KURAM VE UYGULAMADA EĞİTİM BİLİMLERİ EDUCATIONAL SCIENCES: THEORY & PRACTICE Copyright © 2018 EDAM www.estp.com.tr DOI 10.12738/estp.2018.4.0260  2018  18(4)  831–859 Received: March 15, 2018 Revision received: July 29, 2018 Accepted: August 30, 2018 Research Article Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative Study of Hedges and Boosters in Postgraduate Academic Writing 1 Erdem Akbas Erciyes University 2 Jan Hardman University of York Abstract From a cross-linguistic and cross-cultural perspective, this paper reports on the findings of an exploratory study examining the features of the academic texts produced by three groups of postgraduates: native speakers of Turkish (TL1), English (EL1) and Turkish speakers of English (EL2). To this end, the study involves a microdiscourse analysis of a corpus of ninety discussion sections of dissertations to identify and classify the choices made by the authors for expressing commitment/detachment in presenting knowledge claims. The results indicated interesting similarities and differences across the groups in the ways in which writers qualified their level of commitment to a higher level and detachment from the claims in their writing. In other words, this can be described as a cline from the highest to the lowest, even intentionally withholding their commitment. By looking at the hedging and boosting devices contributing to the interactive side of academic writing, the discourse constructed by Turkish L1 writers appeared to be slightly less interpersonal but highly authoritative overall. In contrast, the results suggested that the Turkish writers of English were similar to their English L1 counterparts in terms of building a significantly more cautious strategy for presenting knowledge claims and making use of relatively fewer boosting devices when presenting their claims. It is hoped that the implications of the findings can be useful for teaching of academic writing to postgraduates within the contexts of the study. Keywords Postgraduate academic writing • Metadiscourse • Hedges and boosters • Discourse analysis and corpus linguistics • Commitment and detachment 1 Correspondence to: Erdem Akbas (PhD), Department of English Language Teaching, Erciyes University, Kayseri 38039 Turkey. Email: erdemakbas@erciyes.edu.tr 2 Department of Education, University of York, United Kingdom of Great Britain and Northern Ireland. Email: jan.hardman@york.ac.uk Citation: Akbas, E., & Hardman, J. (2018). Strengthening or weakening claims in academic knowledge construction: A comparative study of hedges and boosters in postgraduate academic writing. Educational Sciences: Theory & Practice, 18, 831–859. http://dx.doi.org/10.12738/estp.2018.4.0260 EDUCATIONAL SCIENCES: THEORY & PRACTICE In this widely explored topic, many researchers have paid considerable attention to how the viewpoints of writers are expressed with differential control over the force of propositions in written discourse. Strengthening or weakening the force of a proposition by means of linguistic items helps the author encode information in a format expected to be received in the way that is intended. That is why the focus of such research has mostly been on subjective or epistemic certainty so that researchers can examine various degrees and functions of writers’ implications regarding the truthfulness status of propositions by means of linguistic signals. The linguistic devices signaling a writer’s commitment to or detachment from knowledge claims can be evaluated as a component which is likely to change the level of confidence of the writer within the immediate context if the item is substituted or removed from the sentence. The following example from the subcorpus of Akbas (2014b) shows that the author of the sentence attempted to assert his/her view regarding teaching efficacy as definitively as possible by employing a very strong verb ‘found’ rather than signaling that “the writer is not prepared to personally guarantee the proposition” (Hyland, 1998, p. 173) by the use of such verbs as suggested, implied or indicated: The current study found that teaching efficacy could not be predicted by whether the participant was an in-service teacher or a student teacher (EL1-1). It can be noticed that this was likely to have been a conscious linguistic choice by the writer of the text, and substituting the strong verb “found” with one of these weakening verbs would simply result in a hedged point of view with a decrease in the level of certainty and confidence. Nevertheless, as can be seen and felt, the writer in the above example presented his/her finding in a confident and indisputable way to underline and boost the importance of the scientific contribution to the academic community. Although this can be regarded as a face-threatening act because it makes him/her fully committed to the proposition, the writer did not attempt to soften the claim and leave some room for the reader to evaluate the possibility of the knowledge claim being true or not. Thomas (1983) suggested that there is a strong association between pragmatic competence and successful communication for particular contexts. With this in mind, potential variations in the linguistic forms used for academic purposes in building an authorial presence, evaluation or stance can result in violating the consensus of the discourse communities over the specific uses of language in the form of accepted conventions. This idea has fueled the investigations of how communication is provided in academic texts with rhetorical choices made in the discourse. Ädel (2018) also pointed out that academic genres with respect to metadiscourse investigations have received much attention by the researchers in the field. To illustrate, a great deal of 832 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... research has explored linguistic elements signalling metadiscourse functions in order to reveal tendencies in writing practices across different discourse communities (regarding language, culture and discipline) and genres (Akbari, 2017; Crawford Camiciottoli, 2010; D’Angelo, 2008; Hatipoğlu & Algı, 2017; Hu & Chao, 2015; Liu & Buckingham, 2018; Molino, 2010). Nevertheless, the research examining academic discourse in terms of such expected and accepted norms has mainly been related to what experienced scholars follow in their academic writing (for example, Koutsantoni, 2004, 2005; Vassileva, 2001). In other words, there are studies which have compared the publications of experienced writers with a focus on various interpersonal relations in academic texts and this can be quite crucial for helping other members of these discourse communities to find what is generally accepted. There are also other studies which have contrasted professional and inexperienced writers without focusing on the question of genre as what is contrasted are generally two different genres, that is, research articles as opposed to dissertations. However, novice writers have rarely been the main focus so far (see Akbas & Hardman, 2017; Andresen & Zinsmeister, 2018; Bogdanović & Mirović, 2018; Gardner & Han, 2018; Ho & Li, 2018; Kawase, 2015; Vergaro, 2011). Considering the fact that novice writers of any discourse community are both novice with regard to their academic performance and managing authorial strategies to meet the expectations of experienced members of the academic community, that is the examiners, postgraduate writers will definitely need more instruction and guidance on how authorial strategies need to be managed in their particular contexts. Therefore, the fact that postgraduate students are novice writers due to having very little experience in corresponding to the expected academic practices has been the main concern for the present research. With an exploratory and comparative design towards the potential effect of language and culture on the writing of postgraduate students, this study aims to fill the identified gap of modelling postgraduate academic writing by thoroughly investigating the rhetorical choices made by writers from selected contexts for strengthening and weakening the force of propositions. To recap, with a corpus-driven approach, the present study was designed to identify the linguistic resources and rhetorical strategies used by three groups of novice writers to qualify their commitment/detachment for the sake of creating a fairly effective ethos3 in order to persuade their examiners about their knowledge claims. By analyzing a reasonably representative corpus of successfully completed dissertations written by novice writers, a range of strategies and preferences for displaying stance could be identified. As far as we are concerned, such a model would also be of great importance in contributing to the understanding of how postgraduate writers achieve signaling their commitment and detachment and express their viewpoints about propositions in order to engage with the target audience. The results of the present 3 Cherry (1988) distinguished ethos and persona in building an authorial presence in texts and by following this distinction, ethos has been used to refer to instances in which the author attempts to attain a level of credibility. 