1. Introduction

A text or discourse is not just a series or combination of sentences for introducing different random topics. It is combining sentences in a logical way, according to their meaning and that helps to create unity for a text. This is what we call cohesion in which sentences stick together to function as a whole. Cohesion, the most important principle and criterion of textuality, is the connection or the connectedness manifested when the interpretation of one textual element in the text (a word usually but not necessarily in another sentence.

The term cohesion has been defined in various ways. Some researchers apply the term “cohesion” to the surface structure of the text. Cohesion sometimes was applied to smaller units of language in the text. According to Hoey (1991), cohesion may be crudely defined as the certain words or grammatical features of a sentence that can connect the sentence to its predecessors (and successors) in a text. It is usually one of the important elements in creating a discourse that communicates effectively and naturally. It is a linguistic phenomenon in a discourse which assists the hearer to understand and perceive the text as a single unit. It is impossible for a collection of unrelated sentences to make sense to the audience unless all the sentences are tied together to form a unified whole or a discourse by cohesion.

Dooley and Levinsohn (2001) state that cohesion is achieved by “using linguistic signals in the text as clues to help hearers in coming up with an adequate mental representation. Within a discourse structure, these linguistic signals function as a link which glues the individual parts of discourse together. In addition, Pickering’s (1978) discussion in the introduction paragraph of Cohesion implies that how much and fast a person’s mind can accommodate new information depends on how much that piece of new information relies on what the person has already known, that is, old information. This statement relates directly to the concept of cohesion.

Pickering (1978) also views cohesion as something which ties a discourse together in a linear way. Halliday and Hasan (1976, p. 4) explain that cohesion occurs where the interpretation of some element in the discourse is dependent on that of another element and that one presupposes the other, in the sense that it cannot be effectively decoded except by recourse to it. Thus, cohesion is the use of linguistic means or linguistic devices to knit the unity of a text. Cohesion, the most important principle and criterion of textuality, is the connection or connectedness manifested when the interpretation of one textual element (a word located in one sentence) is dependent on another element in the text (a word usually but not necessarily in another sentence). Cohesion relates to the “semantic ties” within text whereby a tie is made when there is some dependent link between items that combine to create meaning. The foundations of text linguistics were laid down by Halliday and Hasan’s “Cohesion in English” in 1976. Cohesion is defined as the set of linguistic means we have available for creating texture (Halliday & Hasan, 1976, p. 2), i.e., the property of a text of being an interpretable whole (rather than unconnected sentences). The purpose of the present study is to compare English political articles in mass media written by American native and Iranian non-native authors concerning the use of references as cohesive devices.

This study may help to accelerate the knowledge of ESP students who are going to be the future writers in this domain. Findings of the present research may provide various types of references which are different in the discussion section. It is worth noting that the use of references in political articles can be realized differently among native and non-native authors. By comparing the American native with those of Iranian non-native corpora, one can draw his/her attention to how different authors treat differently with references to make their discussion sections cohesive. Its pedagogical recommendations can lead to enhance the knowledge of those instructors who teach academic political writing and make them be conscious of the standard references of “discussion” section as a subgenre.

This may make their students aware that such references are to be included in their political article discussions. Actually, this will help political students be sensitive to the format of discussion section and the design of references within the text which affect the readers’ comprehension significantly. The role of discussion as an argumentative text could be revealed through analyzing and comparing the native and non-native political discussion section in the present study. In a nutshell, what manifests as the product of this analysis will be utilized for pedagogical implementations including writing and analyzing political papers, especially in ESP courses.

2. Review of the Literature

Cohesion has been studied thoroughly in various aspects. According to Halliday and Hasan (1976), cohesion enables us to create a text. It is defined by Halliday and Hasan (1976) as the “set of semantic configuration that is typically associated with a particular class of context of situation, and defines the substance of the text”. In their view, the function of cohesion is to relate one part of the text to another part of the same text. Consequently, it lends continuity of the text. Most scholars (Halliday & Hasan (1976), Beaugrande and Dressler (1981), Lyons (1995) define cohesion as the network of lexical, grammatical, and other relations.

