Improving the Use of Language Hedges in Academic Writing through Reading Journal Articles

In the arena of English for academic purposes, many nonnative speakers of English in different contexts find it difficult to perform well, because academic genre is alien to them. Current study was based on improving writing skills with a special focus on writing of language hedges in argumentative essays through reading journal articles. The study focused on the two research questions; 1.) Is there a significant improvement in using language hedges in academic writing through reading journal articles? 2.) What are the perceptions of students in reading journal articles to improve academic writing? Methodology of the study was based on quasi experimental and longitudinal design. Mixed method was utilized in collection of data. Participants of the study were 32 first year undergraduates of an English Language Teaching degree programme of a vocational technological university in Sri Lanka. Quantitative data was collected through a questionnaire and intervention through reading journal articles. Qualitative data was collected by interviewing 12 selected participants of the study. Two subject expertise evaluators and AntConc (2019), SPSS (23), MS Excel (Office 365) and thematic analysis were used to analyse data. Findings of the study reveal that there is a significant improvement in using language hedges by reading journal articles in the five categories of language hedges concerned; epistemic hedges, lexical hedges, lexical verbs, modal verbs and possibility hedges according to the descending order of the usage and the rate of improvement. Further, it can be concluded that pleasure and conscious reading of journal articles provide both cognitive and affective insights for novice academic writers of English. Two major implications for further research were drawn; to study the effect of language hedges in the culture of first language affects the usage of language hedges among undergraduates, and to study on the other stance features and engagement features in academic writing among the undergraduates in the Sri Lankan context.


INTRODUCTION
Second language (L2) writing is considered different from first language (L1) writing. L2 writing researchers claim that L2 writing is strategically, rhetorically and linguistically different in important ways from L1 writing. Learners have different writing experiences, different aptitudes and different motivational levels in L2 writing. They have varying metacognitive knowledge of their L1 and different experience of using L1, and especially writing is based on different individual characteristics. According to the constitution of Sri Lanka, at present, in the Sri Lankan context, English is considered as the link language. However, in the Sri Lankan education system, English is considered the L2. Thus, in Sri Lanka, English can be considered as a language which plays a dual role as a L2 as well as a link language.
In Sri Lankan government universities, the undergraduate degree programmes based on science, technology, medicine, management and finances are offered in the medium of English. This includes the medium of instruction and the Published by Australian International Academic Centre PTY.LTD. Copyright (c) the author(s). This is an open access article under CC BY license (https://creativecommons.org/licenses/by/4.0/) http://dx.doi.org/10.7575/aiac.alls.v.11n.3p.17 medium of assessments. However, the medium of instruction of certain degree programmes, mostly related to humanities and social sciences, is in vernacular (Rameez, 2019;Rathnayake, 2013;Shriganeshan, 2017). Even though the students who enter Sri Lankan government universities learn English, which is taught as a subject in the school curriculum of the Sri Lankan education system from Grade 1 to Grade 13, and/or those who study under the National Vocational Qualification (NVQ) framework since Grade 11, the students find difficulties in academic performances when they enter the universities and follow the degrees in the English medium. The medium of instruction of all the undergraduate degree programmes in the vocational technical university in Sri Lanka, from which the sample was selected, is English. Having recognized the importance of English, the students are offered courses related to communication skills in English during their first year. However, the problem of students scoring lower grades for content-based modules remains the same.

Research Problem
Literature provides evidence for the difficulties in academic writing, since academic language is nobody's L1 (Hirvella, 2004). The research problem of the current study seeks to evince to what extent conscious reading of academic texts, specifically journal articles focusing one specific feature of academic genre, i.e. language hedges, can be improved in writing academic essays.

Aim of the Study
The aim of the study is to find out how far conscious reading of journal articles with specific focus on language hedges will enable the undergraduates to use language hedges accurately and appropriately in writing academic essays.

