هدف تحقیق حاضر بررسی تاثیر نشانه های فرا گفتمان متنی بر فهم متون انگلیسی به وسیله ی زبان اموزان فارسی زبان است.این تحقیق علاوه بر این کوشیده است تا میزان اگاهی این زبان اموزان و نحوه ی تعامل انان را با متون خوانده شده در زبان انگلیسی به وسیله ی پرسش نامه ی تهیه شده بررسی کند.بر اساس محتوای یک متن انگلیسی یازده سوال درست /غلط طرح گردید و یک مرتبه با ان متن و یک مرتبه با نسخه ای که نشانه های فرا گفتمان متنی د ان حذف شده بود استفاده شد.بررسی های اماری نشان داد که نمرات زبان اموزان به صورت معنا داری در متن دارای نشانه های فرا گفتمان متنی بهتر بود.علاوه بر این ،نتایج اماری این نکته را نیز اشکار کرد که در زبان دوم،زبان اموزانی که سطح زبان پایین تری دارند بیشتر از افرادی که از سطح بالاتری در زبان دوم بر خوردارند از نشانه های فرا گفتمانی بهره می برند .نتایج پرسش نامه نیز مشخص کرد که سختی متن دارای دو معنی متفاوت است.سختی، در معنای عمومی آن با نوع متن رابطه ی معناداری نداشت ولی سختی در معنای توانایی در فهمیدن نکات کلی متن رابطه معنادار با نوع متن داشت بدین معنی که شرکت کنندگان در این مطالعه بر این عقیده بودند که متون دارای نشانه های فرا گفتمان متنی را بهتر درک کرده اند.به طور کلی این تحقیق بر تاثیر مثبت نشانه های فرا گفتمان متنی بر فهم متون در زبان دوم صحه میگذارد. payamenoor university faculty of humanities the comparison of the effects of metadiscourse markers and proficiency level in reading comprehension of iranian efl learners submitted in partial fulfillment of the requirements for the degree of m.a. in english language teaching by: rozana shamsabadi advisor: dr. m.jafarigohar reader: dr. f.hemmati july 2009 “dedicated to my parents, for their love, support, and encouragement” abstract the present study aimed at investigating the effects of metadiscourse markers on the comprehension of english texts by iranian efl learners. it, moreover, tried to measure the subjects’ awareness of those markers and their interaction with the texts by using a follow-up questionnaire. a set of 11 true/false questions were developed based on the propositional contents of an original english text and its doctored, metadiscourse-removed version. then, it was used as the english reading comprehension test. the participants’ awareness of metadiscourse markers and their interaction with the texts were measured by a set of two multiple-choice questions used as a follow-up questionnaire. the results of the analysis showed that all the participants of the study performed significantly better on the un-doctored text. the study, moreover, revealed that it was the lower learners rather than the higher ones who benefitted more from the presence of metadiscourse markers. the results of the follow-up questionnaire also revealed that difficulty of a text in its general sense had nothing to do with the presence or absence of metadiscourse markers, but difficulty in the sense of being able to comprehend the major points of the texts was closely related to the presence or absence of those markers ,with un-doctored text felt to be understood more by all the participants. keywords: metadiscourse, interaction, text, reading comprehension, awareness, difficulty. acknowledgements i would like to express my gratitude to those who gave me the possibility to complete this thesis. first and foremost, i am deeply indebted to my advisor, dr. m. jafarigohar, and my reader, dr. f.hemmati, for their encouragement and insightful comments. i would also like to thank all the esteemed professors at faculty of humanities of the payamenoor university for their invaluable support, help, and cooperation. moreover, i have to thank prof. avon crismore, and prof. william vande kopple, those who kindly sent me the articles i needed most, and supported me throughout the research process. last but not least, i would like to thank mr. a.sharif and other colleagues from language institute of sadr of isfahan, whose kind cooperation made the data collection possible. table of contents chapter one: introduction 1.1. overview 11 1.2. statement of the problem and significance of the study 13 1.3. research questions 15 1.4. research hypotheses 16 1.5. definition of important terms 16 1.5.1. discourse 16 1.5.2. metalanguage 16 1.5.3. metadiscourse 16 1.5.4. textual metadiscourse 17 1.5.5. interpersonal metadiscourse 17 1.5.6. reading 17 1.6. limitations of the study 17 chapter two: review of literature 2.1. importance of reading 19 2.2. definitions of reading 19 2.3. l1readingvs. l2 reading 21 2.4. reading purposes 22 2.5. mayor’s theory for meaningful learning from prose passages 23 2.6. discourse analysis 23 2.7. rhetoric 26 2.7.1. preliminary points 26 2.7.2. rhetoric and metadiscourse 28 2.8. metadiscourse 30 2.8.1. introduction 30 2.8.2. definitions of metadiscourse 34 2.8.3. metadiscourse: functional or syntactic 38 2.8.4. functions of metadiscourse 39 2.8.5. metadiscourse signals 41 2.8.6. categorizations of metadiscourse 45 2.8.6.1. lautamatti & williams categorizations 45 2.8.6.2. vande kopple categorization 46 2.8.6.3. crismore et al’s categorization 48 2.8.6.4. hyland`s interpersonal model to learners 50 2.9. advantages of teaching metadiscourse markers to learners 54 2.10. metadiscourse studies conducted in iranian context 56 chapter three: methodology 3.1. introduction 57 3.2. participants 57 3.3. instrumentation 57 3.4. procedure and design of the study 57 chapter four: results and discussion 4.1. overview 69 4.2. restatement of the problem 69 4.3. results and findings 70 4.3.1. the first research question 71 4.3.2. the second research question 72 4.3.3. the third research question and null hypothesis 74 4.4. discussion 80 chapter five: conclusions, implications and suggestions for further research 5.1. summary of the findings and conclusions 83 5.2. pedagogical implications 83 5.3. suggestions for further research 85 reference 89 appendices 106 list of tables table2.1. crismore et al. `s (1993) categorization of metadiscourse 39 table2.2.an interpersonal model of metadiscourse 42 table3.1.metadiscursivedifferences between two texts 47 table3.2. the relationship between questions and text paragraphs 51 table3.3. descriptive of nelson proficiency test 52 table3.4. schematic frequency representation of nelson proficiency test 53 table3.5. independent samples t-test for means of higher and lower efl learners 54 table4.1. descriptive statistics for the two english texts 56 table4.2. paired samples t-test for means of the english texts 56 table4.3. descriptive statistics for higher and lower efl learners 57 table4.4. two-way anova for the effects of text type and proficiency level 58 table4.5. crosstabs for question no.1 regarding lower efl learners 59 table4.6. crosstabs for question no.2 regarding lower efl learners 60 table4.7. crosstabs for question no.1 regarding higher efl learners 61 table4.8. crosstabs for question no.2 regarding higher efl learners 62 list of figures figure2.1. sinclair`s planes of discourse model 24 figure2.2. crismore`s model of discourse model 25 figure2.3. potential non-verbal expressions of metadiscourse 34 figure3.1. histogram of nelson proficiency test 53 figure4.1. bar graph for mean scores of the english texts 57 figure4.2. two-way anova for higher and lower efl learners 58 figure4.3. bar graph of question no.1 regarding lower efl learners 60 figure4.4. bar graph of question no.2 regarding lower efl learners 60 figure4.5. bar graph of question no.1 regarding higher efl learners 63 figure4.6. bar graph of question no.2 regarding higher efl learners 63 chapter one: introduction 1. 