a result? Besides comparing different languages, we must keep in mind that disciplinary differences, even within one language, might be even larger than cross-linguistic differences. For example, it has been shown that the author profile in medical research articles is quite different from the manifestation of authors in economics articles (Fløttum et al. 2006). As such, when reporting on language-and culturespecific writing traditions, in our larger study (Jürine et al. 2021), we propose a model (Leijen et al. 2023) which will allow a wholistic approach to mapping writing on large databases of academic texts consisting of different genres and disciplines. The result is a better picture of the national features of academic discourse.
In this regard, a telling aspect of the research on national features of academic discourses is that when comparing various national languages to English, researchers tend to make claims that a certain feature, for example, a particular rhetorical move or step that is common to English is missing or is expressed to a lesser extent in other languages or that some rhetorical conventions are different from those of English. On the one hand, it is only natural, since a lot more is studied and known about English than about many other languages. On the other hand, putting these languages against the English standards might guide researchers to biased conclusions. Therefore, we suggest that researching languages other than English, and comparing them to English academic discourse, can have both a positive and a negative impact. The positive side would be that studies on languages can build a much better understanding of national academic discourse structures. The negative impact of comparative studies would be that, if not positioned properly, and put too much up against English, they diminish the importance and value of these local structures, if not placed on equal footing with English. Furthermore, it might be more beneficial to compare national discourses other than English with national languages and cultures which share greater similarities. For example, considering Kaplan's doodles, which cover very different languages and cultures, the unavoidable outcome is a presentation of stark contrasting patterns. Comparing the academic discourse structure of Lithuanian with languages more closely related structurally, geographically, or historically would be much more insightful, as these cultures might share a similar background or the languages a similar structure. The argument here is that comparing traditions would gain value, when, in fact, we have a greater diversity of resources to draw comparisons from, which includes language and cultural resources as well as disciplinary.
Despite the numerous advantages that English offers as the dominant language of science and comparative studies, it also poses a danger to other academic languages, as has been highlighted by Swales (1997) and Hyland (2016). Swales, in his self-reflective article, English as Tyrannosaurus rex, and Hyland (2016) in his article, Academic Publishing and the Myth of Linguistic Disadvantage, both agree that English as the academic language is the dominant language. There are clear advantages and disadvantages of this in the context of smaller academic cultures and languages; on the one hand, English allows the research of other cultures to be included in the larger discussion of science globally, but at the same time there may be a risk that focusing too much on the English tradition can lead the researchers who represent smaller academic cultures to start imitating this tradition in their own languages without having fully understood the strengths of one's own writing traditions.
Our position is that linguistic injustice, as Hyland proposes it, can become linguistic justice if more diversity and understanding of other writing traditions can be part of academic writing. In other words, English can certainly appear like a Tyrannosaurus rex, to use Swales's analogy, but rather than focusing only on the Tyrannosaurus, in the Jurassic Park we also have the Dilophosaurus and Velociraptor fighting for their recognition and survival. The threat of English may not even be apparent, and it does not have to affect the overall vitality of a language. Instead, the structures influenced by the English-saurus might be hidden in certain areas of language use that seem to be so common and have thus gained little attention from researchers.
In addition to the possible dangers posed by the English language, there can be other language-or culture-specific tyrannosauruses, which may affect the overall perceptions and attitudes of academic texts. For example, as Korotkina (2018) has pointed out in the context of Russian academic writing, the old Sovietera conventions may have a negative effect on the quality and trustworthiness of texts, as this style is not perceived as clear and honest by the international research community. Similar problems may arise in other languages, when the local conventions seem perfectly acceptable among local researchers, but these researchers are hindered in their ability to publish their texts, if at all possible, in international journals, to give an example. Therefore, the question should not be centred only around the English language rhetorical patterns and traditions, but rather more generally around the international standards of research publications. Once a more thorough understanding is obtained, we can build further research questions around features of academic rhetoric which will help guide future research and a better comprehensive model of academic writing in other languages, such as Lithuanian.