Objectives of the Study
Following objectives were formulated for the study; 1) To find out whether there a significant improvement in using language hedges in academic essays by undergraduates by reading journal articles. 2) To explore the perceptions of the undergraduates in reading journal articles to improve academic writing. Hypothesis of the study is that there is a significant improvement in using language hedges in academic writing by reading journal articles. Null hypothesis of the study is that there is no significant improvement in using language hedges in academic writing by reading journal articles. Mauranen et al., (2010) states that, in the present the academic setting, the native speakers of English are outnumbered by non-native speakers who use English for academic purposes. Therefore, English today is used by native speakers and non-native speakers at all levels in the academic world. Hence, academic writing covers textual aspects of writing, and lexico-grammatical phenomena and writing styles, including even learner languages, vernacular universals and varieties of English.

Language Hedges in English for Academic Purposes for Non-native Speakers of English
Moreover, Johns (2009) observed that at an undergraduate level, English for Academic Purposes (EAP) is "more complex and elusive than most other English for Specific Purposes (ESP) categories" because it is shaped by various factors, including differences in higher educational systems at country level that influence the positioning of EAP courses and materials (p.41). For example, Wingate (2012) highlights the lack of research on the genre of the argumentative essay, which is the most common genre found in undergraduate writing. Hyland (2009) explains that in academic discourse, interactions exist between the writer and the reader; and these interactions are achieved through the systems of stance and engagement. By 'stance ' Hyland (2009) refers to the "community recognized personality" of the writer or the "textual voice of the writer". It conveys the writer's judgements, opinions and commitments. Writers' recognition of the presence of their readers "to actively pull them along with the argument, include them as discourse participants, and guide them to interpretations" is known as 'engagement' and it is "more of an alignment function" (p.74). Swales & Feak (2012) note that the stance or perspective in academic writing is important because it allows the writer to reveal not only what the writer knows, but also what the writer thinks. Moreover, Swales & Feak (2012) reveal that, among the stance features, the use of hedges marks the highest average. The way in which the writer reveals his/her stance contributes to author positioning. However, as cited in Hyland (1998), Crystal (1995) notes that the concept of hedging is an area of some neglect.
Regarding hedges in academic writing among L2 learners, Hyland (1998) explains that hedges are complex literary devices for novice writers to handle, for several reasons. One reason is that hedges can convey different meanings simultaneously; these are the referential meaning of the writer and the relationship with the reader. Another reason is that epistemic meanings can be assigned in several ways. Moreover, there are difficulties in the process of learning such as differences in meaning when expressing modality in L1, non-native students' use of hedges in L2 measured by an expert writer of L1 and a speaker of English in the academic community, and the absence of adequate pedagogic materials to teach hedging. Hinkel (2005) found that nonnative speakers tended to have a very limited number of hedging devices that they employed frequently, and these seemed to be less sophisticated than those used by native speakers.

Teaching Writing through Reading in Academic Contexts
Starting from the Scaffolding Theory, several researchers claim how writing is acquired through reading either in L1 or in L2. Swales & Feak (2012), explain that the writing tasks designed at particular universities will vary from one degree programme to another, but that they are similar in two aspects. "First, the tasks become progressively more complex and demanding the further you go in the program. Second, in general, they need to be written 'academically', although certain assigned writing in some fields may require personal reflection" (p.1).
According to Hyland (2003) teaching writing is a process of four stages as; familiarization, controlled writing, guided writing and free writing. In the first stage of familiarization, the students are taught grammar and vocabulary relevant for a particular writing task usually through a text. In the stage of controlled writing, using substitution tables, learners write using fix patterns. In guided writing, writers use model texts as a guide and imitate them in the writing endeavor. In the last stage, i.e. free writing, learners use patterns they have developed in the previous stages of the process and produce written texts. Several theories supporting teachers' efforts to understand L2 writing and learning have developed since research in writing in ESL/EFL context emerged in 1980s. Thus, in L2 writing, theories based on language structures, text functions, creative expression, writing process, content and genre (Hyland, 2003). Further, Hyland (2003) proposes that out of the six orientations, L2 writing teaching methodology should incorporated and extend the insights of each orientation, since writing is a socio-cognitive activity. Since L2 writing is a developmental aspect of language learning, written texts of L2 students are found less effective than native English-speaking writers and texts written by L2 writers are observed as generally shorter, less cohesive and with more errors (Hyland, 2003). Hyland (2003) claims that at different stages of proficiency, reading has shown a positive influence on composing skills, whether reading is voluntary or is an assigned activity. Furthermore, Hirvela (2016) finds that there is an "increased interest in the role of reading as related to writing in line with the fact that in academic contexts, students are not often asked to write without some kind of stimulus or input, usually in the form of reading materials ……… they are reading for writing" (p.127).
It could also be noted that, as highlighted in the stage three of Conduit Hypothesis forwarded by Krashen (2018), i.e. 'narrow academic reading' where the readers are engaged in a "great deal of narrow reading of academic texts in an area of great personal interest to the reader" (p.3-4). He comments further that much of academic language competence cannot be entirely achieved either through only attending the class/lectures or through writing, because writing is the output and not the input, and academic language cannot be learnt but acquired.
Thus, available literature facilitates how reading enacts as a scaffolding to improve writing for academic purposes, whereas the current study is focused on using language hedges in writing for academic purposes through reading, which would yield some significant implications in the arena of features of EAP in the Sri Lankan context.