1. overview nowadays, most researchers or teachers of a second/foreign language pay special attention to the activity of reading, believing that reading is one of the most important skills for learners (see grabe & stoller, 2001; lynch & hudson, 1991). richards and renandya (2002) point out major reasons why language teachers should focus on this activity. first among the reasons is that many efl learners declare reading as one of their most important goals. second, extensive exposure to comprehensible written texts can facilitate language acquisition; and finally, reading provides opportunities to introduce new topics, to stimulate discussion, and to study language (e.g., vocabulary, grammar, and idioms). reading in the view of nunan (2001) is an interactive process which pushes the reader to constantly shuttle between bottom-up and top-down processes. this interactive process means that in addition to decoding the meaning of individual words, prior knowledge of content and relevant schemata should also be activated and used. moreover, aside from a student knowing a number of words, being familiar with context and falling back on background knowledge, research clearly shows that a reader’s knowledge of text structure and discourse cues has significant effects on reading in a foreign language (see carrel, 1985; carrel et al., 1989; raymond, 1993; tang, 1992). parallel to the interaction between the reader and the content in the reading process, an interaction also takes place between the reader and the writer (camiciottoli, 2003). this latter interaction is called metadiscourse, and is defined by crismore et al. (1993, p.40) as “linguistic materials in texts, written or spoken, which does not add anything to the propositional content, but that is intended to help the listener or reader organize, interpret, and evaluate the information given.” vande kopple (1997, p.2), too, defines metadiscourse as “discourse that people use not to expand referential material, but to help the readers connect, organize, interpret, evaluate, and develop attitudes towards that material.” likewise, hyland (2005, p.3) believes that “metadiscourse embodies the idea that communication is more than just the exchange of information, goods or services, but also involves the personalities, attitudes and assumptions of those who are communicating.” although in recent years some researchers have investigated the effects of some types of metadiscourse markers on reading comprehension (see aliasin, 2006; chung, 2000; geva, 1992; mohammad salehi, 2005; ozono, 2002; ozono & ito, 2003), few studies have taken a holistic look at their effects on reading comprehension (see camiciottoli, 2003), and almost no study has focused on their effects as a whole on the comprehension of english texts by iranian efl learners. being in agreement with the above discovery, the purpose of this study is, therefore, an investigation on the effects of metadiscourse on the comprehension of english texts by iranian efl students according to their proficiency levels. the possible degree of the awareness of metadiscourse markers among groups of learners will be investigated by using a follow-up questionnaire. 1. 2. statement of the problem and significance of the study today english is the lingua franca of the world, and according to crystal (1992), non-native users of english comprise more than two thirds of the potential speakers of english. it is the language of commerce, education, and diplomacy in the modern world; and hence reading in that language is an important skill for those who learn it as their second or foreign language. reading is the process of activating the relevant knowledge and related language to accomplish the exchange of information from the writer to the reader. some scholars in the past used to call it a passive skill, but nowadays, because of the priority given to the active mind of the reader, it is referred to as a receptive skill in the sense that the reader receives something from the writer and tries not only to decode the language in the text but also to recreate the writer’s intended meaning, using whatever resources he/she has at his/her disposal (chastain, 1988). in fact, each proficient writer tries as hard as possible to create a dialogue with the reader. this dialogue is brought about by the help of metadiscourse markers. therefore, as observed by hyland (2005), the writer does not simply present information about the suggested route but he/she takes the trouble to see the walk from the reader’s perspective. halliday (1985) also believes that readers must anticipate what the writer is going to say and engage themselves with the text, and in this fashion meaning can not be arrived at only by sequencing unconnected sentences. crismore (1985) also believes that all academic disciplines are conventions of rhetoric personality which influence the ways writers intrude into texts to organize arguments and represent themselves, their readers and their attitudes and all of them can be accomplished by using metadiscourse in the text. metadiscourse markers satisfy the textual and interpersonal functions of language proposed by halliday (1978), but unfortunately, as camiciottoli (2003) and vande kopple (2002) indicate, the effects of these non-linguistic materials on reading comprehension have not been intensely investigated in foreign or second language contexts. in fact the dangerous tendency, as coates (1987, p.113) points out, has always been “to concentrate on the referential function of language at the expense of all the others.” although metadiscourse has been studied from a descriptive point of view (e.g., bondi, 1999; bunton, 1999; crismore, 1989; crismore &farnsworth, 1989; crismore & farnsworth, 1990; hyland, 2000; swales, 1981 & 1990), and some researchers have investigated its use by different writers (e.g., abdollahzadeh, 2001; bandari, 1999; carlson, 1988; cheng and steffensen, 1996; connor, 1987; crismore et al., 1993; ghapanchi, 1998; hyland, 2004; intaraprawat &steffensen, 1995; marandi, 2002; simin, 2003), little work has been done on the effects of metadiscourse on reading comprehension of native or foreign language learners (e.g., camiciottoli, 2003; crismore, 1989; crismore & vande kopple, 1997); and almost no researcher has, up to this point, focused on the effects of metadiscourse as a whole (i.e., not focusing on one aspect of it like ‘connectives’) on the comprehension of english texts by iranian efl learners. therefore, the purpose of this study is to shed light on the effects of metadiscourse markers on the comprehension of english texts by iranian efl learners. this study will, moreover, include proficiency level of the learners as a moderator variable; and, finally, in this study, the awareness of the students’ of the metadiscourse markers and the degree of their interaction with the texts will also be investigated and analyzed by using a follow-up questionnaire. 1. 3. research questions in the present study an attempt was be made in order to answer the following questions regarding the effects of metadiscourse on the comprehension of english text: 1. is there any difference between the performances of efl learners on the un-doctored english text (the text with the original metadiscourse markers) and the performances of students on the doctored english text (the text whose metadiscourse markers have been removed)? 