833 EDUCATIONAL SCIENCES: THEORY & PRACTICE exploratory study could therefore characterize some strategies and be used to guide future postgraduate writers in the contexts chosen. Review of the Literature There is increasing awareness among writers that academic practices vary based on the genres and the norms of the discourse community being contributed to. This essentially results in a case that the authors are expected to follow a range of accepted practices in presenting knowledge through scholarly work. Apart from the quality of their work, this involves orientating their own writing to the norms of a targeted discourse community (Hyland, 2005; Varttala, 2001), not just at the textual level (organization) but also at the level the propositional content. At the same time, the rhetorical choices would basically influence the interpretation and acceptance of the propositional content by the intended audience as far as the reliability and accuracy of the content are concerned. Assuming this, how the propositional content is conveyed seems to allow writers to gain credibility by projecting their writer-self (Hyland, 2002). This is perfectly in line with the argument of Coates (1987), suggesting that propositional content is presented after being epistemically-qualified (for example, it is possible that, this might be, it is obvious that). As far as the epistemic qualification is concerned, this chiefly allows the intended audience to assess the reliability and accuracy of the claims presented and the writer’s stance. The linguistic and rhetorical choices made by the writers, therefore, need to match the expectations of the discourse community so that the intended audience can interpret the propositional content easily from the way in which it is conveyed. Strengthening or weakening the force of propositions in academic knowledge construction is of enormous importance in terms of qualifying and packaging the information in the way in which the writers intend it to be comprehended by the audience. Such practices are labelled and discussed in a variety of ways in the literature. As an example, Stubbs (1986) refers “modality markers” to explore evaluative elements in texts, whereas Hunston and Thompson (2000) use the term “evaluation” in a broader sense to characterize discrete expressions signalling a writer’s beliefs, judgements and attitudes; Silver (2003) discusses in terms of a writer’s stance, to examine the linguistic items stressing the degree of confidence over propositions with the help of epistemic certainty. No matter what terms have been used to explore such relations in discourses, the qualification of a noticeable degree of commitment/detachment while conveying meaning through utterances can be used in packaging the knowledge claims and the representation of stance in academic writing. The linguistic resources employed to highlight a degree of commitment/detachment are considered primarily to be hedges and boosters. Aull and Lancaster (2014) suggest that “hedging and boosting allow writers to express more or less commitment to their claims, and they are regularly 834 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... featured in research on academic stance” (p. 159). Although the functions of hedges and boosters could well be linked to issues of politeness, authorial caution (Varttala, 1999), vagueness, modesty of claims (Crompton, 1997) and/or (un)certainty, both elements signal a noteworthy level of commitment/detachment. Following Stubbs (1986), Akbas (2014b) clarified the distinction between commitment and detachment with the concepts of hedges and boosters as follows: Expressing a degree of commitment occurs when the author attempts to signal a confident voice of authority and indicate a higher level of certainty towards the truthfulness status of the propositions. This can also be regarded as reinforcement of the truth value with a boosting effect in the statements via a range of linguistic items that can also be classified as boosters. On the other hand, expressing a degree of detachment occurs when the author withholds commitment so that a degree of doubt and hesitancy can be included in the presentation of the propositions. This can also be regarded as avoiding the presentation of definitive and factual knowledge claims, to open up the alternative voices for the reader’s consideration. The linguistic items classified as hedges can be used for explicitly qualifying a degree of detachment from what is asserted (p. 39). Even though Crismore, Markkanen, and Steffensen (1993) and Grabe and Kaplan (1997) stated that hedges and boosters are inseparable concepts, various researchers have intentionally undertaken studies related to the hedging concept only (for example, Atai & Sadr, 2006; Crompton 1997, Falahati, 2004; Hyland, 1996; Kranich, 2011; Lewin, 2005; McLaren-Hankin, 2008; Peterlin 2010; Šeškauskiene, 2008; Varttala 1999, 2001). Conversely, research examining the concept of boosters for expressing a high level of certainty has been limited to very few studies (such as, Bondi, 2008; Heiniluoma, 2008; Koutsantoni, 2005; Vázquez & Giner, 2009). Hyland and Milton (1997) carried out a comparative study with regard to hedges and boosters in the written discourse (exam scripts) of native and non-native speakers of English with a corpus totaling approximately 1,000,000 words.4 They found that the non-native speakers failed to employ epistemic commitment by representing a more authoritative stance in English whereas the L1 writers of English portrayed a more balanced presentation of their degree of commitment/detachment. In particular, one of their most significant results was related to the relationship between tone of writing and band scores of the L2 students: the lower the band of the student, the more authoritative and the less tentative the voice. Nevertheless, this finding has not been linked to any potential cultural transfer issue (if any) as L1 texts of these writers were not included in their study. Vassileva’s (2001) crucial study elegantly highlighted the general routes of the expert writers of English (L1), Bulgarian (L1) and Bulgarian English (L2) in terms of the notions of commitment and detachment by limiting the study to three important parts of research articles, the Introduction, Discussion and Conclusion. Despite the concept of interlanguage 4 As the participants were from different contexts and the writing tasks seemed to be not identical, such issues can reduce the validity of the comparable corpus. 835 EDUCATIONAL SCIENCES: THEORY & PRACTICE theories, the overall results revealed that the English L15 and Bulgarian L1 texts appeared to have more hedges resulting in detachment compared with the Bulgarian English texts. Regarding commitment, the Bulgarian English texts seemed to present a highly authoritative style with far more boosting devices than hedges whereas the English L1 writers preferred to construct a more tentative discourse in negotiating knowledge claims. What is more interesting is related to the extraordinary route of the Bulgarian English (L2) writers’ texts: they started with a highly committed style and closed with an intensely hesitant style. In other words, they seemed to rely on the convention of employing highcertainty resources –that is, boosters– in their introductions and discussions; however, they offered relatively more softened and tentative knowledge claims in their conclusion sections. This cross-sectional analysis suggested that Bulgarian English writers simply contradicted what English L1 writers did and Bulgarian (both L1 and L2) expert writers equipped their knowledge presentations with a rather assertive nature in general, which highlighted a cultural tendency of Bulgarian writers. Exploring the effect of culture on the avoidance of uncertainty or on employing certainty markers, Koutsantoni (2005) attempted to characterize rhetorical variations across three groups, Greek L1, Greek speakers of English and English L1, in the field of engineering by looking at research articles and conference papers.6 The inclusion of L1 texts to understand the nature of the L2 texts contributed to the finding of the Greek writers’ (L1 and L2) high-certainty style and confirmed that the English L1 writers avoided making too authoritative claims in their discourses with fewer boosters. However, a potential question to be addressed is linked to the idea of whether a writer’s authoritative and high-certainty style can be explored by ignoring the notion of weakening claims with hedges in the data of the study. It could be the case that the Greek writers balanced their willingness to express their certainty by hedging their bets through uncertainty devices in their articles; but it is hard to draw such a conclusion as the researcher did not search for such expressions in her corpus. This is why the current study treated certainty and uncertainty equally and merged them to probe the phenomenon of expressing commitment/detachment. Considering the previous research designs (mostly on expert texts and exploring only one side of the coin) and issues (ignorance of L1s, mismatch between genres, and groups), a relatively detailed study is deemed to be essential to determine how language and culture can give a direction to the writing conventions of different 5 The articles gathered from English L1 speakers were from British and American writers, and this could be quite speculative in a study in which the researcher is attempting to differentiate conventions across cultures as British and American writers may potentially follow different rhetorical strategies. 6 The data of Koutsantoni (2005) seemed to be troublesome and not representative as the three data sets did not match very well in order to be comparable, when comparable corpus design by Moreno (2008) is taken into account. There were research articles written by the English L1 and Greek speakers of English whereas the Greek L1 texts included unpublished conference papers which were four times shorter, according to the given numbers of lines in her study. In addition, the data collected from the English L1 writers did not seem to belong to one group of writers, as happened in Vassileva’s (2001) study and was highlighted in the previous note. The English L1 texts were from British, American, Australian and Canadian English speakers and they were treated as native texts although the indicated group of L1 writers might also follow different conventions in scientific writing and it could be quite hard to draw conclusions by treating them as the same group. 836 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... groups regarding expressing certainty towards propositions. Before elaborating on the methodological considerations of the present study in the next section, it is useful first to stress that ‘the postgraduate writers’ labelled in the present study consisted of a representative sample of English L1 writers from the UK, Turkish L1 writers from Turkey and Turkish writers of English from Turkey at masters’ level. Corpus and Methodology of the Study The intention in this section is to shed light on a range of significant points and considerations in the data collection, analyses (pilot and main analyses) as well as an analytical framework. The Research Procedures of the Study The present study was exploratory in nature and a triple comparative approach was used for investigating how postgraduate students from different discourse Figure 1. Overall view of the research procedures (Akbas, 2014b). 