Cohesion has gained prominence in studies on discourse analysis as well as L1/L2 writing research following Halliday and Hasan (1976) seminal work on Cohesion in English. Halliday and Hasan (1976) defined cohesion as “the set of possibilities that exist in the language for making text hang together” (p.18). In a similar vein, Hinkel (2003) conceptualized cohesion as “the connectivity of ideas in discourse and sentences to one another in text, thus creating the flow of information in a unified way.” (p. 279). For Halliday and Hasan (1976) and other researchers in Hallidain tradition, the organization of text is made up of relationships among items in the text and those relationships are realized through the use of cohesive devices. Reid (1992) further extrapolated cohesion devices as “words or phrases that act as signals to the reader; those words or phrases make what is being stated relate to what has already been stated or what soon will be stated” (p. 81). Cohesion is divided into two subcategories: grammatical cohesion and lexical cohesion. Grammatical cohesion in then divided into five main categories: reference, ellipsis, substitution, conjunction. Reference occurs whenever an item indicates that the identity of what is being talked about can be retrieved from the immediate context. Pronouns, determiners, definite articles, and comparatives such as he, this, the, less are reference items. The interpretation of the reference elements depends upon presupposed information contained in the sentences immediately above it.

According to Halliday and Hasan (1976), such cohesive devices which mentioned previously serve to contribute to text cohesion. They believe that text cohesion leads to greater text coherence which in turn enhances quality of writing. Although Halliday and Hasan (1976) did not consider issues of language pedagogy in their research, the effective use of cohesive devices has been identified as one of the important criteria for good writing and thus considered as something to be treated in a pedagogical context (Hinkel, 2001). The theory and practice of grammatical cohesion has been widely analyzed by many foreign linguists such as Halliday and Hasan (1976), Beaugrande and Dressler (1981), Lyons (1995), Baker (1992), Yule (1996) as well as Lithuanian scholars, Verikaite (1999), Valeika (2001), Valeika and Buitkiene (2006).

2.1. The Concept of References

Brown and Yule (1983, pp. 27-28) see the nature of reference in text and in discourse as an action on the part of a speaker/writer. It describes what they are doing “not the relationship which exists between one sentence or proposition and another.” Mc Carthy (1991, p. 37) states that we must consider the notion of discourse segments as “functional units, rather than concentrating on sentences and to see the writer/speaker as faced with a number of strategic choices as to how to present them to the receiver.” He adds that reference items can refer to segments of discourse or situations as a whole rather than to any one specified entity in that situation. The main feature that characterizes reference is that the information signals for retrieval. The identity of particular thing that is being referred to has a referential meaning and cohesion is found then the same thing occurs a second time. Reference has the semantic feature of definiteness or specificity. Because of that there has to be reference to the context of situation. Referencing items do not have to match the grammatical class they must have semantic properties (Halliday & Hasan, 1976, p. 31).

Linguistics provides various definitions of the concept of reference. Lyons (1995, p. 293) defines reference as the relation that exists between “Linguistic expressions and what they stand for in the world (or the universe of discourse) on particular occasions of utterance.” Yule (1996, p. 17) describes reference as “an act in which a speaker, or writer, uses linguistic forms to enable a listener, or reader, to identify something.” To quote Baker (1992, p. 181) “reference is a device which allows the reader/hearer to trace participants, entities, events, etc. in a text.” According to Schiffrin et al. (2001, p. 36) “reference refers to resources for referring to a participant or circumstantial element whose identity is recoverable.”