2. is there any difference between the performances of students on the two kinds of english texts across different proficiency levels? 3. do the students find the doctored texts more difficult to understand than the un-doctored ones considering different proficiency levels? 1. 4. research hypotheses based on the above research questions, the following null hypotheses are suggested: h01. there is no difference between the performances of students on the un-doctored english text and the performances of students on the doctored english text. h02. there is no difference between the performances of students on the two kinds of english texts across their different proficiency levels. h03. the students do not find the doctored texts more difficult to understand than the un-doctored ones considering different proficiency levels. 1. 5. definition of important terms 1. 5. 1. discourse according to crystal (2003, p.141), discourse refers to “a continuous stretch of (especially spoken) language larger than a sentence. mccarthy (2001, p.48) also believes that discourse is “the organization of language beyond the level of the sentence and the individual speaking turn, whereby meaning is negotiated in the process of interaction.” 1. 5. 2. metalanguage richards et al. (1992, p.227) believe that metalanguage is “the language used to analyze or describe a language.” for example in english the sentence ‘an adjective is a word that describes the thing, quality, state, or action which a noun refers to’ is in metalanguage. 1. 5. 3. metadiscourse williams (1981) defines metadiscourse as “writing about writing, whatever does not refer to the subject matter being addressed.” mao (1993, p.265) also believes that metadiscourse can “refer to discourse about discourse, to any kind of critical interpretation or theoretical exposition of a given (or target) discourse or theory.” 1. 5. 4. textual metadiscourse vande kopple (1985, p.87) defines textual metadiscourse as metadiscourse that “shows how we link and relate individual propositions so that they form a cohesive and coherent text and how individual elements of those propositions make sense in conjunction with other elements of the text.” 1. 5. 5. interpersonal metadiscourse according to vande kopple (1985, p.86-87), interpersonal metadiscourse can be defined as “communication about communication”, and is used to “express our personalities and our reactions to the propositional content of our texts and characterize the interaction we would like to have with our readers about that content.” 1. 5. 6. reading richards et al. (2001, p.306) define reading as “perceiving a written text in order to understand its contents.” widdowson (1979, p.173), too, has his own view of reading. according to him, reading is “a reasoning activity whereby the reader creates meaning on the basis of textual clues.” likewise, urquhart and weir (1998, p.22) believe that reading is “the process of receiving and interpreting information encoded in language form via the medium of print.” 1. 6. limitations of the study 1. this study deals with the effects of metadiscourse on the comprehension of english texts by higher intermediate and lower intermediate students; therefore, other proficiency levels are not included. 2. this study will be conducted in a language institute context not in a university chapter two: review of literature 2. 1. importance of reading reading according to chastain (1988) is the process of activating the relevant knowledge and related language to accomplish the exchange of information from the writer to the reader. in the past, some scholars used to call it a passive skill. but, today, because of the priority given to the active mind of the reader, it is referred to as a receptive skill in the sense that the reader receives something from the writer and tries not only to decode the language in the text but also to recreate the writer’s intended meaning. the ability to read in a foreign language is one of the most important skills required of people in academic, professional, and international settings. rivers (1981), in a similar vein, considers reading as “the most important activity in any language class, not only as a source of information, but also as a means of consolidating and extending one’s knowledge of language” (p. 259). this is particularly true of english, for today it is the language of education, commerce, and tourism, and this is why for most of the efl learners reading is the main reason why they want to learn english. 2. 2. definitions of reading although widdowson (1979, p.173) believes that generally reading is “a reasoning activity whereby the reader creates meaning on the basis of textual clues”, and urquhart and weir (1998, p.22), too, define reading as “the process of receiving and interpreting information encoded in language form via the medium of print”, carrel and grabe (2002) contend that reading is difficult to define precisely because different people read for different purposes and vary the cognitive processes and knowledge resources that they use while reading. in the same fashion, grabe (2002, p.53) believes that any comprehensive understanding of the activity of reading should take into account the following points: 1. different purposes for reading 2. definitional criteria for fluent reading 3. procedures that underlie reading as an individual skill 4. social context influences on l2 reading 5. unique features of l2 reading as opposed to l1 reading he mentions ‘reading to find information’, ‘reading to learn’, ‘reading to critique and evaluate’, and ‘reading for basic comprehension’ as the main purposes for reading in a second or a foreign language. he believes that in the first the crucial skill is scanning for a specific word, phrase or number; in the second, in contrast, the necessary skill is looking for the main ideas and awareness of many of the details of the text and an organizing frame; in the third, reflections and connections to prior knowledge and an integration with prior knowledge are needed; and, finally, in ‘reading for general understanding’, a large recognition vocabulary, automaticity of word recognition, rapid overall speed, and the ability to build overall text comprehension under relatively rapid time demands are the skills needed most. similarly, grabe (1999) claims that any definition of reading should take into account the fact that: 1. reading is a rapid process, implying that various bits of information being activated at any moment in working memory. 2. reading is an interactive process, implying the activation of both automatic and intentional skills and abilities, and also the interaction of textual, and background knowledge. 3. reading is strategic and flexible in that readers assess whether they are achieving their purposes for reading or not. 4. reading is purposeful, implying that behind each act of reading does exist a purpose and without purpose no reading is carried out by readers. 5. reading is a linguistic process, as opposed to a reasoning process 2. 