837 EDUCATIONAL SCIENCES: THEORY & PRACTICE communities qualified their commitment or detachment in their academic writing. Figure 1 shows the overall research design followed while carrying out the PhD research (Akbas, 2014b), starting with building the corpus of the study and ending with comparisons across the groups. As can be seen, the study followed reasonably detailed steps in order to achieve a better understanding of the phenomenon under investigation. The study and consequently the findings gained greater significance after a pilot analysis using Nvivo 10 and a way of compiling a list of linguistic resources unique to the postgraduates, and the application of second-coder analysis in order to provide solid and consistent results. Corpus of the Study and the Research Question7 Relying on a combination of the comparable corpus design by Moreno (2008) and maximum similarity across sub-corpora by Chesterman (1998), a corpus consisting of 30 discussion sections from successfully completed master’s dissertations (in the Social Sciences) for each sub-corpus was compiled. The thesis center in Turkey (http:// tez2.yok.gov.tr) was used to download the open-access dissertations of Turkish L1 (TL1) and English L2 (EL2) with a traditional format8 in order to include sections with the same communicative purposes. The English L1 (EL1) texts of British students who had studied in the UK, on the other hand, were accessed through White Rose eTheses Online (http://etheses.whiterose.ac.uk) as well as personal contacts using the snowballing method. Then the discussion sections of the dissertations collected were separated. As shown in Table 1, the sub-corpus of Turkish L1 (Tcorp) writers had 71,581 words, the English L2 (TEcorp) texts had 122,161 words and the English L1 (Ecorp) texts had 102,361 words, making a total corpus of nearly 300,000 words. Table 1 The Sizes of the Sub-corpora (TL1, EL2 and EL1) Total number of words Average number of words Tcorp 71.581 2386 TEcorp 122.161 4072 ECorp 102.361 3412 Average number of sentences 103 159 126 Considering the fact that since the emergence of the concept of genre there have been many studies (Akbas, 2012; Bruce, 2014; Cakir, 2016; Hu & Cao, 2015; Gillmore & Millar, 2018; Kafes, 2017; Karahan, 2013; Martı́n, 2003; Samraj, 2002; Tanko, 2017; Tessuto, 2015) which have looked at the rhetorical organizations of scientific writing in general or have focused on particular sections of genres (such as research article abstracts, introductions and results), there has been relatively less attention 7 This study used the corpus of a successfully completed PhD project at the University of York (UK) in 2014, titled ‘’Commitment-detachment and authorial presence in postgraduate academic writing: A comparative study of Turkish native speakers, Turkish speakers of English and English native speakers’’ 8 The traditional format comprises “Abstract, Introduction, Literature, Methodology, Results, Findings, Discussion, and Conclusion or similar communicative purposes with different labels” (Akbas, 2014b, p. 78) 838 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... given to the exploration of the nature of discussion sections (Akbas, 2014a; Akbas & Hardman, 2017; Basturkmen, 2009; Hopkins & Dudley-Evans, 1998; Samraj, 2013) in dissertation/thesis writing. For this reason, only discussion sections were chosen to be examined in the present study. The rationale behind limiting the study to discussion sections is also strongly linked to a few factors. Discussion sections, in the traditional dissertation format, are significant in the sense that the stance and voice of the author are unique, given the communicative purpose of the section. Other parts of the dissertations are generally designed with a role more of giving information and reporting. To illustrate, introduction sections are structured to introduce the research and the topic in a brief way; literature sections generally review and report what is already in the literature in a critical manner, whereas the writers are expected to exhibit the interpretation of the results and present potential links critically to the available literature to discuss the findings. This essentially requires a relatively more persuasive and communicative style on the way to justifing claims before a writer finalizes his/her study with a proper conclusion section. The research questions designed to be answered through this exploratory study were: RQ1. How do postgraduate students (L1 writers of Turkish, of English, and Turkish writers of English) display their commitment/detachment towards their propositions in their academic texts? 1.1. What are the most commonly employed linguistic means of qualifying commitment/ detachment in the postgraduate texts? 1.2. Are there any similarities or differences across the groups in terms of commitment/ detachment in achieving different discourse acts? As already explained, the investigation of discussion sections in the present study is heavily based on the fact that writers are expected to present a unique stance and voice in their discussion sections in comparison with the other parts of the dissertations which have a more informative nature. In addition, other parts of academic texts, such as abstract, introduction and conclusion, have received a great deal of attention since the notion of genre appeared, which has left the discussion sections in academic writing relatively unexplored. In the next section, we shall give details of the analytic framework followed in the present study in order to highlight how linguistic items were categorized for the pilot and the main analyses. Analytic Framework One of the earliest models of certainty categorization was that proposed by Holmes (1982; 1984), who set out a scale of linguistic resources as Certain, Probable and Possible and used this to indicate the level of commitment shown by writers. Rubin, Liddy, 839 EDUCATIONAL SCIENCES: THEORY & PRACTICE and Kando (2006) added two extremes to Holmes’s continuum (absolute certainty and uncertainty) in order to explain “certainty” in English. Taking into consideration both the continuum which Holmes used to categorize a wide range of linguistic devices and the modification made by Rubin et al. (2006), a broader approach was preferred in this current study for reassessing the level of certainty conveyed by a writer by using a free and not a predetermined scale. This was mainly due to the fact that the present research involved two languages (Turkish and English) and it was considered that a broader conceptualization would contribute to our understanding of how writers modify the illocutionary force by emphasizing or weakening it. A great many terms have been used to refer to such strategies of language use by writers, such as “hedges and boosters”, “certainty markers”, “downtoners”, “epistemic modality”, and “emphatics”. The connection between such labels and concepts demonstrates that they are mainly used to signal a particular degree of commitment/ detachment even though there are linguistic devices which can be used to highlight vagueness, tentativeness, uncertainty or positive/negative politeness strategies. The approach followed in this study for exploring such strategies resulted in a broader distinction of the phenomenon of commitment/detachment, as is also suggested by Stubbs (1986), formulating the propositional content with varying degrees of certainty ranging from (1) very weakened propositions, resulting in detachment from what is presented, to (2) very assertive propositions, resulting in a higher commitment of the writer. Following the concept of hedges and boosters established by Aull and Lancaster (2014), various ways of enhancing or mitigating the propositional content can well be explained by the concept of commitment/detachment to indicate interpersonal functions, as each linguistic resource can signal a noticeable degree of commitment or detachment. The examples below (cited from Akbas, 2014b, p. 36) provide a straightforward clarification of what is proposed: (1) The bird flu might spread if the infected birds are shedding the virus in their nasal secretions. (2) It is clear that the bird flu will spread when the infected birds shed the virus in their nasal secretions. As can be seen, the information in (1) and (2) has clearly been packaged and qualified in two different styles in two context-free sentences. It is highly plausible to interpret the propositional content within its authentic context; however, by analyzing the linguistic choices in the examples signaling certainty/uncertainty, we can propose that example (1) (modified by a modal verb and a conditional sentence) demonstrates a rather detached stance towards the truthfulness status of the proposition in comparison with example (2) (modified by the expression it is clear that) displaying a higher degree of certainty and commitment. In line with this perspective, Akbas (2014b) followed 840 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... Crismore, Markkanen, and Steffensen (1993) and used a cline for the sake of locating propositions signaling commitment/detachment, as shown in Figure 2. Figure 2. ‘The cline of commitment and detachment’ of Akbas (2014b). It can be a reasonable argument to state that propositions with varying degrees of detachment through hedging resources are likely to exhibit a partial level of commitment; nevertheless, the leading goal of the author can be assumed to be withholding the proposition by the use of linguistic clues. Following this, on the one hand, the propositions are classified and labelled as ‘commitment’ with the help of linguistic resources, namely boosters, when the author deliberately flags a higher degree of commitment to present “assured and reinforced information” or to “make his/ her perspective prominent within the discourse appeal to the reader’s attention” (Akbas, 2014b, p. 110). In such cases, the author is assumed to take full responsibility for what is being asserted rather than mitigating it, as is shown in the following example: (3) It is evident that each participant has developed both their classroom practice and their organizational presence and confidence significantly since starting their course.9 On the other hand, with the help of linguistic choices of hedges by the writers, the propositions were labelled “detachment” when the writers deliberately toned down their assertions to signal hesitancy or a lack of certainty and confidence. Such a strategy of disclosing one’s distance from a higher level of commitment simply places the propositions in the middle of the scale or closer to “full detachment”, as shown in Figure 3. This helps writers to present opinions rather than actual information, to show complete or a little doubt and hesitancy over the content and to open up other possibilities and voices for achieving dialogic expansion. The following example (4) indicates how a writer showed detachment from the proposition in order to implicitly underline that the claim seemed good-looking but might stay unproven because of potentially inadequate evidence. (4) Overall, the data would suggest that all participants provided an adequate and relatively comparable learning experience, using Mohan, Leung and Davison’s (2001) suggestions for evaluation. The occurrences signaling various degrees of commitment and detachment based on the cline were categorized by analyzing the contexts as they appeared qualitatively, 9 Due to inexperience, some postgraduate writers may prefer to sound more confident in order to make the reader accept what is presented as taken for granted without supplying enough evidence to support the knowledge claim. However, this study did not intend to evaluate how appropriately the writers used the evidence to support the knowledge claim, but focused only on explicit markers to indicate the truth-value of the propositions in postgraduate writing. 