2017, Cilt 7, Sayı 1, Sayfa(lar) 132-138
The Use of Discourse Markers in the Writings of Turkish Students of English as a Foreign Language: A Corpus Based Study
Namık Kemal Üniversitesi, Yabancı Diller Yüksekokulu, Tekirdağ, Türkiye
Keywords: Corpus based study, Discourse markers, Writings of Turkish students, Foreign language, English
The aim of this study was to investigate the discourse markers used by 104 elementary-level prep class students studying at Namık Kemal
University in Turkey. Students were required to write a paragraph with 80-100 words as part of their mid-term exam in the academic year
of 2013-2014. A small-size corpus was constructed by using these writings. The corpus was analyzed via a software program called as
AntConc 3.2.4. in order to find out the types and frequency of discourse markers. It was revealed that 180 discourse markers were used by
elementary-level students: and was used 98 times, but occurred 51 times, because was written 18 times and other discourse markers of
then, so, also, too and still were used 7, 2, 2, 1, 1 times respectively. Furthermore, according to Frasers (1999) taxonomy of discourse
markers, 180 discourse markers were grouped into four categories. It was found out that 101 markers were elaborative markers, 52 were
contrastive markers, 18 were causative markers and 9 were inferential markers.
According to Prommas and Sinwongsuwat (2011), the purpose
of writing is not only to note down the sentences in the
accurate way but also to create cohesion and coherence in the
texts using cohesive markers such as discourse markers. For
this reason, discourse markers have been investigated for the
past fifteen years for the purpose of understanding what they
are, what they mean and what their functions are from the
perspective of different approaches (Lahuerta Martinez, 2004).
Therefore, among various terms used to refer to cohesive
devices, the term discourse marker was chosen in this study.
Rahimi (2011) states that the use of discourse markers were
examined in many languages such as Chinese, Danish, Finnish,
French, German and Hebrew and they were also investigated
in different genres such as classroom, newspaper, radio talk,
political interviews, and tutorial sessions. There are also studies
that have investigated the use of discourse markers by L2
learners but the writings of low-level learners have not studied
before. For this reason, this study examines the discourse
markers used by elementary level foreign language learners in
their writings since writing in a target language is much more
complicated due to this groups low proficiency level (Jalilifar,
The Aim of the Study
The aim of this study was to examine the frequency and function
of discourse markers used by elementary-level students in
Significance of the Study
Norrish (1983) mentions that writing has been known as the
most difficult language skill, even for native speakers of one
language. Hence, this skill has been the center of many studies.
However, the writings of low-level learners have not been
investigated until this study. This study attempts to find out the
uses of discourse markers and their frequency and functions in
The following research questions are examined for the purpose
of this study.
1. Which discourse markers are used by elementary-level students
in their writings?
2. What are the frequency and function of discourse markers
in students writings?
Cohesion and Coherence in a text
Brown and Yule (1983: 190) defines the text as the verbal
record of communicative event while Halliday and Hasan
(1976: 1) interpret it as any passage, spoken or written, of
whatever length, that does form a unified whole. According
to these definitions the text is not a collection of unrelated or
disconnected sentences and it is not just like putting the parts
together and making a whole out of it; there should be relationship between the sentences (Sadeghi and Kargar, 2014:
329). Thus, a written text requires cohesion and coherence created
by cohesive devices and it is achieved by constructing sentences
properly by cohesive ties (Halliday and Hasan, 1976). On
the other hand, coherence is the semantic relations that allow
a text to be understood and used and it is based on writers
purpose, the audiences knowledge and expectation (Witte
and Faigley, 1981: 202). Apparently, it could be stated that
coherence does not have as clear definition as cohesion (Wang
and Guo, 2014). As a consequence, this study mainly employs
the construction of cohesion by cohesive devices rather than
coherence. Moreover, it will also study and analyze discourse
markers as cohesive devices which create a meaningful text by
tying sentences since to communicate appropriately in written
texts, it is essential for students to learn about cohesive
and coherent devices (Sadeghi and Kargar, 2014:329).
According to Sadeghi and Kargar (2014: 329) Discourse markers
are lexical terms and link the segments in discourse.