Halliday and Hasan (1976, pp. 32-33) present situational and textual references. Textual reference is reference to another item within the text derived from situational reference which is considered to be the prior form of reference. That means reference in a linguistic context is only secondary or derived from a reference in situational context. There are special terms for situational and textual reference, situational reference is named exophora or exophoric and general term for reference within a text is endophora. Endophora is divided into anaphora and cataphora. An exophoric item does not name anything and does not give a significantly greater amount of information. It signals that the reference must be made to the context of situation from elsewhere. Exophoric reference refers to the creation of text; it links the language with the context of situation. Its reference is not cohesive because it does not bind two elements into text; it takes us outside the text for interpretation. Only endophoric reference is cohesive and it contributes to the integration of the passage with another to form coherent text. These two instances demonstrate that a reference item is not exophoric or endophoric itself, it simply has the property of reference, and it suggests looking somewhere else (Halliday & Hasan, 1976, p. 37). Endophoric relations are of two types, those which look back in the text for their interpretation (anaphoric relations) and those which look forward to the text for their interpretation (cataphoric relations). For instance, the following sentences show the use of the references.

Ÿ There was an orange on the Table. So I ate it.

Ÿ The woman prepared the dinner. She used a lot of seasoning.

In the first sentence above, ‘It’ refers back to ‘an orange’ while ‘she’ in the last sentence refers back to “the woman”. This kind of references is referring to an anaphora (i.e. looking backward). The other kind of reference, where the pronoun is given first and then kept in suspense as to its identity, which is revealed later, is known as cataphora (i.e. looking forward). Examples:

Ÿ He was aggressive. My Boss.

Ÿ He made tremendous impact. The Provost.

Yule (1996, p. 17) claims that words themselves do not refer to anything. The speaker uses words to refer to entities in the world by using appropriate expression. Those linguistic forms that are being used to refer to something can be defined as referring expressions (Yule, 1995, p. 17). Every referring expression has a range of reference, which is a number of possible referents of the referring expression. The referential range is defined by their meaning in the language and contextual factors (Lyons, 1995, p. 294). When we chose one type of referring expression rather than another we must assume the context is known for the listener otherwise we must use more elaborate noun phrases. The listener must infer correctly what the speaker wants to identify by using a particular referring expression. Sometimes it is not clear which referring expression would be the best to use. The expressions must be invented. We can only rely on the listener’s ability to understand what referent we have in mind (Yule, 1996, pp. 17-18).

Many of earlier cohesion studies conducted in the 1980s within a pedagogical context attempted to discover whether there is a significant correlation between use of cohesive devices and quality of writing in L1 texts. Inspired by the studies on cohesion in L1 writing, a number of studies have been conducted in L2 contexts as well in an attempt to investigate the relationship between cohesion and quality of writing. Zhang (2000, 2004) in a study investigated the use of cohesive features in the expository compositions of Chinese undergraduates, using both quantitative and qualitative methods. One hundred and seven essays were collected from two Chinese universities and assessed by three raters. Halliday and Hasan’s (1976) taxonomy of cohesive devices and their framework for analysis were used. Students employed in their writing a variety of cohesive devices with some categories of ties used more frequently than others. Lexical devices were the most frequently used, followed by conjunctions and reference devices. In terms of tie distances, the majority of the cohesive ties were either immediate or remote. There was no statistically significant relationship between the number of cohesive ties used and the quality of writing. Certain cohesive features were identified in the expository writing of Chinese undergraduates which included ambiguity in reference, overuse and misuse of conjunctions, and restricted use of lexical cohesion.

3. Methodology

3.1. Data

The research sample that is used in this study consists of 200 political articles and news i.e., American native researchers (100 written by NRs of English and 100 written by Iranian non-native researchers of English). They were randomly chosen from recent publications in prestigious international online newspapers and magazines such as (US news, Yahoo news, Newsweek magazine, Jame Jamonline, & Keihan International) in the field of politics. It should be mentioned that they were news articles of media, not the research articles published in political journals. In order to specify whether the native articles are written by American writers, the CV of each writer, that is, the country of birth and the University of their Studies were checked online. The data set were developed by selecting articles published internationally and nationally from 2011 to 2014. In order to compile more reliable and recent corpora, only the articles published since 2011 onwards were selected. There are some rationales behind this selection:

1. Firstly, many natives and non-native authors of English have written significant articles regarding this domain in various journals, etc.