841 EDUCATIONAL SCIENCES: THEORY & PRACTICE and this led to a better differentiation and grouping of what such linguistic means has accomplished within the texts of postgraduate writers. A Closer Look at the Pilot, Second-coders and Main Analyses Initially, a random-sampling manual analysis –with seven texts from each group, twenty one texts in total– was applied in order to identify and categorize the means of hedges and boosters in the sub-corpora of the study. With the help of computer-assisted qualitative data analysis software (Nvivo 10), a systematic and detailed contextual analysis assigning nodes to different occurrences was completed manually. Not only did the manual analysis enable us to come up with a reliable list of search items to be used in the main analyses of the whole corpus, it also gave the opportunity to test the items in their contexts in order to examine whether they performed the functions looked for. Rather than compiling a list of potential linguistics devices functioning as hedges or boosters from the available literature, this more efficient way helped to create a list of items unique to the corpus of the study, employed by postgraduates, to use in the main investigation of the study. Before the main analyses, an intercoder analysis was considered to be fundamental in order to validate how effective the coding system and categorization would be. This was simply because it was crucial to decrease the subjectivity of the assessment of the decisions of the linguistic items performing hedge and booster effects as well as the categorization in the researcher’s coding system. Therefore, as shown in Table 2, five people with previous experience in corpus studies as researchers/second coders were invited to participate in an inter-rater reliability phase and to code sample extracts. A codebook was developed for this purpose as the second-coders needed some training in the notions of the research and the coding scheme. These five independent coders were asked to code instances, in total 700, within the original context of these extracts, and the coding process was completed online after they had studied the codebook. Table 2 Intercoder Agreement Results Regarding Commitment-detachment Coder 1 & Coder 2 & Coder 3 & Coder 4 & Coder 5 & All Coders & Researcher Researcher Researcher Researcher Researcher Researcher Number of extracts 100* 150** 150** 150** 150** 700 Matched choices 87 135 143 136 141 642 Unmatched choices 13 15 7 14 9 58 Agreement on choices 87.0% 90.0% 95.3% 90.6% 94.0% 91.7% (%) Cohen’s Kappa 0.736 0.798 0.906 0.813 0.879 0.826*** Agreement Significance 0.000 0.000 0.000 0.000 0.000 0.000 * Coder 1 coded only English extracts (100) ** Coders 2 to 5 coded both English and Turkish extracts (150) ***The kappa was computed by comparing the arithmetic mean of all coders with that of the researcher, as suggested by Light (1971) 842 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... As suggested by Lombard, Snyder-Duch and Bracken (2005), the evaluation of the independent coders for the sample extracts was compared with our identification system in order to calculate the intercoder agreement. The overall agreement among the independent raters and the researcher indicated the consistency of the values or functions assigned (Green, 1997). In line with this, we carried out Cohen’s kappa statistics for calculating the degree of agreement, instead of just relying on the simple percentage of matches among raters. As revealed by the intercoder agreement results shown in Table 2, there was a promising agreement both among raters and between raters and the researcher. This result not only validated the consistency among the raters in assigning linguistic items with their functions but also showed that our categorization was sufficiently reliable and practical to use for the main study. Figure 3. The immediate context analysis for “may” in the data. The main analyses of the corpus in relation to linguistic means of signaling commitment/detachment were conducted using WordSmith Tools 5.0 and the compiled list of linguistic devices from the pilot study. Detailed analyses were then conducted to validate whether the items performed the functions of hedges or boosters. A closer examination of the occurrences was essential for identifying lexical items of certainty, uncertainty or none. As can be seen from Figure 3, may was used as a hedging resource to weaken the force of the claim; nevertheless, there were some cases of May as the month of the year and they needed to be excluded in order to finalize the raw number of occurrences in each group. There was a range of comparisons across groups from a quantitative perspective in relation to observed and normalized distributions, mean frequency and statistical differences. In order to see whether the differences were statistically significant, a non-parametric test, the Kruskal Wallis test, was used and statistically significant or not significant results across the three groups were determined. Nevertheless, the test 843 EDUCATIONAL SCIENCES: THEORY & PRACTICE did not reveal which group had caused the significant results. In consequence of this, regarding the statistical analysis, there is also a further point to be considered and the groups were regrouped by the variables of culture and language and compared using the Mann-Whitney U test. In other words, another test was applied to the groups sharing the same culture as opposed to the other culture (Turkish L1 writers + Turkish writers of English as opposed to English L1 writers) and the groups writing in the same language as opposed to the other (Turkish L1 writers as opposed to Turkish writers of English + English L1). An equally significant aspect of the comparisons across the groups was related to the qualitative consideration of the findings. Many researchers (Hyland, 1996, 1998; Martin-Martin, 2008) have managed to identify some functions of hedges (such as signaling a lack of complete commitment) and boosters (such as indicating higher confidence) in academic prose; nevertheless, it could be rather difficult to connect particular functions with the linguistic expressions as far as the polypragmatic nature of resources is concerned. Noting the compelling nature of this, a relatively bold strategy was applied: stressing commitment or decreasing it to signal a lack of commitment (resulting in detachment). With the assistance of this perspective, it became possible to identify a pattern in relation to how authoritative the three groups of postgraduates sounded while accomplishing pragmatic functions in the discussion section. Results and Discussion In the light of the discussions in the previous sections, this section now presents the quantitative and qualitative results of the commitment/detachment choices made by postgraduates while performing discourse acts along with a discussion of the findings and responses to research questions. Quantitative Findings The quantitative analysis of the linguistic items indicated that Turkish writers of English employed relatively more hedges and boosters (24.2 per 1000 words) to strengthen or weaken the knowledge claims in the discussion sections. Not only did the Turkish L1 writers differ in terms of relatively fewer instances in general but also they seemed to favor a completely different style in comparison with the English L1 and L2 writers. As can be seen in Figure 4, similar to the results of Akbas (2012), the Turkish L1 writers mostly preferred to present their knowledge claims with a more definitive and authoritative nature through the use of more boosters than hedges whereas the balance of hedges and boosters in the discussion sections of the English L1 and L2 writers was observed to be greater with respect to hedges. In other words, the latter group of writers showed a more tentative style in presenting knowledge claims. Even though the place of interlanguage users (EL2) regarding the 844 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... use of boosters was somewhere between the Turkish L1 and the English L1 writers, it should be noted that the case of the Turkish writers of English in signaling academic modesty seemed to be different, with a greater number of hedging resources compared with their linguistically-linked peers (EL1). Figure 4. Mean frequency of hedged and boosted sentences (per 1000 words). In terms of commitment signaled by boosting resources, the Kruskal Wallis test revealed that there was a statistically significant difference across the three groups of postgraduates (H (2): 22.198, p = .00); however, it was not clear which group(s) created the significant difference. Therefore, two Mann-Whitney U tests were run over the data grouped as cultural (T1+T2 vs E1) and language (T1 vs EL2 + EL1) pairs to determine whether any of these variables had had an influence on the use of boosters across groups. According to the results of the Mann-Whitney U tests based on the culture variable, a statistically significant difference was found between Turkish postgraduates (TL1 and EL2) and British postgraduates. Similarly, the test regarding the language variable also resulted in a statistically significant difference between Turkish L1 writers and English L1 and L2 writers. Considering both of these results, it is highly possible to say that it was the Turkish L1 writers who caused the difference across the groups regarding the resources used for strengthening the claims. This slightly higher use of boosters over hedges by the TL1 writers constructed a distinctive style in their discussion sections. Under research for almost four decades, the concept of weakening a propositional meaning seems to help writers to achieve a variety of rhetorical functions ranging from stating doubt to academic modesty and avoiding preciseness. Martin-Martin (2008) clearly stated that hedges as the linguistic means of such functions contribute to the voice of the authors of the texts. In line with this, the results of the current research highlighted that both the EL2 and the EL1 writers preferred to follow a more 845 EDUCATIONAL SCIENCES: THEORY & PRACTICE detached way of qualifying their claims with the help of varied means such as modal verbs (5-6), full verbs (7) or some formulaic expressions (8-9). (5) Applying these two ideas to the situation in Greater Manchester, we might expect linguistic features to spread