3. l1 reading vs. l2 reading most scholars believe that some differences between l1 and l2 reading are matters of degree, but other strong qualitative differences exist between them. to summarize, the differences between l1 and l2 reading can be related to ‘linguistic and processing’, ‘individual and experiential’, and, finally, ‘socio-cultural and institutional’ factors which influence both kinds of reading. therefore, it can be argued that reading in l2 might cause difficulties which are not available for l1 readers and this is why l2 readers should be given more practice and help regarding the above three factors influencing their reading. 2. 4. reading purposes one of the most influential classifications of reading purposes has been proposed by davies (1995, p.102). he, based on researches done on second language reading, classifies reading into four major categories: 1. receptive reading: the rapid and automatic reading of narratives. 2. reflective reading: the kind of reading in which the reader pauses often and reflects on what has been read. 3. skim reading: rapid reading of a text in order to establish in a general way what the text is about. 4. scanning: the kind of reading in which the reader searches for specific information. however, he further argues that it is difficult to distinguish between the last two types of reading because scanning includes some skimming and skimming, in turn, includes some scanning. carrel and grabe (2002, pp.233-234), too, elaborate on the major reading purposes in the following way: the many purposes for reading, although drawing on the same cognitive processes and resources, do so in differing combinations and with varying emphasis on these processes and resources. for example, when we want information from a manual, we will search for that information by some combination of scanning for key terms and skimming small segments for meaning to see if we are in the right area of the text. when we read a newspaper we read headlines and often skim news stories to see if we want to slow down and read more carefully. when we read a good novel at night, we generally do not skim (unless we get bored), but we usually do not read carefully to remember details either. when we are trying to learn new information, we read more slowly, thinking about how information fits with prior information in the text and with our own background knowledge. 2.5. mayer’s theory for meaningful learning from prose passages mayer (1984) suggests a cognitive theory of reading comprehension (see figure2.1). according to this theory, four memories are available to the reader: sensory memory (sm), which temporarily holds incoming sensory information which are images of printed words; short-term memory (stm), which holds information that has been attended to (such as words); working memory (wm), in which information from stm can be manipulated or rearranged; and, finally, long-term memory (ltm), which is permanent store for information. 2.6. discourse analysis in the last thirty years there has been a major shift in what john swales (1990) has called ‘centre of gravity’ of linguistic concerns. broadly speaking, this can be seen as a shift from the clause to the text, and from a focus on language as a set of syntactic structures in isolation to a focus on language as a set of resources for functional use. this new perspective is called discourse analysis and its main aim is, as scott and thomson (2001, p.1) claim, “to describe communication via communicative events.” for the discourse analyst the concept of text is of utmost importance. text is defined by brown & yule (1983, p.6) as the “verbal record of a communicative act.” in discourse analysis both spoken and written texts are studied, although some scholars believe that the discourse analysts should mainly focus on spoken texts and leave the written texts for text linguists to analyze. brown and yule (1983) argue that what should be studied in discourse analysis is just ‘discourse as process’ view not ‘sentence as object’ or ‘text as product’ (cohesion) views of language. they claim that in ‘sentence as object’ view sentences need not be considered in terms of functions. this view is best seen in the following statement from chomsky (1968, p.62) where he says: “if we hope to understand human language and psychological capacities on which it rests, we must first ask what it is, not how or for what purposes it is used.” in ‘text as product’ view although both producers and receivers of sentences exist, the main focus is on the product, that is, the words on the page. this view of text is what exists in text linguistics. typical of such an approach is what can be seen in the work by halliday and hasan (1976), which can be named cohesion view of language. yule (2000, p.84) calls this view the “structural perspective of discourse” and believes that: within this structural perspective, the focus is on topics such as the explicit connections between sentences in a text that create cohesion, or on elements of textual organization that are characteristic of storytelling, for example, as distinct from opinion expressing and other text types. in contrast to those two approaches, in ‘discourse as process’ view (what is pursued in true discourse analysis) the discourse analyst considers words, phrases, and sentences which appear in the textual record to be evidence of an attempt on the part of a producer to communicate something to the receiver (brown and yule, 1983). in this view the main focus is on text sentences, those sentences derived from performance data, not on the grammarians’ sentences written with the object view of language. yule (2000) calls this “pragmatic perspective” of discourse. to put it another way, in discourse analysis, coherence which is the analysis of how stretches of language used in communication assume meaning is of utmost importance; but it is also believed that coherence is not a property solely derived from the linguistic forms and propositional content of a text, though they may contribute to it. the belief is that coherence derives from an interaction of text with given participants; and, it is, therefore, not an absolute quality, but relative to context (cook, 1998). the importance of context in discourse analysis has been emphasized by brown and yule (1983, pp.25-26) in this way: here we simply remark that in recent years the idea that a linguistic string (a sentence) can be fully analyzed without taking ‘context’ into account has been seriously questioned. if the sentence grammarian wishes to make claims about the ‘acceptability’ of a sentence in determining whether the strings produced by his grammars are correct sentences of the language, he is implicitly appealing to contextual considerations. after all, what do we do when we are asked whether a particular string is ‘acceptable’? do we not immediately, and quite naturally, set about constructing some circumstances (i.e. a ‘context’) in which the sentence could be acceptably used? in this view, the discourse analyst is no longer interested in finding rules across sentences, but in theories and techniques allowing the examination of text in context; and this is why he/she aligns himself/herself with functional analysis, conversational analysis, speech act theory, schema theory, and critical discourse analysis. 2. 7. rhetoric 2. 7. 