Furthermore, Zarei defines them as words or phrases that
function within the linguistic system to establish relationships
between topics or grammatical units in discourse (as with the
use of words like because, so, then) (Zarei, 2013:108). However,
defining the term discourse marker is so complex as Sadeghi
and Kargar (2014) interpret that discourse marker is too
complicated to make a clear definition and to state its functions
easily since their functions may change according to the scholars
view. Consequently, many terms are employed instead of
discourse markers such as comment clause, connective, continuer,
discourse connective, discourse-deictic item, discourse
operator, discourse particle, discourse-shift marker, discourse
word, filler, fumble, gambit, hedge, initiator, interjection,
marker, marker of pragmatic structure, parenthetic phrase,
(void) pragmatic connective, pragmatic expression, pragmatic
particle and reaction signal (Brinton, 1996: 29). Brinton also
uses pragmatic marker in her book. Bell (2010) mentions
some terms such as pragmatic connectives (van Dijk, 1979),
discourse particles (Schourup, 1985) and discourse connectives
(Warner, 1985; Blakemore, 1987) which are used to refer
to discourse markers (Bell, 2010: 515). All these differences
demonstrate that discourse markers are studied for different
linguistic approaches (Urgelles-Coll, 2010). Moreover, it is also
stated that discourse markers are one of the most ambiguous
phenomena in linguistics (Polat, 2011:3746). According to
Fraser all these terms have a common feature. They impose
a relationship between some aspect of the discourse segment
they are a part of, call it S2, and some aspect of a prior discourse
segment call it S1 (Fraser, 1999: 938).
Brinton summarizes the characteristics of discourse markers
used as pragmatic markers (1996:33-34). They are generally
used in oral discourse due to informality and spontaneity of
speech. However, in the written discourse the structure and
the reasons of use might be totally different. The markers may
be used more than once in a sentence in informal or spoken
discourse. In spite of their frequent usage in spoken discourse,
discourse markers should be used appropriately and carefully in written and formal discourse. These generally favour sentence-
initial position but they may occupy mid or end position
as well. The translation of pragmatic markers into another
language is highly complex due to their semantic shallowness
(Svartvik, 1979 and Stubbs, 1983:69 cited in Brinton,
1996:34). The speakers or the writers do not have to use them
in their discourses but they help them to create cohesion and
coherence in their writing and their speaking (Brown & Yule,
1983). Furthermore, Zarei (2013: 108-109) also exemplifies
and explains some characteristics regarding discourse markers
as follows: connectivity, optionality, non-truth conditionality,
weak clause association, literality, morality, multi-categoriality.
Classification of discourse markers
Brown and Yule (1983: 191) summarize the taxonomy of types
of discourse markers provided by Halliday and Hasan (1976) as
a) additive: and, or, furthermore, similarly, in addition
b) adversative: but, yet, however, on the other hand, nevertheless
c) causal: so, consequently, for this reason, it follows from this
d) temporal: then, after that, an hour later, finally, at last
Halliday and Hasan (1976: 238-239) exemplify each category
with a sentence in order to clarify the concept of taxonomy:
(1) For the whole day he climbed up the steep mountainside,
almost without stopping.
a. And in all this time he met no one. (additive)
b. Yet he was hardly aware of being tired. (adversative)
c. So by night time the valley was far below him. (causal)
d. Then, as dusk fell, he sat down to rest. (temporal)
Halliday and Hasan (1976: 239) add that the words (and, yet,
so, then) in the examples above clearly show the general conjunctive
relations ensuring to approach a text easily so as to
understand and analyze the cohesion.
As Halliday and Hasan (1976: 238) mention, there is no single,
uniquely correct inventory of the types of conjunctive relation;
different classifications are possible, each of which would highlight
different aspects of the facts. Hence, two more classifications
of discourse markers suggested by Quirk et al. (1985) and
Fraser (1999) will be outlined.
The other taxonomy of discourse connectives proposed by
Quirk et al. (1985:634-640) includes seven conjunctive roles
some of which have subdivisions as presented in Figure 1
Click Here to Zoom
|Figure 1: The classification of discourse markers by Quirk et al.
Quirk et al. (1985) classify the conjuncts according to their
functions and list them as listing (first, second, firstly, secondly,
in the first place, in the second place, first of all, on the one
hand, to conclude, finally, last of all, correspondingly, equally,
likewise, by the same token), summative (altogether, further, also, furthermore, moreover, in addition, above all, on the top
of it all), appositive (namely, thus, in other words, for example,
that is), resultative (accordingly, hence, so, therefore, as a consequence,
as a result of, of course), inferential (else, otherwise,
then, in that case), contrastive (better, rather, more precisely,
again, on the other hand, worse, instead, on the contrary, by
contrast, anyhow, anyway, however, nevertheless, still, yet, in
spite of, that said) and transitional (incidentally, by the way, by
the by, meantime, eventually).