from urban Manchester/Salford to the suburban towns of Greater Manchester (EL1-12). (6) So, this may lead teachers to soften or change their comments related [to] children in [the] evaluation part (EL2-29). (7) This finding seems fairly reasonable as far as [the] characteristics of the region are taken into consideration (EL2-25). A range of formulaic sequences (such as “is likely due to”, “it is possible that”) appeared to signal the perspective of the writers towards the accuracy of the information, no matter how restricted such uses were to a combination of a few adjectives or nouns to create multi-word units as in: (8) In an evolved network, it is probable that two similar agents possessing similar traits belong to a similar group –i.e. are close together in the network (EL1-9). (9) From the perspective of teachers, lack of science centers and related materials prevents them from properly implementing science activities (EL2-29). Interestingly, however, both groups employed modal verbs so frequently as the major means of expressing detached meaning that the use of modal verbs in the EL2 and EL1 texts constituted more than half of all hedged sentences (53.1 % and 51.5% respectively). Conversely, for the TL1 writers, a particular suffix (-ebilir/-abilir as in (10) below) used for expressing detachment accounted for more than 70% of all hedging cases in the Turkish sub-corpus (5.72 per 1000 words). This can be linked to what (2018) discussed with respect to multi-functional linguistic items with relatively more precise semantic meanings in L2. Since the rest of the hedging instances were very limited, this can indicate a relatively monotonous style of marking tentativeness over knowledge claims by the TL1 writers. (10) Özetle şiddet içerikli bilgisayar oyunu oynayan oyuncu “bir başkası” tarafından engellendiğini düşünüp daha fazla stres yaşamış olabilir.10 (TL1-10). (A player, especially playing a computer game containing violence, “can may/might have had” more stress by thinking s/he was stopped by “anyone else”.) As can be seen in Figure 5, similar to the case with hedging resources, the Turkish L1 writers relied heavily on suffixes (–mIştIr, -mAktAdIr, -AcAktIr) rather than lexical words (8.8 per 1000 words), without leaving any room for the reader to form an opinion, to close down any other potential interpretations and boost the knowledge 10 The suffix –ebilir/-abilir in Turkish is represented by “can”, “could”, “would”, “may” or “might” in English with different strengths of epistemic meaning. However, as this is a translation of the original extract, it is thought that it should be free of bias. 846 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... claims presented as reliable. A typical example of such an assertive tone is presented in (11), which simply strengthens the illocutionary force with a definite meaning and indicates that the writer prefers to sound authoritative: (11) Bu nedenle yukarda saydığımız eksiklikler bir an önce çözülmesi mevcut kaygı ortamının da ortadan kalkmasını sağlayacaktır (TL1-26). (For this reason, correcting the deficiencies, which are stated above, immediately “is going to enable” the present anxiety environment to come an end.) Figure 5. Linguistic realizations of boosters across groups (per 1000 words).11 In contrast, for both the EL1 and the EL2 writers, there was a widespread preference for lexical verbs over other means of expressing certainty such as auxiliary verbs, adverbs or multi-word constructions. Among many, the three most employed epistemic verbs in these groups were find, show and reveal to signal a notably higher level of commitment about the knowledge claims, as in (12) and (13): (12) The study did find that overall experience was a predictor of teaching efficacy and that with more experience efficacy increased (EL1-1). (13) The study revealed that materials provide the basis for language input, and choosing the materials is a vital phase of curriculum planning (EL2-26). When summing up the quantitative findings and means of expressing commitment/ detachment, we should note that the three groups of postgraduates involved in this current study showed different tendencies in producing knowledge claims and negotiating them with the intended reader. We reached a statistically significant keycontrast between the Turkish postgraduates (TL1 and EL2). This indicated that the Turkish writers of English sounded more tentative and withheld their commitment towards their propositions whereas the TL1 writers preferred to qualify a considerably higher level of commitment. In addition, the tone of the EL2 writers appeared to be 11 Modal auxiliaries in Turkish do not occur as in English; however, some of the suffixes from Turkish are included as modal auxiliaries for comparison purposes. 847 EDUCATIONAL SCIENCES: THEORY & PRACTICE reasonably similar to that of the EL1 writers, deviating from their counterparts writing in Turkish. On the other hand, the findings regarding the EL1 writers are in line with the idea (Atai & Sadr, 2006; Hyland, 2005) that native writers of English tend to employ more hedging resources so as to protect themselves from potential reader-criticism. Qualitative Findings Noting the difficulty of matching linguistic choices with specific acts, linking the hedges and boosters in the sub-corpora of the study to particular functions seemed to be relatively easier with respect to the rhetorical features of the discussion sections. In order to boost the quantitative results of the study, the linguistic occurrences signaling commitment or detachment were also analyzed with a special focus on their pragmatic functions accomplishing similar discourse acts. Keeping the communicative essence of discussion sections in mind, a qualitative analysis of occurrences was implemented and a range of discourse acts, for which postgraduates favored committing or detaching themselves, were identified. The commonly performed discourse acts12 with varying degrees of commitment/detachment in the discussion sections of postgraduate writers consisted of (i) presenting and interpreting the results, (ii) evaluating previous research findings and comparing results, (iii) promoting research and particular findings, (iv) mentioning methodological considerations, and (v) elaborating an argument. Presenting and interpreting the results. The first of the moves identified in the corpus, namely, presenting and interpreting the results, is thus of great importance. However, the way in which the postgraduates increased or decreased their levels of certainty as to presenting and interpreting the results has a vital role in the acceptance of the knowledge claim by the readers. In order to achieve ratification by the target audience, the writers may prefer to modify the certainty degree of the knowledge claims which they present depending on the evidence with which the propositions are put forward. According to Varttala (2001), the nature of the knowledge claims in a discussion sections calls for them to be relatively hedged by means of the linguistically detached stance taken by the writers. This is essentially in parallel with the idea that the section seems to have a dynamism of speculative inferences leading to further conclusions drawn from the data. Example (14) illustrates how the writer managed to present his/her proposition as “left open to readers’ judgement” (Hyland, 1998, p.182) in order to flag a lower level of certainty in rationalizing a particular case: (14) It seems to me that the more proficient L2 group was indeed exposed to negative evidence in certain ways, i.e. in class or through explicit instruction, but the low level L2 learners were not. I propose that the low level L2 learners may be making use of the Avoidance Strategy (Dörnyei & Scott, 1995a, 1995b) (EL2-1). 12 The rhetorical discourse acts were based on the preliminary examination of the sample texts in the pilot study and realised in the texts to develop a convincing overall argument, through the discussion of findings and elaborating claims. 848 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... It is also evident that the writer in the example above developed his/her stance towards interpreting a very specific case by highlighting his/her putative contribution to the readership without marking it as factual. Even so, there have been various instances in relation to indicating a higher degree of commitment while making a deduction confidently, as in (15): (15) Clearly, this is a practical and understandable tendency as their initial teaching responsibilities will be relatively narrow and contained (EL1-2). Evaluating previous research findings and comparing results. Regarding evaluating previous research findings and comparing results, the data suggested that the TL1 writers overwhelmingly treated knowledge claims rooted in the literature as accepted factual information. To put it another way, as (16) shows, the level of certainty and assurance with respect to the work of others in the field was substantially higher compared with the English L1 and L2 writers: (16) Hand ve Prain (2002) konuyla ilgili yaptıkları çalışmada, yazmanın kavramlara ilişkin yanılgıları ortaya çıkardığı ve kavramsal öğrenmeyi gerçekleştirmede etkili olduğu sonucuna varmıştır. Benzer şekilde Bulloc (2006) yaptığı çalışmada, yazmanın kavramlara ilişkin ön bilgileri ortaya çıkardığını ve kavramsal değişimi sağlamada etkili olduğunu ispatlamıştır. Reilly (2007) de öğrencilere matematik dersinde konu ile ilgili yazı yazdırılmasının, öğrencilerin matematiksel kavramları öğrenmelerinde büyük katkı sağladığını ortaya çıkarmıştır (TL1-29). (Hand and Prain (2002) revealed in their study about this subject that writing reveals delusions/errors regarding concepts and it is effective in performing of conceptual learning. Similarly, Bulloc (2006) proved in his study that writing reveals foreknowledge about concepts and it is effective in providing conceptual change. Reilly (2007) revealed that making students write about the subject in maths classes contributes hugely in learning mathematical concepts.) The discussion sections of the TL1 writers seemed to have a dominant use of the same pattern as in (16) by evaluating previous studies by means of the use of strong positioning verbs combined with suffixes in Turkish, such as ispatlamıştır (proved that), ortaya çıkarmıştır (revealed that) and sonucuna varmıştır (concluded that). This significantly leads to their presentation of reported content as if it were accredited knowledge, rather than strategically indicating a weaker positioning towards it. In contrast, the EL1 and EL2 texts seemed to operate widespread use of speculative language in order to imply that the content reported is somewhat true, but that the authors are relatively hesitant about it. The examples below (17-18) clearly demonstrate how the authors tackled the presentation of other people’s work by underlining their detachment from the source claims, which produces a conveyance of a partial agreement. (17) Although related literature tended to report somewhat similar results, slight variations can be found with respect to age, socio-economic status, values, culture, location, occupations, and knowledge about environmental issues (EL2-25). 