2. Secondly, a bundle of theses have conducted studies on other corpora such as economics, chemistry, physics and so on.

3. Thirdly, the availability of these political articles in Mass Media like magazines, newspapers, or even on-line websites relevant to politics (e.g., News Week, Iran Daily, Tehran Times, JameJam online, US news, Iran news, Keihan International, etc.) makes data collection much more feasible.

When conducting studies in contrastive analysis, Connor, Nagelhout, and Rozycki (2008) pointed out that it is important to establish a valid criterion of comparison between data, in other words, to examine sets comparable original texts with “maximum similarity” written in two or more languages. Although one set of articles in the current study were written by Iranian non-native authors, their English writing proficiency can be taken to be at a native or near-native level since the articles have been published in these widely-known national magazines and journals.

As mentioned above, this study relied on just English political articles in two set of corpus written by Iranian natives and American natives, which were compared in terms of references. The first section, political articles written by native authors, comprised of 100 articles with 64,046 words. The latter part, articles written by Iranian writers, comprised of 100 articles with 64,057 words in the articles. On the whole, 200 articles which included 93.019 words were analyzed. The reason for selecting political articles published in media was that writing articles appears to be a very complicated activity with many visible and invisible layers. Moreover, in order to communicate effectively with English language countries via media (e.g., newspapers, magazines, etc.), knowledge of lingua franca is a pre-requisite and lack of this knowledge leads to misunderstanding. Thus, this comparative analysis enhances native authors’ proficiency in writing highly-qualified political news articles.

3.2. Instrumentation

The present study employed a framework for the analysis of references. It was extracted from Halliday’s (1994) classification of references based on Halliday (1994). This checklist contains three or four types of references concerned with cohesive devices including cohesive, demonstrative, comparative and personal references. The references were compared to other sources (e.g., Comparative Analysis by James, 1980 & Discourse Analysis by Brown & Yule, 1983).

3.3. Procedure

As mentioned before, the non-native corpus consists of 100 published articles or stories contributed by Iranian authors. Meanwhile, the native corpus contains 100 articles published in prestigious international journals mentioned earlier (see Appendix B). Based on the above-mentioned checklist, we extracted all the references used by writers. This researcher read the articles completely to find the full references. The selected RAs which were obtained directly from online magazines and newspapers, converted to Word Format. So, we could count the total number of words in each article and find the references one by one with (Find) section in the Microsoft Word 2007 in order to calculate the distribution of references. Our main focus was only to analyze the function of references not the structures. At first, we highlighted all the references in (Find) section of Microsoft Word, then we began counting those which merely had the function of reference and skipped those which had other functions. For example, sometimes that had the function of a cohesive reference, but sometimes a conjunction. e.g., Friends of former president George Bush are worried that his health may be in a dangerous decline. The latter one was ignored in counting. Or it mostly was a cohesive reference and sometimes acted as an expletive. For example: It’s inappropriate for Trump to moderate the event. In addition, the references which equaled to zero or one, were not taken into consideration because their frequency was low. Some references overlapped. In other words, they had more than one function and belonged to two or three categories. Thus, it made the decision challenging for the researcher to put them in which category. Because of that, among the nine categories, only four of them which were the most frequent were chosen in order to narrow down the scope of analysis because this research cannot cover all of them. For instance, ambiguous pronoun reference was totally removed from the checklist due to multifunction of some references. The references were calculated twice to estimate inter-rater reliability. The results are presented in Table 1.