METHOD
Both qualitative and quantitative data was used in carrying out the study. In collecting quantitative data, essays written by the participants before and after the intervention were analyzed. In collecting qualitative data, semi structured interviews were conducted.

Research Design
A longitudinal, quasi experimental research design was carried out, utilizing the mixed method comprising of both quantitative and qualitative methods.
The quantitative data were clustered by participants under the independent variable of journal of journal articles, focusing on the usage of language hedges. For the quantitative study, the hypothesis presented in section 1.1 was tested at the -5 to +5 significance level.
The qualitative method involving a semi structured interview was used at the end of the study. The interview focused on the reflections of the participants who were selected based on their performance in the post test and the pre test, in order to find answers for the research question number ii.

Setting, Participants and Sampling Procedure
The setting of the research was based on a vocational technological university in Sri Lanka. In the university, all the undergraduate degree programmes are offered in English medium. The majority of undergraduates who follow degree programmes at this university have completed a diploma of at least one year duration.
The target population of the study was all undergraduates of the university. The accessible population was all undergraduates of the Bachelor of Education in English Language Teaching (B. Ed. ELT) degree programme. The sample population was selected from the first year, second semester undergraduates of English Language Teaching degree programme. The participants are prospective teachers of English language teaching. The sample population under investigation was selected randomly by inviting the undergraduates to participate in the study on a voluntary basis, by explaining to them that the study is an attempt to improve their academic writing through reading. From the total number of 47 students in the class, 39 participants volunteered for the study at the beginning, but due to irregular attendance and drop out during the course of the study, seven participants were removed from the analysis. Therefore, the total sample size was 32.
Thus, the participants of the study are 32 first year second semester undergraduates who follow the B. Ed. ELT degree programme during weekdays.
Out of the 32 participants of the study, 12 selected participants were subjected to semi-structured interviews after the post test. The randomization of participants was done based on the lowest, average and highest marks scored at the pre test and post test.