1. preliminary points rhetoric has had different meanings in its long history. at one time, for example, it referred to one of the most important of all academic subjects; and, was literally defined as using words impressively in speech and writing in order to persuade people (ong, 1983). crystal (2003, p.400), too, defines it as “the study of effective or persuasive speaking and writing, especially as practiced in public oratory.” in other words, it can be referred to as the art of persuasion, the art which at some time in the past carried with itself a negative connotation (hyland, 2005). but today it refers to the way content is presented, and, therefore, is a central concept in discourse and discourse analysis. as mauranen (1993, p.20) observes: the study of rhetoric has been rediscovered not only as a means of improving efficiency in verbal presentation, but as an analytical tool that can be used by different disciplines for uncovering certain aspects of discourse. but even in the contemporary research into oral communication and current theories of composition, we can see the traces of classical rhetoric, especially those of aristotle. hyland (2005, p.64) briefly reviews aristotle’s main ideas regarding rhetoric: aristotle distinguishes rhetoric from the more speculative form of argument of dialectic, defining it as the art of finding available ways of establishing persuasive proof. since people are not persuaded until they are convinced that something is true, the rhetoric involves demonstrating how something is true or how it can be shown to be true. he argued that persuasion has to be adjusted for differences in three major components of communication: the speaker, the hearer and the content of the argument. further he suggested that to make an argument speakers had to attend to three points: the means of persuasion, language and the organization of the argument. richards et al. (1992, p.316), briefly summarize the major goals of the modern interpretation of rhetoric. according to them, rhetoric is the study of how effective writing achieves its goals. the term “rhetoric” in this sense is common in north american college and university courses in rhetoric or “rhetorical communication”, which typically focus on how to express oneself correctly and effectively in relation to the topic of writing or speech, the audience, and the purpose of communication. in modern times the concept of rhetoric discussed above has been taken by many scholars in order to identify criteria for a successful text. de beaugrande and dressler (1994), for example, believe that a successful text should have several criteria of (a) cohesion or grammatical relation between sentences, (b) coherence or the relationships which link the meanings of utterances in a discourse or of the sentences in a text, (c) intentionality or the author’s intention to produce cohesive and coherent texts, (d) situationality or match between text and context, (e) intertextuality or relation between two texts, and (f) informativity or anticipation of writers with respect to the amount of information shared by readers. 2.7. 2. rhetoric and metadiscourse nowadays three main means of persuasion, namely ethos, pathos, and logos are distinguished. hyland (2005, p.64) defines ethos as the “personal appeal of one’s character”, pathos as the “appeal to emotions”, and, finally, logos as the “appeal to reason.” in persuasion all three should work in combination towards persuasive ends. ethos concerns the character of the speaker and his or her credibility (hauser, 1986); pathos concerns affective appeals, and focuses on the characteristics of the audience rather than the speaker; and logos, in turn, concerns the speech itself, its arrangements, length, complexity, types of evidence, arrangements, and so on (hyland, 2005). in short, the relationship between ethos, pathos, and logos and metadiscourse can best be seen in the following words by hyland (2005, p.65): …we can see metadiscourse projecting the rational appeals of logos when it explicitly links elements of the argument; it conveys an ethos where it refers to the writer’s authority and competence; and it relates to pathos when it signals respect for the readers’ viewpoint or that the message has direct relevance to the audience. crismore (1989), too, discusses rhetorical form based on four levels, namely the knowledge level, the metadiscourse level, the text level, and the disciplinary level. the first level, the knowledge level, deals with the structure of knowledge and the structure of particular actions and events. the second one, the metadiscourse level, refers to the author’s intrusion in the text, which can be defined as “discourse about discourse, to any kind of critical interpretation or theoretical exposition of a given (or target) discourse or theory” (mao, 1993), or as “linguistic material in texts, written or spoken, which does not add anything to the propositional content” (crismore et al. 1993, p.40). this level satisfies the interpersonal function of language; and is, thus, regarded as the most important level of rhetorical form. the third one, the text level, focuses on the textual function of language, and deals with cohesion and coherence conventions in texts. finally, in the fourth one, the disciplinary level, the belief is that the more consistent and systematic the organization and presentation of ideas in a text are, the more likely it is that the ideas will be learned. 2. 8. metadiscourse 2. 8. 1. introduction as writers write, they work on two different levels. on the first level they focus on what it is that they are communicating to the readers, and on the second level they focus on how they are communicating with the readers. this first level is called “primary/discourse level” and the second level is called “metadiscourse level” (vande kopple, 1985). urmson (1952, p.231), too, distinguishes between “what is said” and an “understanding and assessment of what is said.” this understanding and assessment can also be called as metadiscourse. this level satisfies the textual and interpersonal functions of language proposed by halliday (1973, 1985). sinclair (1981) also believes that metadiscourse is at the service of the “interactive plane” of discourse, the plane which deals with the ways people use language to negotiate with others and present their texts interactively in order to create a relationship with the reader. he distinguishes this plane of discourse from what he calls the “autonomous plane”, the plane which refers to the gradual unfolding of a record of experience through the organization of text structure (see figure 2.1). real world language autonomous interactive figure 2.1. sinclair’s planes of discourse model (adapted from hyland, 2005) in the same fashion, crismore (1989) claims that as writers write, they work on two levels. on one level they convey information about their subject matter (propositional content), and, on the other level, they show their readers how to read, react to, and evaluate what was spoken or written about the subject matter. she calls the first level as “primary discourse level” and the second level as “metadiscourse level” (see figure 2.2). this second level, the “metadiscourse level”, is therefore called “talk about talk”, “writing about writing”, “discourse about discourse” or “text about text” (crismore, 1989). subject discourse (own/others) discourse metadiscourse (oral/written) (oral/written) author audience author audience discourse model metadiscourse model figure 2.2. crismore’s levels of discourse model (adapted from crismore, 1989) as the figure shows, the source of metadiscourse can be the