Table 1. Inter-rater Reliability of the References

By this descriptive analysis, the researcher can determine the frequencies of the selected references in native and non-native articles and stories within the mentioned field. The quantitative analysis also enables the author to come to an appropriate conclusion of any difference or similarity on the using of the references. All the references in all the articles were calculated. As a result, the total word count performed by Microsoft Word Office for the non-native data was about 64,057 words and 64,046 words for the native corpus. In the final stage, various references will be identified using Non-parametric Chi-square procedure, which is a technique for looking at how references function in the corpora. The sole objective of this test here is to identify the probable significant concordance among references in these two separated collection of articles. And in case of no compatibility, or at least slight correspondence, we can figure out how and why non-natives prefer some specific references and overlook some others which natives do not.

In summary, 200 articles from one discipline in two languages were randomly chosen from recent published leading journals. First, the overall organization of the articles was analyzed based on Halliday’s (1994) model. Then the frequency of references was analyzed in the English political articles. The results of these analyses are presented in Chapter Four.

4. Results

4.1. Results of Descriptive Analysis

The first step taken in the analysis of references in the mentioned articles was to run word count to determine the length of the two corpora. A total of 4353 references were identified among 64.057 words in the Iranian non-native corpus, and 6308 references were recognized among 64.046 words in American native data.

4.2. Frequency and Percentage of References

The two sets of political articles investigated in this study were analyzed concerning the frequency of occurrence of references in each of the 4 categories of the checklist used in this study. The data presented in the following tables show the statistics which were obtained after the analysis of the articles; rows numbered 1-4 represent the taxonomy applied here in the present study. The frequency and percentage of all references in each of the two groups (data) of political articles under study are shown under the columns of native and non-native corpora (data) as well. Total number of references is also given.

Table 2. Native American and Non-native Iranian Articles Using References

4.3. Frequency of Occurrence of References in Articles

1. Cohesive References

The first category in the classification used here represented the cohesive references. Different cohesive references were found which were mostly situated in texts on politics by native writers. However, the least amount of this category was found in political articles written by Iranian non-native authors. The sum of the occurrence of this category in the two groups of articles equaled 8851 cases (83.02% of all the references) among which the definite article “the” had the highest frequency in both native and non-native articles while “yesterday” had the lowest.

2. Demonstrative References

The most frequent demonstrative reference was found in political articles written by native writers. This category had the frequency of 172 cases in all the two groups (1.61% of all the references) among which the demonstrative reference “those” had the highest frequency and clearly “then” the least frequency in both native and non-native articles (Appendix B).

3. Comparative References

This category occurred 681 times in the two groups (6.38% of all the references). The most frequent type of this category was found in articles by American native authors and clearly the least amount of this category was found in articles by Iranian non-native writers. Among which, in native corpora “other” had the highest and “identical”, “differently” and “equally” had the lowest frequency. Nevertheless, “more” had the most and five items, that is, “additional”, “else”, “identically”, and “likewise” had the lowest frequencies in non-native articles.

4. Personal References

The fourth category of references in the classification of references in the present study deals with the personal references. The overall number occurrence of this category was 957 cases equal to 8.97% of the references highlighted in this study. The most frequent type of this category was found in articles on politics by non-native writers, while the least amount of this category was found in American native articles (Appendix B).

4.4. Descriptive Statistics

There were some differences in the frequency of references used by American and Iranian writers of the news articles.

Table 3. Descriptive Statistics

In this Table 3, N is the number of four items of references we considered in each group. The obtained mean was 1577 for American native articles and 1088 for Iranian non-native articles. It shows that Iranian writers used fewer references in comparison to American native ones. Since descriptive statistics cannot show the significant difference between the groups, data were analyzed through Chi-square analysis. According to Table 3, there are four categories of references in both groups which are compared. The American NSs and Iranian NSs used references differently in their political news and editorials.

4.5. Chi-Square Calculator

The contingency table was used to do the Chi-square calculation. The significance level was at (p < 0.05.) which indicated the difference between each category of the references between American NRs and Iranian NNRs regarding the use of references in the political texts.

Table 4. Frequency of References

The Table 4 shows that cohesive references have the highest frequency and demonstrative references have the lowest frequency in both American native and Iranian non-native articles. It also indicates that Iranian non-native writers used fewer references in comparison to American native writers in writing news articles.