Instrumentation
The instruments used in the study were the marking rubric for essays and the semi-structured interview.
The rubric was given to two subject expertise evaluators who were experts in the subject in the field of academic writing at the tertiary level, to record accurate language hedges at the pre-test and post-test, and to categorize the types of language hedges used by the participants so as to create a common corpus for analysis. Thus, the rubric was created according to criterion-referenced practices where "the quality of each essay is judged in its own right against some external criteria such as coherence, grammatical accuracy, and contextual appropriateness." (Hyland, 2003, p.226) Within the criterion referenced practice in marking, traitbased scoring, which defines the "specific and genre features of the task being judged" (Hamp-Lyons as cited in Hyland, 2003, p. 229) was focused on in preparing the rubric. The trait considered in scoring in the developed rubric was language hedges.
The categories under language hedges which were defined by Hyland (1995) and Hinkel (2002;2005), were analyzed and compared to the usage of hedges among the participants, and a combination of the above three definitions were utilized for the rubric created by the researcher.
Thus, the rubric focused on the following categories of language hedges; i. Lexical verbs: indicate, suggest, appear, propose, etc. ii. Modal verbs: can, may, could, would, etc. iii. Epistemic hedges: adjectives, adverbs and collocational phrases such as according to (+noun), actually, apparent(-ly), approximate(-ly), etc. iv. Lexical hedges: refers to the terms which represent prepositional modifiers and/or lexical phrases which are relatively lexically simple but can be syntactically complex such as 'in a way', 'kind of', 'maybe', 'like', 'sort of', etc. v. Possibility hedges: by (some/any) chance, perhaps, possibly, etc. The instructions to the essays given at the pre-test and post-tests did not indicate that the participants needed to use language hedges.
Semi structured interviews were conducted with 12 selected participants. Since the interviews were designed to draw the reflective experiences of the participants, the interviews contained more open-ended questions and the interviews evolved according to the experiences shared by the participants.

Procedure
Procedures used at the levels of the intervention, semi structured interviews, data processing and analysis and ethical considerations are discussed in this section.

Procedure of the intervention process
The essay topics provided at the pre-test and post-test were piloted with 06 sophomores pursuing the same degree programme to avoid possible difficulties in analyzing the data, if the topics given under discussion were to be written without language hedges. Moreover, piloting the essays was done to check whether the essay topics had similar scope and style at both test levels, in order to avoid inconsistencies in the analysis.
The topics used for the post-test were selected depending on students' familiarity with the themes discussed during the intervention process. The intervention period consisted 06 weeks and during each week one text was read and discussed by the group. Further, an essay type answer was written by group members and the answers were discussed by the group. Therefore, during each week, one session of one hour and another session of two hours were conducted. The intervention period was limited to 06 weeks, with a maximum of 03 contact hours per week, and the texts were based on the most preferred themes of the participants.
The journal articles were selected based on the theme, social media which was suggested by the majority participants at a preliminary discussion. The articles were distributed among the participants one week before the discussion, so that the participants were able to read them before coming to class. After reading the text, group discussions got underway on the conclusion/summary of the text. The participants were guided towards identifying the language hedges used in each text, and the role played by the identified language hedges in the text were discussed in the group. Then the participants were given an argumentative essay question based on the text to answer in class within 35 minutes, reflecting on the content and academic writing style.
In selecting journal articles, brevity and the simplicity of the articles were also considered. Since the participants are in their first year, they were encouraged to read only the introduction and conclusion, excluding the other sections of the article. However, the participants were provided with the complete article, for those who would like to read the complete article, no restrictions were implemented.
After the intervention period where extracts from journal articles were read, the post-test essay was given to the participants.

Procedure of the interviews
Interviews were conducted among 12 selected participants through stratified sampling. The participants were encouraged to provide reflections on their learning experiences of reading different kinds of texts and reflections on academic writing with a special focus on language hedges as one of the stance features.

Data processing and analysis
Data gathered through the pre-test and the post-test were processed based on the uses of language hedges. In analyzing the categories of hedges, the face-value frequency of hedges occurring in an individual essay through reading a previously identified set of written texts is somewhat deceiving because a word or a phrase might function as a hedge or not, depending on the context in which it occurs. Therefore, counting raw instances of use in a set of vocabulary provided in a list might be deceptive. Especially, when considering hedges of L2 writers whose vocabulary is yet to be developed, students' limited knowledge of grammatical features, register, or subtle connotations of a word might yield erroneous results. Moreover, the lexical items which were misused can also be identified as hedging devices because even the readers may have differential ideas about the meaning and use of a particular hedging device. Therefore, context is essential in identifying language hedges. Thus, the evaluator's role became very important in deciding the accuracy of the results depending on the context and creating a corpus of all types of hedges that occur in all essays.
Then, to compute the number of categories of hedges used, basic analysis tool of AntConc (2019) was used. Then the numbers of occurrences were analyzed by comparing the results of pre-test and post-test and were compared through the paired sample T test using SPSS. The hypothesis was tested according to the results. The statistical data was analyzed using SPSS (23) and graphical representations of graphs and charts were done using MS Excel (Office 365).
Data gathered through the semi structured interviews were recorded and analyzed based on the themes revealed by the participants and represented the qualitative part of the analysis.