self or others. this means that the writer can write about his text as professional or student writers often do, or he/ she can discuss and comment on someone else’s text, as writing teachers and students or professional literary critics do. similarly, hyland (2005) distinguishes between ‘informative’ and ‘interactional’ functions of language, and believes that the latter is fulfilled by appropriate use of metadiscourse by the author. the term metadiscourse was, at first, proposed by zellig harris in 1952 in order to offer a way of understanding language in use. metadiscourse options are the ways people articulate and construct their interactions. this term is closely related to the term metalanguage which can be defined as the language people use to talk about language (richards et al., 1992). in other words, it can be said that metalanguage is used in order to designate a language which speaks about another language. crismore (1989) has shown how metadiscourse has always been a ubiquitous aspect of writing from the antiquity to the present; and hyland (2005, p.11) reasons why this has been so: …not every thing that occurs in such informational genres works to convey information. speakers and writers seek to ensure that the information they present is not only distinct and intelligible, but also understood, accepted and, in many cases, acted upon. addressees have to be drawn in, engaged, motivated to follow along, participate, and perhaps be influenced or persuaded by a discourse. to do this, senders endeavour to shape their texts to the anticipated expectations and requirements of receivers. supporting his view, he lists the major metadiscourse studies which have demonstrated the importance of metadiscourse. this list includes schiffrin’s (1980) study of metadiscourse in casual conversation, crismore’s (1989) study of metadiscourse in school textbooks, norrick’s (2001) study of metadiscourse in oral narratives, crismore and farnsworth’s (1990) study of metadiscourse in science popularizations, hyland’s (2000) study of metadiscourse in undergraduate textbooks, bunton’s (1999), swales’s (1990), and hyland’s (2004) studies of metadiscourse in postgraduate dissertations , crismore and farnsworth’s (1989) study of metadiscourse in darwin’s origins of species, fuertes-olivera et al.’s (2001) study of metadiscourse in advertising slogans, hyland’s (1998b) study of metadiscourse in company annual reports, camiciottoli’s (2003) study of the effects of metadiscourse on esp reading comprehension, mauranen’s (1993b), crismore et al.’s (1993), and valero-garces’ (1996) investigations of metadiscourse in different genres and cultures, intaraprawat and steffensen’s (1995), and cheng and steffensen’s (1996) studies of metadiscourse use by native-speaker students, and, finally, crismore and farnsworth’s (1990) and hyland’s (1998a) study of metadiscourse use in persuasive and argumentative discourse. therefore, metadiscourse is a potentially useful device in the hands of the writer; and, it is said that, if used appropriately, it can guide and direct readers through a text by helping them interpret and understand the text content (vande kopple, 1985). 2. 8. 2. definitions of metadiscourse metadiscourse has always been a fuzzy term in discourse analysis and applied linguistics. this fuzziness has been remarked by nash (1992, p.100), who believes that the boundaries of this term are difficult to establish: the word ‘metadiscourse’ may have a reassuring objective, ‘scientific’ ring, but its usage suggests boundaries of definition no more firmly drawn than those of, say, ‘rhetoric’ or ‘style’. one reader may perceive a clear stylistic intention in something which another reader dismisses as a commonplace, ‘automatized’ use of language. this fuzziness is intensified more by some scholars having their own views of metadiscourse. for example, some writers like bunton (1999), mauranen (1993a) and valero-garces (1996) restrict the term to features of rhetorical organization by including only those text elements which refer to the text itself; and still others like beauvais (1989) narrow the term to explicit illocutionary predicates like ‘i believe that’ and ‘we demonstrate that’ in the text. fairclough (1992), too, has his own definition of metadiscourse. he sees metadiscourse as a kind of “manifest inertextuality” where the writer interacts with his or her own text. moreover, geisler (1994) refers to both metadiscourse and rhetorical processes, contrasting these with what she calls domain content (see hyland, 2005). however, in spite of all the points mentioned above, hyland (2005) claims that metadiscourse scholars generally agree to use metadiscourse in wider sense, the sense which refers to the various linguistic tokens employed to guide or direct a reader through a text so both the text and the writer’s stance is understood. in this way, one of the most influential and widely quoted definitions of metadiscourse is that of schiffrin’s (1980). she defines metadiscourse as the author’s manifestation in a text to “bracket the discourse organization and the expressive implications of what is being said” (p. 231). other influential definitions also exist. crismore (1983, p.2), for example, defines metadiscourse as “the author’s intrusion into the discourse, either explicitly or non-explicitly, to direct rather than inform, showing readers how to understand what is said and meant in the primary discourse and how to take the author.” in another place (crismore, 1985) she defines metadiscourse as the author’s discovery about the discourse and metacommunicative phenomena. in a similar way, vande kopple (1985) defines metadiscourse as those linguistic materials which does not add to propositional information, but indicate the author’s presence. williams (1981, p.226), too, defines metadiscourse as “whatever does not refer to the subject matter being addressed.” halliday (1978) also claims that metadiscourse includes discourse elements that refer not to aspects of reality but to the organization of discourse, and the relationship that develops between the author and the reader. in another place (halliday, 1994) he sees proposition as something that can be argued about, affirmed, denied, doubted, regretted, and so on. however, it should also be taken into consideration that this distinction between propositional content and metadiscourse which sees the role of metadiscourse as secondary to that primary role of propositional content has been criticized by hyland (2005, p.19), who calls this distinction as “disturbing.” he believes that this distinction unnecessarily elevates the transactional function above the interactional one. hyland and tse (2004) also argue that it is difficult to see how metadiscourse can constitute a different level of meaning. they believe that it is certainly possible, even commonplace, to distinguish the propositional content of each text from the way it is expressed, but, this, according to them, does not imply that metadiscourse can be easily and inadvertently omitted from a text without changing its meaning. hyland (2005, p.24) briefly summarizes the point made in the previous paragraph that metadiscourse is not a separate level of meaning: …meaning is not synonymous with ‘content’ but dependent on all the components of a text. both propositional and metadiscoursal elements occur together, generally