5. Discussion and Conclusions

5.1. Discussion

In this section, the results were discussed to give the possible reasons for the results obtained from the study. Moreover, it gives answers to the research questions.

1. Does the frequency of references differ significantly across English political articles written by American native and Iranian non-native researchers?

In answering the first question, the researcher proposed the following discussion: The results of the Chi-square analysis showed that there are some differences and similarities between native and non-native news articles. The most frequent types of references used by American researchers were cohesive references, and the least used ones were demonstrative references. For instance, the following examples show this phenomenon.

1. In the end, Gingrich emerged where he started under sharp attack from his rivals and under increasing security by the media as he tries to maintain his position as the GOP front-runner.

2. People who heard the debate on the radio thought Nixon won but those who saw it on TV thought he lost.

In Iranian non-native articles, the most frequent items of references were cohesive references, but the least used items were demonstrative references. For example, the use of “he and those” in the following examples.

1. On Iran’s nuclear program, Gates has said he still believes Iranian leaders are intent on building a nuclear weapon and are “getting closer.”

2. A senior advisor to the Islamic Revolution leader says the European Union serves as a tool in the hands of the U.S., adding the policies of the bloc are influenced by those of Washington.

In native news articles, 6308 words out of 64046 were references whereas in non-native news articles they were 4353 words out of 64054. Generally, the writers of American news articles used more references than those of Iranian ones. The probable reasons and justifications for the obtained results may be due to the mastery of native researchers in their writings. Due to the fact that there are some differences between English and Persian grammar, and overgeneralization which sometimes occur by non-native researchers in writing English political news. In other words, Iranian writers feel free to overuse some references in their own language and may omit some of them without clear reasons. The other reason may be due to transfering this habit in writing English texts coming from their L1. In fact, overuse of the references is open to criticism heavily reference-based might appear “laymanlike” and also readers may lose their trust in what writers try to express. On the other hand, because of uncertainty and lack of sufficient knowledge, some other writers prefer to underuse them. It may expose them with a lot of criticism on the part of readers. Thus, learning how to use references is a must. Difficulty with cohesive devices is either because of discourse conventions or because of the lack of familiarity of Persian writers with a wide variety of techniques in applying these devices. Differences in the use of cohesive references may be due to the knowledge of grammar Iranian researchers may have in using grammatical patterns which help them to feel self-confident in using grammatical elements without referring to the English native samples in using the pre-fabricated structures. Other references such as comparison, personal and demonstrative were not significantly different from Persian researchers’ use of references. This may be due to the lack of enough knowledge on using such references since they hold less frequency and Persian researchers may follow the English native researchers in using such structures.

The result of the present study is compatible with the following works including Abdul Rahman (2013) who has emphasized that there was a vast difference between the natives’ and the non-natives’ use of cohesive devices such as reference in frequency, variety, and control. Similarly, Ghasemi (2013) in his study concluded that there were distinct linguistic differences in the use of cohesive devices by native and non-native learners. In a similar vein, Hessamy and Hamedi (2013) in an attempt to compare and contrast the frequency of the use of cohesive devices in independent and integrated essays found out that there was a significant difference in the use of almost all types of cohesive devices between the two conditions.

2. Do non-native researchers use references functionally similar as American native ones in the political articles?

In American native news and articles, the frequency of cohesive references was 5464 cases, 91 demonstrative references, 306 comparative references, and 447 personal references. On the whole, the total number of used references was 6308. The most frequent types of references used by American writers were cohesive references, and the least used ones were demonstrative references. The following examples present the use of these references.

1. In the end, Gingrich emerged where he started under sharp attack from his rivals and under increasing security by the media as he tries to maintain his position as the GOP front-runner.

2. People who heard the debate on the radio thought Nixon won but those who saw it on TV thought he lost.

Meanwhile, the frequency of these four categories of references in Iranian non-native news articles were 3387 cohesive references, 81 demonstrative references, 375 comparative references, 510 personal references. Totally, Iranian non-native writers used 4353 references in writing news and articles. In Iranian non-native articles, the most frequent items of references were cohesive references, but the least used items were demonstrative references.