Ethical considerations
A consent form was obtained from the students who volunteered at the initial stage of the study. In the consent form, the participants were made aware of the purpose, instrumentation, and process of the study, their freedom to leave the intervention since participation was voluntary, and the confidentiality and anonymity of their information was mentioned.
Furthermore, the participants were made aware that the evaluation of tests and providing feedback has no any direct or indirect relation with the examinations or evaluations conducted by the university.

Comparison of the Use of Language Hedges in Academic Essays at the Pre-test and the Post-test
Results and discussion of the first objective set for the study, i.e.: "to find out whether there a significant improvement in using language hedges in academic essays by undergraduates by reading journal articles" is discussed in this subsection.
Results of the paired sample t test of the pre-test and the post-test after the intervention is given in Table 1. Table 1 illustrates the results of the paired sample t test conducted at the pre-test and the post-test conducted after the intervention made with reading journal articles. T test results showed that the paired differences between the mean values of the pre-test and the post-test is -3.6656 with a p value of -9.932 and a level of significance less than 0.005. The result reveals that there is a significant improvement in students' use of language hedges in argumentative essays by reading journal articles. Therefore, the hypothesis that there is a significant improvement in using language hedges in academic writing by reading journal articles is accepted. The null hypothesis that there is no significant improvement in using language hedges in academic writing by reading journal articles is rejected.
Though Swales & Feak (2012) reveal that, among the stance features, the use of hedges marks the highest average, in the current study this scenario cannot be discussed because the study was conducted only based on language hedges and other stance and engagement features used by undergraduates in writing argumentative essays were not focused.
The comparison was further extended to categories of language hedges used by participants at the pre-test and post-test, after the intervention and the result is presented in Figure 1. Figure 1 shows that there is an improvement of all the categories of language hedges in the study after the intervention of reading journal articles. However, the improvement varies from one category to the other. The highest percentage of usage at the pre-test and the post-test are the epistemic hedges and that category shows a higher percentage of improvement of 1.4%. The second highest category of language hedges at both tests are lexical hedges and the post-test results shows 1.3% of improvement, which is slightly below the improvement of epistemic hedges. Use of lexical verbs as language hedges, use of modal verbs as language hedges and use of possibility hedges show the third highest to the least percentage of usage of language hedges at the pre-test and the post-test. While showing an improvement of usage of all the categories of language hedges in academic writing, the descending order of the usage of categories of language hedges i.e. epistemic hedges, lexical hedges, lexical verbs, modal verbs and possibility hedges remain unchanged in both the pre-test and post-test.
Moreover, it could be noted that, most of the language hedges which were there in the journal articles read, were used by the participants at the post-test, showing the impact of acquisition of language hedges through reading to writing. This finding supports the observations of Hirvela (2016) that, in academic contexts, there is a higher interest for reading related to writing, because they are asked to write based on an input provided through a reading material, and the students are reading for writing.