in the same sentences, and we should not be surprised that a stretch of discourse may have both functions. such integration is common, with each element expressing its own ‘content’: one concerned with the world and the other with the text and its reception. an influential definition which, according to him, preserves the distinction between proposition and metadiscourse without assigning the former the primary role is that of beauvais (1989) based on austin’s theory of speech acts (1962). according to this definition metadiscourse can be regarded as explicit markers which help readers to identify how a writer’s arguments are to be understood. metadiscourse can also be looked at from the view point of natural language philosophers such as bertrand russell, who believe that it is the emphasis on logical structures and problems that should be regarded as important. such philosophers separate language into ‘object language’ and ‘metalanguage object’. for them language is used to refer to the world or reality, but metadiscourse is used to refer to language. hyland (2005, p.13) also defines metadiscourse as “an important link between a text and its context as it points to the expectations readers have for certain forms of interactions and engagement.” he further argues that it is by revealing a writer’s understanding of an audience through the ways that he/she addresses readers that the dialogic role of the discourse is fulfilled. in this way, a text has to talk to readers or hearers in ways that they find familiar and acceptable. in sum, it can be argued that what is common in almost all the definitions of metadiscourse available is the notion that metadiscourse does not add anything to the propositional content at question, it only guides the readers how to engage with the text and interact with what is being presented. 2. 8. 3. metadiscourse: functional or syntactic an important issue in metadiscourse studies has always been whether metadiscourse is a syntactic-functional category or only a functional category. hyland (2005) mentions the major scholars in both directions. according to him, while crismore et al. (1993) believe that metadiscourse is simultaneously both functional and syntactic, lautamatti (1978), meyer (1975), and williams (1981) have adopted a functional approach to metadiscourse, classifying metadiscourse markers according to the functions they perform in a text. he himself prefers the latter approach and argues that in functional analysis of metadiscourse what is important is how language is used to fulfill certain goals. what counts in this kind of analysis is how the language is used in relation to the context at question (hyland 2005, p.24): functional analyses recognize that a comprehensive and pragmatically grounded description of any text must involve attending to the use of language in relation to its surrounding co-text and the purpose of the writer in creating a text as a whole. the emphasis is therefore on meanings in context, how language is used, not what a dictionary says about it. so, when considering any item as a candidate for inclusion as metadiscourse, the question is not ‘what is the function of this item?’ but ‘what is this item doing here at this point in the text?’ in this way it is possible that one item may function as metadiscourse in one text but as proposition in another. the following examples adapted from hyland (2005) may clarify the point: 1. i want to agree about the data, then we can talk about the venue. 2. i was waiting an hour then he told me the train had already left. while in the first sentence ‘then’ is an example of metadiscourse because it explains how the speaker is going to organize his discussion, in the second one it is an example of proposition because it tells the reader how events followed. 2.8. 4. functions of metadiscourse metadiscourse researchers and scholars have always looked at systemic functional theory of language within systemic functional linguistics of halliday for insights (cf. de joia & stenton, 1980; halliday, 1978; halliday, 1994; halliday & matthiessen, 1999). in this theory, language is organized around three broad functions of ideational, interpersonal, and textual: 1. ideational function which is used in order to organize the speakers’ or writers’ experiences of the real or imaginary world. 2. interpersonal function which is used to indicate, establish, or maintain social relationships between people. 3. textual function which is used to create written or spoken texts which cohere with themselves and which fit the particular situation in which they are used. metadiscourse analysts believe that metadiscourse is used in order to satisfy the second and the third functions of language, namely the interpersonal and textual ones discussed above. in this way, vande kopple (1985, p.87) believes that textual metadiscourse “shows how we link and relate individual propositions so that they form a cohesive and coherent text and how individual elements of those propositions make sense in conjunction with other elements of the text.” he also regards interpersonal metadiscourse markers as communication about communication in that they can help us express our personalities and our reactions to the propositional content of our texts and characterize the interaction we would like to have with our readers about that content. any truth value that these conveyors of interpersonal meanings carry would relate not to state of affairs in the world but to such things as our actual personalities, our own evaluations of propositional materials, our role in the situation in which the text functions, and our hopes for the kinds of responses readers might make (pp. 86-87). but although this way of looking at those language functions proposed by halliday, that is looking at them as separable and discreet has been an influential one in metadiscourse analysis for years, recently it has been criticized by hyland (2005). he believes that halliday tends to look at those functions not as separate entities but as a whole, and metadiscourse is about interpersonal dimension and utilizes both interactive and textual resources: discourse is a process in which writers are simultaneously creating propositional content, interpersonal engagement and the flow of the text as they write. but in this process the creation of text is a means of creating both interpersonal and ideational meanings, and textual features can not be seen as ends in themselves. if metadiscourse is the way writers engage their readers and create convincing and coherent text, then we have to acknowledge that it is about interaction in text. it expresses the interpersonal dimension and how both interactive and textual resources are used to create and maintain relations with readers (p.27). to summarize, systemic functional theory of language has always been an important theory in metadiscourse analysis. most metadiscourse analysts believe that metadiscourse satisfies the interpersonal and textual functions of language proposed in this theory, although some scholars tend to see it as fulfilling just the interpersonal function of language, believing that both interactive and textual resources are used in this regard. 2. 8. 