849 EDUCATIONAL SCIENCES: THEORY & PRACTICE (18) Huckle (2008) argued upon four reasons why he believed New Labour’s eight doorways had failed. I would agree with Huckle to a certain degree that the systems of competition and privatization at face value do appear not to promote sustainability (EL1-25). Promoting research and particular findings. The help of hedges and boosters, as noted by Halliday (1978), in building a firmer relationship between writer and reader can also be considered to be highly key to another discourse act identified as promoting research and particular findings. Nevertheless, both the TL1 and the EL2 writers were in favor of self-promotion whereby they relied on particular lexical verbs to create a sense of conviction as to what they had achieved: (19) I have shown that causal wh-phrases in Turkish have a weakening effect on intervention effects. Furthermore, I have shown that lexically marked focus phrase with the focus particle sadece “only” provides evidence for Göksel and Özsoy’s (2000) claim (EL2-6). (20) All these findings clearly reveal the need for a change in function of environmental education from just transmitting ecological knowledge to bringing out the emotions of learners regarding the value of nature and its elements (EL2-23). The way in which the Turkish writers created such a sense of conviction contributed to a less tentative discourse with no reservation to hesitation when presenting their research and findings to their discourse community. Even so, as exemplified by (21) and (22), the English L1 writers signaled comparably more mitigation for the purpose of politeness in the course of expressing what their research had achieved: (21) This study was an attempt to explore the use of wikis in L2 academic writing workshops (EL1-11). (22) This study attempted to simulate very simplistic models of language contact situations in groups of artificial agents (EL1-7). In addition, for the EL1 writers, it was another prevailing use of hedging resources to decrease the force of the propositions while coming up with a non-factive reasoning for disproving previous results, views and/or hypotheses: (23) I suggest that the reasons for this are that these dyads had established a successful method of constructing tangrams without the need for dialogue; therefore, introductory mentions of referents in the speech part of the experiment did not need to be as intelligible, and this hypothesis is rejected. (EL1-17) Mentioning methodological considerations. Another evident discourse act identified throughout the corpus was mentioning methodological considerations, in which the postgraduates directly or indirectly evaluated their study in terms of the methodology (the participant(s), method or approach) in order to open up a dialogue for recommendation for further research. Signaling their level of commitment or detachment by linguistic realizations also played a vital role in coding the information 850 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... for the readership. As an example, (24) attempts to indicate that the qualitative findings to some extent were fallible because of the non-existence of some methodological issues, which can be, in a way, considered as a suggestion to future researchers: (24) The classroom practices of more teachers working at different educational settings and with different student levels may provide us with more insights about their beliefs and practices. In addition, the qualitative findings of the study could have been more reliable if teachers kept diaries and the observed lessons were video-recorded and then followed by a think-aloud procedure (EL2-14). The writer in (24) did not refer to this methodological consideration sufficiently strongly to be protecting his face; instead, the recommendation-oriented self-criticism was produced by displaying tentativeness about what ‘could have been’ achieved and not committing him/herself to such a consideration. There were similar cases in the texts of the Turkish L1 writers, when the writers chose to be rather less confident in explaining methodological considerations and their outcomes. For instance, the following example (25) can be presented as an illustration of how a writer linked a particular finding to a previous methodological consideration by being tentative enough to get ratified: (25) Öğrenmenin kalıcılık düzeyinin deney grubu lehine olmasının beşinci nedeni olarak araştırmada ontest ve sontest olarak kullanılan akademik başarı testinin kalıcılık testi olarak da kullanılması gösterilebilir (TL1-22). (The fifth reason why the experimental group had a higher level of permanence of learning “can/could/would/may/might be explained” by the academic success test which was used as pretest and posttest and also used as permanence test in the research). On the other hand, among very few examples in the EL1 data, some postgraduates sounded highly confident so as to strengthen the truth value of the propositions and appeal to the target reader’s acceptance. To illustrate, the next excerpt (26) explicitly demonstrates the commitment of the writer to the way of asserting his/her projection by employing a strong verb followed by a construction boosted with an auxiliaryverb pattern, that is “do vary”: (26) Through conducting a range of biographical case studies with people across different age groups I have established a number of areas where influences on career choices and aspirations do vary across generations, and also some areas where these differences are less obvious (EL1-22). Elaborating an argument. The last of the strategies found in the discussion sections of the postgraduate texts under investigation here is elaborating an argument. That is, the postgraduates in the study attempted to create a rhetorical effect in the text through the employment of markers signaling certainty and doubt in their claims. This component of the discussion section is relatively essential for writers to be able to gain the credibility of the audience by means of presenting their knowledge 851 EDUCATIONAL SCIENCES: THEORY & PRACTICE claims as needing to be decoded and accepted as possible depending on the tone of the claims. As highlighted by Akbas (2014b), postgraduate writers can provide their readers with the “established and confidently presented knowledge claims to contentiously worded and low committed opinion-based claims in the texts” (p. 183). There was a striking difference among the groups in elaborating arguments which can lead to building a disciplinary knowledge; the TL1 writers mostly sounded quite authoritative by relying on more boosters to signal a confident tone (as in 27) whereas the EL1 and EL2 writers marked their involvement in as detached a way as possible as in (28) and (29): (27) 2005 eğitim öğretim programının uygulamaya başlamasının üzerinden her ne kadar 5 yıl geçmiş olsa da hali hazırda daha yapılandırmacılık yaklaşımını tanımayan öğretmenler bulunmaktadır. Bu da hizmet içi eğitim ile bu açığın en kısa sürede kapatılması gerektiğini göstermektedir (TL1-26). (Although it has been five years since the application of the 2005 educational curriculum, currently there are teachers who do not recognize the constructivist approach. “This shows that” there is an urgent requirement for eliminating the deficit with in-service training.) (28) The problems that preschool teachers face in the curriculum implementation showed no significant difference with respect to preschool teachers’ educational level. This situation may be due to [the] level of education studied, in other words, it is a consequence of dealing with early childhood education (EL2-29). (29) By looking at this data, it is possible to argue that the use of the online forum affects several aspects of pupils’ opinions and perceptions of learning (EL1-8). The qualitative analysis in this exploratory study with a special focus on discourse acts indicates that the Turkish L1 writers were more prone to producing fairly assertive claims whereas the sub-corpora of the EL1 and EL2 writers preferred to promote a higher level of deference with their more detached style towards presenting claims and achieving the intentions of the discussion section. We can therefore suggest that the quantitative and the qualitative findings are parallel in showing the level of commitment/detachment across the texts of the postgraduates. Concluding Remarks Taking a closer look at the three different groups both quantitatively and qualitatively, we attempted to investigate how the meanings of their knowledge claims were strengthened or weakened. Adapting the view of Varttala (2001), a broader treatment and categorization of hedges (a reduced degree of commitment) and boosters (a strengthened commitment) was explored with the help of morphological (only for Turkish), lexical and multi-word-unit linguistics resources. The data suggest that, when TL1 writers are compared with EL1 and EL2 writers, the ways in which 852 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... the writers qualified their level of commitment or withheld it to signal detachment showed major differences. Hyland (2000) raised our awareness with respect to the contribution of hedges and boosters to creating a more interpersonal discourse; in line with this, it seems that the Turkish L1 writers (19.1 per 1000 words) created a discourse which was comparatively less interpersonal than those of the EL1 and EL2 writers (20.4 and 24.2 per 1000 words respectively). With respect to signaling commitment or detachment, the TL1 and EL1 writers constructed overwhelmingly divergent academic prose texts. In particular, it was surprising to find that the EL2 writers, despite sharing a cultural background with the TL1 writers, seemed to favor a significantly more modest tone of expressing their knowledge claims, similar to what the EL1 writers did by employing fewer boosters and signaling commitment to the propositions. In other words, it is obvious that the tone adopted by the EL2 writers was statistically detached and more cautious than that of the TL1 writers. This is likely to be linked to a prevalent academic convention descending from Anglophone practices and the potential familiarity of the EL2 writers with such practices through