1. On Iran’s nuclear program, Gates has said he still believes Iranian leaders are intent on building a nuclear weapon and are “getting closer.”

2. A senior advisor to the Islamic Revolution leader says the European Union serves as a tool in the hands of the U.S., adding the policies of the bloc are influenced by those of Washington.

A comparative analysis between these two sets of articles indicates that Iranian non-native writers used fewer references than American native writers. This incompatibility may be due to several reasons proposed as follows:

1) Lack of mastery on cohesive markers can lead to inappropriate use of references.

2) Iranian students’ lack of variety in using references. For instance, some of comparative references are unique in use like otherwise, less, identically, etc. They may not have known the function of each reference and also the context they should use.

3) Teachers are not competent enough to teach references to students appropriately.

4) Sometimes students cannot distinguish form from the function. For example, students have difficulty comprehending the difference between the two functions of the same form such as it, this, or there which may be references or expletives in terms of function in different contexts.

5.2. Conclusions

Cohesive devices are important features of academic writing which connect the sentences together in a text and make it cohesive. The main purpose of cohesive devices is to help readers make logical connections between sentences. Writers use these devices in their academic writing in order to make their writing more accurate and comprehensible. The results of this study showed that the researchers of Iranian political news and articles used fewer references than those of American news. This underuse problem may be related to the impact of first language and culture of non-native writers on the use of references in the second language. Following Atai and Sadr (2006), results of this study showed that familiarizing and involving students with the rules of academic writing may improve their reading ability and can help them to know what kind of discourse they have to produce and understand in academic settings.

Finally, it can be said that even when Iranian non-native researchers have a good knowledge of references, they may be influenced by their first language and culture. Contrastive studies on the use of references in two or more languages could help the writers to be familiar with the differences between the Persian and English structures and cultures and the language use of discourse markers when they write the text.

The most important contribution of this study to English for academic purposes classrooms is that students seem to have little awareness of these cohesive devices and the interactional nature of reading in general. In Iranian context, special instruction should be integrated into ESP or EAP for political courses specially and for writing courses generally to help students become competent and successful writers. There are discrepancies in different disciplines in applying references. A better understanding of these discourse devices by Persian writers can help them to have a good command of using these cohesive devices to be able to write their essays articles with standard academic format and to introduce themselves as members of the academic discourse community. Despite the growing demand for English for specific purposes instruction in Iran, ESP courses are still limited to learning specific cohesive texts rather than a variety of these linguistic elements. With the continued expansion and participation in the international specific arena, much attention should be drawn to the design of ESP courses and the roles of references in binding the sentences together in an ESP passage. This can help to prepare learners for future professional communication. The materials should help learners to be familiar with different strategies which are helpful to comprehend the grammatical structures of the passages deeply. The results of the present study can be useful for both ESP developers and ESP text designers to explain the structure of disciplines which have specific characteristics in terms of using grammatical cohesive devices specially references. Finally, the results of this study can be useful for every EFL teacher and learner who is interested in learning the nature of English political news and articles and their linguistic structures.

For future research, it is hoped that more large-scale, corpus-based studies on non-native and native writers’ articles will be done to enable us to gain a more comprehensive picture on the interlanguage development of references by different group of learners and nations. Another research can be an investigation on the reasons of over-use or under-use of references by Iranian native authors. With regard to pedagogical implications, address to the effective practices in helping the learners to improve cohesion of their writing.

Other studies can provide a list of references in Persian and compare and contrast in order to predict what problems students might encounter based on the differences and to discover whether students over generalize and transfer the use of references from their first language to English. Further studies can be done to compare the frequency and position of references in native and non-native essays, compositions, and letters to provide important implications for the teaching of references in those genres. The focus of some other studies can be on those cohesive references which were not included in the proposed checklist.