5. metadiscourse signals there is lack of clarity in the metadiscourse literature about what should be counted as metadiscourse. in fact, metadiscourse can be regarded as an open category to which, as hyland (2005, p.27) says, “writers are able to add new items according to the needs of the context”. all the categories of metadiscourse are, in fact, complex, fuzzy, and non-discrete since metadiscourse is a subject that is extremely disheveled (appendix i shows 300 potential expressions which are usually regarded as metadiscourse). although most researchers only regard the explicit textual devices as metadiscourse, argyle (1972), for example, regards non-verbal signals, too, as metadiscourse. moreover, crismore et al. (1993) also regard forms of punctuation and typographical marks to be instances of metadiscourse (see figure 2.3). paralanguage oral proxemics potential non-verbal metadiscourse signals kinesics printing written genre/media punctuation figure 2.3. potential non-verbal expressions of metadiscourse (adapted from hyland, 2005) according to argyle (1972), paralanguage includes intonation, stress, volume, and voice quality; proxemics includes orientation, touch, and physical distance; and, finally, kinesics includes gesture, posture, dress, appearance, and facial expression. concerning written non-verbal metadiscourse signals, crismore et al. (1993), regard binding quality, paper quality, colour, font size and type, and publisher status as the sub-categories of printing; book, letter, post card, and so on as examples of genre and media; and scare quotes, underlying, italics, bolds, exclamation marks and emoticons as examples of punctuation. besides those controversies discussed above, hyland (2005, pp.30-31) mentions the following three reasons why sometimes it is difficult to identify whether one item counts as metadiscourse or not: 1. there is invariably a certain amount of insider opacity in metadiscourse use. 2. writer’s expression of affect or lexical evaluation is an important consideration of what should be regarded as metadiscourse. in this way the analyst should distinguish between evaluative lexis, used to qualify individual items, and stance markers, which provide an attitudinal or evaluative frame for an entire proposition. only those items in the latter group are an example of metadiscourse. 3. there are opaque features of language (like metaphors and allusions) which might signal metadiscoursal meanings. but, in spite of these difficulties, some categorizations have been proposed by scholars in the field; and, in the next section the most important and influential of them will be discussed in detail. 2. 8. 6. categorizations of metadiscourse 2. 8. 6. 1. lautamatti and william’s categorizations as it was mentioned in the last section, because of the fact that wide ranges of meanings are realized by metadiscourse markers, different scholars have different categorizations of metadiscourse markers. the first important categorization belongs to lautamatti (1978). she refers to metadiscourse as ‘non-topical material’, contrasting it with ‘topical material’ which is the name of the material related to discourse topic. she identifies five types of metadiscourse: 1. metatextual markers: material used to comment on the discourse itself and may refer to the language of the text or the properties of the organization of the text itself. 2. illocutionary markers: material used to comment on and make explicit to the readers the illocutionary force of the utterance. 3. commentary markers: material used to comment directly to the readers. 4. modality markers: material used to comment on the truth value of the ideas expressed. 5. attitude markers: material used to make explicit the author’s attitude towards what is being discussed. another important categorization of metadiscourse markers belong to williams (1981). he identifies three broad categories of metadiscourse: 1. hedges and emphatics: expressions which indicate the status or truth value what is being said. 2. sequences and topicalizers: expressions which lead the reader through the text. 3. attributors and narrators: expressions which tell the reader where the facts expressed come from. 2. 8. 6. 2. vande kopple’s categorization perhaps one of the most widely used categorizations belongs to vande kopple (1985). in fact, most taxonomies of metadiscourse are more or less based on this kind of categorization. this categorization is itself based on those proposed by lautamatti (1978) and williams (1981) discussed above. in his categorization, vande kopple has compared those two categorizations and expanded upon their presentations. he divides metadiscourse into two broad categories of textual metadiscourse and interpersonal metadiscourse . according to him, textual metadiscourse includes four sub-categories of text connectives, code glosses, validity markers, and narrators; and interpersonal metadiscourse includes three subcategories of illocution markers, attitude markers, and commentaries. although this kind of categorization is one of the most widely used categorizations of metadiscourse markers, it is not without its critics. one scholar (hyland, 2005), for example, criticizes vande kopple on the grounds that: 1. it is difficult, in that categorization, to distinguish narrators and attributors particularly in academic writing where citation is used to perform a variety of rhetorical functions. the belief is that not only can citations provide propositional warrants (validity markers in vande kopple’s terms) and meet conventions of precedence (narrators), but they might also be used to offer a narrative context for the research (berkenkotter & huckin, 1995) or establish an intertextual framework to suggest a cumulative and linear progression of knowledge (hyland, 2000). 2. in that categorization, functions are not performed in isolation, and the writer may be trying to achieve several purposes at once in selecting a citation; and, as a result, it is not clear how far either the analyst or the reader can determine which function may be intended. 3. it is also difficult, in vande kopple’s categorization, to distinguish between illocution and validity markers, and commentary and attitude markers: similar problems occur when we try to disentangle examples of illocution and validity markers where cases such as ‘we suggest that’ and ‘i demonstrate that’ seem to indicate both the degree of commitment that the writer wishes to invest in a statement and simultaneously the act that the discourse is performing at that point. nor is it obvious what the category of commentary might include or how this is likely to differ from that which might be labeled as attitude markers (hyland 2005, p. 33). because of the criticisms leveled against vande kopple’s categorization of metadiscourse markers, some scholars have proposed their own categorizations of metadiscourse markers. they have reorganized vande kopple’s categories in the ways that will be elaborated on in the next section. but, finally, it should also be mentioned that recently vande kopple (2002), in order to amend his categorization, himself has made some changes to his original categorization by re-labeling validity markers as epistemology markers and including narrators in that category. 2. 8. 6. 3. crismore et al.’s (1993) categorization crismore et al (1993) modified and reorganized vande kopple’s (1985) categorization of metadiscourse markers. their revision is perhaps one of the most substantial ones made to that of vande kopple