instruction or self-development. This contradicts the best articulated assumption of Contrastive Rhetoric (Kaplan, 2000) in which the rhetorical organization and choices followed by the learners are noted to stem from native culture/language. Thus, the present study contributes to our understanding of how the rhetorical practices of L2 writers of English can be in parallel with the norms in the target language if supported, since it is wise to argue that the EL2 writers in the study were assumed to have been instructed or to have developed themselves in terms of the target language practices/conventions to produce such an important piece of academic writing. In contrast with a few studies (see Hu, Brown, & Brown, 1982; Hyland & Milton, 1997; Koutsantoni, 2005; Vassileva, 2001) in which L2 learners of English were claimed to have constructed a more strongly committed discourse than native speakers of English, the knowledge claims of the Turkish writers of English (L2) at postgraduate level sound fairly detached, as was also found by Onder-Ozdemir and Longo (2014), resulting in developing more tentative epistemic strategies. It appears to be a completely contradictory style in comparison with the style of their peers writing in Turkish. This highlights not only the fact that the EL2 writers showed an awareness of academic conventions divergent from their native language, but also that they adopted it themselves in order to accomplish more interpersonal academic prose for the readership. Implications and Limitations of the Study The writer of an academic text is expected to “construct a pseudo-dialogue with readers in order to gain their acceptance of the argument” (Hyland, 2012, p. 146), for the intended readers to be able to follow in the footsteps of the writer by designing a 853 EDUCATIONAL SCIENCES: THEORY & PRACTICE space for negotiation. One of the ways of doing this is for writers to express caution or confidence over their own propositions, and it has been evidenced by the previous studies that such practices in different academic genres vary. Nevertheless, to the best of our knowledge, there have been very few investigations on Turkish academic discourse, and postgraduate writers have not received a great deal of attention, which has made this study an initial endeavor to shed light on the textual and rhetorical practices of Turkish postgraduate writers as far as the commitment/detachment phenomenon as a part of stance-making is concerned. As can be anticipated, each study has its own limitations; however, before considering the limitations of this current study, it is important to note the implications of the study. Given the significance of constructing stance in written academic prose, a range of authentic academic materials can be designed and used to assist postgraduates as novice writers in the field in order that they can accomplish interpersonal relations in their dissertation writing, especially for L2 writers, who really need guidance in accomplishing the communicative purposes of this particular genre. This is essentially because, as Molino (2018) suggested, “activities that draw from authentic experiences with the aim of stimulating reflection on appropriate uses in specific setting” (p. 952) can contribute to the use of such devices for particular purposes. One of the central issues to be considered here is to let novice writers be aware of the particular practices and expectations of the academic community to which they are about to contribute. In relation to linguistic markers signaling commitment/detachment and revealing stance, Hyland (2000) suggested that “a clear awareness of the pragmatic impact of hedges and boosters, and an ability to recognize them in texts, is crucial to the acquisition of a rhetorical competence in any discipline” (p. 193). This can also be achieved by providing authentic materials with in relation to metadiscourse devices (see Alotaibi, 2018; Bogdanović & Mirović, 2018) so that writers can acquire particular linguistic patterns as well as their functions and integrate them into the rhetorical organization of their own texts. Indeed, there is no shortage of disagreement among scholars that novice writers will simply follow some rhetorical organizations of their native language and culture, and this might sometimes result in the rejection of their style by the intended audience (examiners, in this case). In order to see whether Turkish writers of English follow some rhetorical and linguistic styles of Turkish, more three-angled-research (TL1, EL2 and EL1), as in the current study, is needed. In particular, the more academic work of Turkish writers of English and English L1 writers is scrutinized from different perspectives, the easier it would be to design a course assisting EL2 writers to match their style with that of native writers through potential writing courses comparing practices and general tendencies. Also, a writing course facilitated through corpus-informed teaching would essentially provide various insights by presenting distinctive language practices and applications from authentic texts written by previous novice writers. Considering the advantages of presenting authentic 854 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... language uses and choices in teaching, integrating corpora into teaching academic writing could become an effective instructional tool and a trigger for learner autonomy by making novice writers more aware of discipline-sensitive writing conventions. References Ädel, A. (2018). Variation in metadiscursive ‘you’across genres: From research articles to teacher feedback. Educational Sciences: Theory & Practice, 18, 777–796. http://dx.doi.org/10.12738/estp.2018.4.0037 Akbari, N. (2017). Analysis of stance in the writing of non-native speaker university students in business communication. In C. Hatipoglu, E. Akbas, & Y. Bayyurt (Eds.), Metadiscourse in written genres: Uncovering textual and interactional aspects of texts (pp. 201–222). Bern, Switzerland: Peter Lang. https://doi.org/10.3726/b11093 Akbas, E. (2012). Exploring metadiscourse in master’s dissertation abstracts: Cultural and linguistic variations across postgraduate writers. International Journal of Applied Linguistics & English Literature, 1(1), 12–26. http://dx.doi.org/10.7575/ijalel.v.1n.1p.12 Akbas, E. (2014a). Are they discussing in the same way? Interactional metadiscourse in Turkish writers’ texts. In A. Łyda, & K. Warchał (Eds.), Occupying niches: Interculturality, crossculturality and aculturality in academic research (pp. 119–133). London/Berlin: Springer. http://dx.doi.org/10.1007/978-3-319-02526-1_8 Akbas, E. (2014b). Commitment-detachment and authorial presence in postgraduate academic writing: A comparative study of Turkish native speakers, Turkish speakers of English and English native speakers (Doctoral dissertation). Retrieved from http://etheses.whiterose.ac.uk/7083/ Akbas, E., & Hardman, J. (2017). An exploratory study on authorial (in)visibility across postgraduate academic writing: Dilemma of developing a personal and/or impersonal authorial self. In C. Hatipoglu, E. Akbas, & Y. Bayyurt (Eds.), Metadiscourse in written genres: Uncovering textual and interactional aspects of texts (pp. 139–174). Bern, Switzerland: Peter Lang. https://doi.org/10.3726/b11093 Alotaibi, H. (2018). Metadiscourse in dissertation acknowledgments: Exploration of gender differences in EFL texts. Educational Sciences: Theory & Practice, 18, 899–916. http://dx.doi. org/10.12738/estp.2018.4.0247 Andresen, M. & Heike Zinsmeister (2018). Stylistic Differences between Closely Related Disciplines: Metadiscourse in German Linguistics and Literary Studies. Educational Sciences: Theory & Practice, 18, 883–898. http://dx.doi.org/10.12738/estp.2018.4.0042 Atai, R. M., & Sadr, L. (2006, July). A cross-cultural genre study on hedging devices in discussion section of Applied Linguistics research articles. Paper presented at the 11th Conference of PanPacific Association of Applied Linguistics, Kangwon National University, South Korea. Aull, L. L., & Lancaster, Z. (2014). Linguistic markers of stance in early and advanced academic writing: A corpus-based comparison. Written Communication, 31(2), 151–183. https://doi. org/10.1177/0741088314527055 Basturkmen, H. (2009). Commenting on results in published research articles and masters dissertations in Language Teaching. English for Academic Purposes, 8(4), 241–251. https://doi. org/10.1016/j.jeap.2009.07.001 Bogdanović, V., & Mirović, I. (2018). Young researchers writing in ESL and the use of metadiscourse: Learning the ropes. Educational Sciences: Theory & Practice, 18, 813–830. http://dx.doi.org/10.12738/estp.2018.4.0031 855 EDUCATIONAL SCIENCES: THEORY & PRACTICE Bondi, M. (2008). Emphatics in academic discourse: Integrating corpus and discourse tools in the study of cross-disciplinary variation. In A. Ädel & R. Reppen (Eds.), Corpora and discourse: The challenges of different settings (pp. 31–55). Amsterdam: John Benjamins. https://doi. org/10.1075/scl.31.04bon Bruce, I. (2014). Expressing criticality in the literature review in research article introductions in applied linguistics and psychology. English for Specific Purposes, 36, 85–96. https://doi. org/10.1016/j.esp.2014.06.004 Cakir, H. (2016). Native and non-native writers’ use of stance adverbs in English research article abstracts. Open Journal of Modern Linguistics, 6, 85–96. http://dx.doi.org/10.4236/ ojml.2016.62008 Coates, J. (1987). Epistemic modality and spoken discourse. Transactions of the Philological Society, 85(1), 110–131. https://doi.org/10.1111/j.1467-968X.1987.tb00714.x Crawford Camiciottoli, B. (2010). Discourse connectives in genres of financial disclosure: Earnings presentations vs. earnings releases. Journal of Pragmatics, 42(3), 650–663. http://dx.doi. org/10.1016/j.pragma.2009.07.007 Crismore, A., Markkanen, R., & Steffensen, M. S. (1993). Metadiscourse in persuasive writing: A study of texts written by American and Finnish university students. Written Communication, 10, 39–71. http://dx.doi.org/10.1177/0741088393010001002 Crompton, P. (1997). Hedging in academic writing: Some theoretical problems. English for Specific Purposes, 16(4), 271–287. http://dx.doi.org/10.1016/s0889-4906(97)00007-0 D’Angelo, L. (2008). Gender identity and authority in academic book reviews: An analysis of metadiscourse across disciplines. Linguistica e Filologia, 27, 205–221. Falahati, R. (2004). A contrastive study of hedging in English and Farsi academic discourse (Master’s thesis). Retrieved from https://dspace.library.uvic.ca/bitstream/handle/1828/714/ falahati_2004.pdf Gardner, S., & Han, C. (2018). Transitions of contrast in Chinese and English university student writing. Educational Sciences: Theory & Practice, 18, 861–882. http://dx.doi.org/10.12738/estp.2018.4.0067 Gilmore, A., & Millar, N. (2018). The language of civil engineering research articles: A corpus-based approach. English for Specific Purposes, 51, 1–17. https://doi.org/10.1016/j.esp.2018.02.002 Grabe, W., & Kaplan R., B. (1997). On the writing of science and the science of writing: Hedging in science text and elsewhere. In R. Markkanen & H. Schroder (Eds.), Hedging and discourse: Approaches to the analysis of a pragmatic phenomenon in academic texts (pp. 151–167). Berlin/ Boston: De Gruyter. https://doi.org/10.1515/9783110807332.151 Green, A. (1997, March). Kappa statistics for multiple raters using categorical classifications. Paper presented at the 22nd Annual Conference of SAS Users Group, San Diego, CA. Retrieved from http://www2.sas.com/proceedings/sugi22/POSTERS/PAPER241.PDF Halliday, M. A. K. (1978). Language as a social semiotic: The sociological interpretation of language and meaning. London, UK: Arnold. Hatipoğlu, Ç., & Algı, S. (2017). Contextual and pragmatic functions of modal epistemic hedges in argumentative paragraphs in Turkish. In C. Hatipoglu, E. Akbas, & Y. Bayyurt (Eds.), Metadiscourse in written genres: Uncovering textual and interactional aspects of texts (pp. 85–108). Bern, Switzerland: Peter Lang. https://doi.org/10.3726/b11093 Hatipoğlu, Ç., & Algı, S. (2018). Catch a tiger by the toe: Modal hedges in EFL argumentative paragraphs. Educational Sciences: Theory & Practice, 18, 957–982. http://dx.doi.org/10.12738/estp.2018.4.0373 856 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... Heiniluoma, M. (2008). Boosting future prospects of softening promises of success? The use of emphatics and hedging in the letter to shareholders sections of annual reports (Master’s thesis). https://www.utupub.fi/bitstream/handle/10024/39022/gradu2008heiniluoma.pdf Ho, V., & Li, C. (2018). The use of metadiscourse and persuasion: An analysis of first year university students’ timed argumentative essays. Journal of English for Academic Purposes, 33, 53–68. https://doi.org/10.1016/j.jeap.2018.02.001 Hu, G., & Cao, F. (2015). Disciplinary and paradigmatic influences on interactional metadiscourse in research articles. English for Specific Purposes, 39, 12–25. https://doi.org/10.1016/j. esp.2015.03.002 Hu, Z., Brown, D. F., & Brown, L. B. (1982). Some linguistic differences in the written English of Chinese and Australian students. Language Learning and Communication, 1(1), 39–49. Hunston, S., & Thompson, G. (2000). Evaluation: An introduction. In S. Hunston & G. Thompson (Eds.), Evaluation in text: Authorial stance and the construction of discourse (pp. 1–27). Oxford, UK: Oxford University Press. Hyland, K. (1996). Writing without conviction? Hedging in science research articles. Applied Linguistics, 17(4), 433–454. https://doi.org/10.1093/applin/17.4.433 Hyland, K. (1998). Hedging in scientific research articles. Amsterdam/Philadelphia: John Benjamins. Hyland, K. (2000). Hedges, boosters and lexical invisibility: Noticing modifiers in academic texts. Language Awareness, 9(4), 179–197. https://doi.org/10.1080/09658410008667145 Hyland, K. (2002). Authority and invisibility: Authorial identity in academic writing. Journal of Pragmatics, 34, 1091–1112. https://doi.org/10.1016/S0378-2166(02)00035-8 Hyland, K. (2005). Metadiscourse: Exploring writing in interaction. London, UK: Continuum. Hyland, K. (2012). Undergraduate understandings: Stance and voice in final year reports. In K. Hyland & C. Sancho-Guinda (Eds.), Stance and voice in written academic genres (pp. 134– 150). Basingstoke: Palgrave Macmillan. https://doi.org/10.1057/9781137030825_9 Hyland, K., & Milton, J. (1997). Qualification and certainty in L1 and L2 students’ writing. Journal of Second Language Writing, 6(2), 183–205. https://doi.org/10.1016/s1060-3743(97)90033-3 Kafes, H. (2017). An intercultural investigation of meta-discourse features in research articles by American and Turkish academic writers. International Journal of Languages’ Education and Teaching, 5(3), 373–391. https://doi.org/10.18298/ijlet.1823 Kaplan, R. (2000). Contrastive rhetoric and discourse analysis: Who writes what to whom? When? In what circumstances? In S. Sarangi & M. Coulthard (Eds.), Discourse and social life (pp. 82–102). Harlow, UK: Longman. Karahan, P. (2013). Self-mention in scientific articles written by Turkish and non-Turkish authors. Procedia - Social and Behavioral Sciences, 70, 305–322. https://doi.org/10.1016/j.sbspro.2013.01.068 Kawase, T. (2015). Metadiscourse in the introductions of PhD theses and research articles. Journal of English for Academic Purposes, 20, 114–124. https://doi.org/10.1016/j.jeap.2015.08.006 Koutsantoni, D. (2004). Attitude, certainty and allusions to common knowledge in scientific research articles. Journal of English for Academic Purposes, 3(2), 163–182. https://doi.org/10.1016/j. jeap.2003.08.001 Koutsantoni, D. (2005). Certainty across cultures: A comparison of the degree of certainty expressed by Greek and English speaking scientific authors. Intercultural Pragmatics, 2(2), 121–149. https://doi.org/10.1515/iprg.2005.2.2.121 857 EDUCATIONAL SCIENCES: THEORY & PRACTICE Kranich, S. (2011). To hedge or not to hedge: The use of epistemic modal expressions in popular science in English texts, English–German translations, and German original texts. Text & Talk, 31(1), 77–99. https://doi.org/10.1515/text.2011.004 Lewin, B. A. (2005). Hedging: an exploratory study of authors’ and readers’ identification of ‘toning down’ in scientific texts. Journal of English for Academic Purposes, 4(2), 163–178. https://doi. org/10.1016/j.jeap.2004.08.001 Liu, Y., & Buckingham, L. (2018). The schematic structure of discussion sections in applied linguistics and the distribution of metadiscourse markers. Journal of English for Academic Purposes, 34, 97–109. https://doi.org/10.1016/j.jeap.2018.04.002 Lombard, M., Snyder-Duch, J., & Bracken, C. (2010). Practical resources for assessing and reporting intercoder reliability in content analysis research projects. Retrieved from http:// matthewlombard.com/reliability/ Martin-Martin, P. (2003). A genre analysis of English and Spanish research paper abstracts in experimental social sciences. English for Specific Purposes, 22(1), 25–43. https://doi. org/10.1016/S0889-4906(01)00033-3 Martin-Martin, P. (2008). The mitigation of scientific claims in research papers: A comparative study. International Journal of English Studies, 8(2), 133–152. Retrieved from http://revistas. um.es/ijes/article/view/49201 McLaren-Hankin, Y. (2008). ‘We expect to report on significant progress in our product pipeline in the coming year’: Hedging forward-looking statements in corporate press releases. Discourse Studies, 10(5), 635–654. https://doi.org/10.1177/1461445608094216 Molino, A. (2010). Personal and impersonal authorial references: A contrastive study of English and Italian Linguistics research articles. Journal of English for Academic Purposes, 9(2), 86– 101. https://doi.org/10.1016/j.jeap.2010.02.007 Molino, A. (2018). “What I’m speaking is almost English…”: A corpus-based study of metadiscourse in English-medium lectures at an Italian university. Educational Sciences: Theory & Practice, 18, 935–956. http://dx.doi.org/10.12738/estp.2018.4.0330 Onder-Ozdemir, N., & Longo, B. (2014). Metadiscourse use in thesis abstracts: A Cross-cultural Study. Procedia - Social and Behavioral Sciences, 141, 59–63. https://doi.org/10.1016/j. sbspro.2014.05.011 Peterlin, A. P. (2010). Hedging devices in Slovene-English translation: A corpus-based study. Nordic Journal of English Studies, 9(2), 171–193. Retrieved from http://ojs.ub.gu.se/ojs/index. php/njes/article/view/419 Rubin, V., Liddy, E., & Kando, N. (2006). Certainty identification in texts: Categorization model and manual tagging results. In J. Shanahan, Y. Qu, & J. Wiebe (Eds.), Computing attitude and affect in text: Theory and applications (pp. 61–76). Dordrecht: Springer. https://doi.org/10.1007/1-4020-4102-0_7 Samraj, B. (2002). Introductions in research articles: variations across disciplines. English for Specific Purposes, 21(1), 1–17. https://doi.org/10.1016/S0889-4906(00)00023-5 Samraj, B. (2013). Form and function of citations in discussion sections of master’s theses and research articles. Journal of English for Academic Purposes, 12(4), 299–310. https://doi. org/10.1016/j.jeap.2013.09.001 Šeškauskiene, I. (2008). Hedging in ESL: A case study of Lithuanian learners. Studies about Languages (Kalbų Studijos), 13, 71–76. Retrieved from https://www.kalbos.lt/zurnalai/13_ numeris/10a.pdf 858 Akbas, Hardman / Strengthening or Weakening Claims in Academic Knowledge Construction: A Comparative... Silver, M. (2003). The stance of stance: A critical look at ways stance is expressed and modeled in academic discourse. Journal of English for Academic Purposes, 2, 359–374. http://dx.doi. org/10.1016/S1475-1585(03)00051-1 Stubbs, M. (1986). ‘A matter of prolonged fieldwork’: Notes towards a modal grammar of English. Applied Linguistics, 7(1), 1–25. https://doi.org/10.1093/applin/7.1.1 Tankó, G. (2017). Literary research article abstracts: An analysis of rhetorical moves and their linguistic realizations. English for Specific Purposes, 27, 42–55. https://doi.org/10.1016/j. jeap.2017.04.003 Tessuto, G. (2015). Generic structure and rhetorical moves in English-language empirical law research articles: Sites of interdisciplinary and interdiscursive cross-over. English for Specific Purposes, 37, 13–26. https://doi.org/10.1016/j.esp.2014.06.002 Thomas, J. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4(2), 91–112. https://doi. org/10.1093/applin/4.2.91 Varttala, T. (1999). Remarks on the communicative functions of hedging in popular scientific and specialist research articles on medicine. English for Specific Purposes, 18(2), 177–200. https:// doi.org/10.1016/s0889-4906(98)00007-6 Varttala, T. (2001). Hedging in scientifically oriented discourse. Exploring variation according to discipline and intended audience (Doctoral dissertation). Retrieved from http://acta.uta.fi/ pdf/951-44-5195-3.pdf Vassileva, I. (2001). Commitment and detachment in English and Bulgarian academic writing. English for Specific Purposes, 20(1), 83–102. https://doi.org/10.1016/s0889-4906(99)00029-0 Vázquez, I., & Giner, D. (2009). Writing with conviction: The use of boosters in modelling persuasion in academic discourses. Revista Alicantina de Estudios Ingleses, 22, 219–237. Retrieved from http://rua.ua.es/dspace/handle/10045/13822 Vergaro, C. (2011). Shades of impersonality: Rhetorical positioning in the academic writing of Italian students of English. Linguistics and Education, 22(2), 118–132. https://doi.org/10.1016/j. linged.2010.11.001 859 View publication stats