Perceptions of the Participants Regarding the use of Journal Articles to Improve Academic Writing
Results and discussion of the second objective set for the study, i.e.: "to explore the perceptions of the undergraduates in reading journal articles to improve academic writing" is discussed in this subsection. In anlysing the perceptions of the participants, it could be categorized into two themes as motivation and awareness.
Under the theme of motivation, it could be noted that many participants mentioned at the interview that they were motivated to read journal articles, "though reading the journal articles are not that exciting" (Participant 6B) because the participant stated that, "I understood that I have to express my stance which is a responsibility. Other writers have done so." Participant 2A noted that "When I read my previous answers, now I feel that they are 'not fully dressed' for the  In discussing the theme of awareness, it can be discussed under several types of awareness such as awareness of the importance of reading in improving writing, career awareness of the participants and awareness of their prior leaning experience.
Participant 7A mentions that "I understand that more reading makes writing sensible, though reading is a bit boring at times", revealing her experience in learning to write through reading, proving the awareness she gained regarding the importance of reading in improving writing. Moreover, as participant 2C mentioned, "What I understood was, the journal article writers have used a lot of hedges in the introduction, discussion and conclusion sections. Now I'm doing the same", showing the extent that the participants' awareness has increase as to search for in which areas the journal article writers have used language hedges in their articles.
Similar scenario on the themes of both motivation and awareness could be discussed based on the stage three, i.e. 'narrow academic reading' stage of Krashen's Conduit Hypothesis (2018) academic language competence can be achieved through narrow reading of academic texts of great personal interest to the learner and in this case, the undergraduates subjected to study. In comparing the participants' perceptions revealing their high motivation and awareness with their performances at the two test levels evinces Krashen's (2018) viewpoint that, writing is the output and not the input and academic language cannot be learnt but acquired because academic language cannot be totally achieved through only attending to lectures or though writing.
Furthermore, participant 6B noted that "I know I will be a non-native speaking teacher of English in the future, but I should learn now to have good language skills when I become a teacher", showing their awareness regarding their career that they will be engaged in after graduation, and how the intervention has indirectly motivated them to learn the written language skills through reading, as second language speakers. Additionally, participant 5C highlights regarding their prior learning experience, before entering the university as undergraduates, as; "In our ……. Diploma, if we were made aware of language hedges and boosters, it could have been beneficial to everyone. Not all who hold our Diploma could follow this degree." What the participant notes here was that, if such academic writing skills could be exposed when gaining their entry qualification, it will be beneficial for them as undergraduates as well as for those who cannot join the degree, but complete the diploma and leave after the graduation and pursue mostly in different levels of teaching English as a second language.
However, it should be noted that the group of participants were in their first year second semester, when the intervention was made and they were novice to the discipline of academic writing. Moreover, as mentioned in the methodology, the group of participants in the study were undergraduates of English Language Teaching degree programme thereby, the percentage of usage of language hedges might differ from other different degree programmes, because Hyland (2009) highlights the fact that 75% of stance and engagement features occur in the humanities and social sciences.

CONCLUSIONS
Through the analysis of quantitative and qualitative data, as a conclusion to the first research objective, it could be concluded that reading journal articles, which belong to the category of nonfictional texts, can be used effectively in improving the accurate use of hedges when participants write argumentative academic essays. Choice of journal articles to suit the level as well as preferences of the undergraduates also becomes crucial in using journal articles to improve participants' awareness and to improve the production of hedges in written texts for two reasons. One is that undergraduates participating in the study are not L1 speakers of English, and it is only after becoming undergraduates that reading and writing in English became mandatory for them. The second reason is that prior research conducted on the topic of academic reading and writing reveals that even native speakers of English find reading and writing in English difficult at undergraduate level because the academic genre is new to them.
According to the findings, as the conclusion for the second research objective, it could be concluded that, with conscious reading of journal articles focusing on features of academic writing such as language hedges, the undergraduates who are non-native speakers of English, yet have to learn in English as the medium of higher education are able to feel that they are 'fully dressed' when they acquire the awareness and the usage of language hedges through reading journal articles, showing a higher level of motivation and awareness of language hedges in academic texts.
As implications for further research, since literature related to the use of language hedges among undergraduates in Sri Lanka was not available so far, the findings are useful in developing EAP curriculum at the undergraduate courses in Sri Lanka, by focusing on hedging in L2 writing instruction, and expanding the exploration further towards improving other stance features and engagement features in academic writing among the undergraduates in the Sri Lankan context. Moreover, it could be suggested that, to study the impact of the culture of L1 of the participants towards the usage of language hedges in academic writing in English among the undergraduates in the Sri Lankan context could also be focused.