The taxonomy of Fraser differentiates from Quirks et al. in
terms of the number and the name of categories. Fraser (1999)
classifies the discourse markers as contrastive markers (but,
however, al(though), in contrast (with/to this/that), whereas,
in comparison ( with/to this/that), on the contrary, contrary to
this/that, conversely, instead (of doing) this/that, rather (than
(doing) this/ that),on the other hand, despite (doing) this/that,
in spite of (doing) this/ that, nevertheless, nonetheless, still)
elaborative markers (and, above all, also, besides, better yet,
for another thing, furthermore, in addition, moreover, more
to the point, on top of it all, too, to cap it all off, what is more,
I mean, in particular, namely, parenthetically, that is (to say),
analogously, by the same token, correspondingly, equally,
likewise, similarly, be that as it may, or, otherwise, that said,
well) and inferential markers (so, of course, accordingly, as a
consequence, as a logical conclusion, as a result, because of
this/that, consequently, for this/that reason, hence, it can be
concluded that, therefore, thus, in this/that case, under these/
those conditions, then, all things considered). He also reveals
some additional subclasses including after all, because, for
this/that reason, since and they are called reason (causative)
markers. The other subclass consists of topic-relating markers
such as incidentally, to return to my point, with regards to
(Fraser, 1999: 947-949).
Here is an example for each subcategory proposed by Fraser
(1999, p. 947-949).
(1) We left late. Nevertheless, we got there on time. (contrastive
(2) The picnic is ruined. The mayonnaise has turned rancid.
The beer is warm. Furthermore, its raining. (Elaborative
(3) The bank has been closed all day. Thus, we couldnt make
a withdrawal. (Inferential marker)
(4) Take a bath right away, because we have to get going. (Reason/
(5) I am glad that is finished. To return to my point, Id like to
discuss your paper. (Topic change marker)
The contrastive marker (nevertheless) in the example (1)
illustrates that two sentences contradict one another in that
they present contrasting views. In the example (2) above, the
elaborative marker (furthermore) provides a quasi-parallel
relationship between the second sentence and the first sentence,
which adds further meaning to the discourse. In the
inferentials marker category it could be deduced that the discourse
marker (thus) in the example (3) presents a conclusion
for the first sentence.
In the other subclass (reason/causative) marker discourse
marker (because) in the example (4) emphasizes a reason for
the first sentence. On the other hand, in the last example (5)
the discourse marker (to return to my point) attempts to manage
the discourse using a topic-relating discourse marker.
Corpus and Corpus-based approach
The term corpus comes from Latin, which means body. It
can be a body of any kind of written or spoken text. Özhan and
Zeyrek (2012: 16) point out the texts are compiled either as
written texts or as a transcription of recorded speech. However,
a text could also be regarded as a corpus provided that
it has such features as sampling and representativeness, finite
size, machine-readable form and a standard reference (McEnery
& Wilson, 2001).
The corpus-based approach helps the researchers to identify
and classify language items by using high-powered computers,
robust software, and large electronic collections of written or
spoken texts obtained from the real world (Gardner, 2007).
Grant (2010: 2282) supports that corpus studies enable both
linguists and language teachers to investigate aspects of written
and/or spoken English by analyzing the authentic language
collected in a variety of both small specialized and large general
databases. Although the extent of this study is small in
number, using a corpus-based approach enabled the study to
be conducted in an easier and more reliable manner since it is
difficult to count all the discourse markers manually.
Data Collection Procedure
The 104 students who participated in this study were studying
English in prep-class at Namık Kemal University. They studied
different majors. They wrote a descriptive composition as a
part of their mid-term exam in the fall semester of the 2013-
2014 academic year by choosing one of the following topics:
1. You study English and you want to practice English with
online friends. Introduce yourself and your family.
2. Bob Simpson has a busy life. Look at the pictures and write
about his typical day.
Only 13 students chose the second topic and the remaining
91chose chose the first one since their proficiency level was
elementary and it was easy to write about themselves rather
than the third persons life using Simple Present Tense. They
were requested to write a paragraph with 80-100 words as a
part of the exam by which a small-size corpus was compiled. It
consisted of 8.500 tokens.
The obtained data were typed and the corpus was constructed
by the researcher herself. Indeed, it was a small size corpus
when compared to other corpora such as the British National
Corpus, Corpus of Contemporary American English and Michigan
Corpus of Academic Spoken English. However, this corpus
is unique in a sense that there is no similar example investigating
The gathered data were analyzed in order to find out answers
of the following research questions:
1. Which discourse markers are used by elementary-level students
in their writings?
2. What are the frequency and function of discourse markers
in students writings?
The corpus was conducted both for quantitative and qualitative
analysis. Quantitative analysis was investigated in order to
determine the frequency and percentage of discourse markers.
The concordance program AntConc 3.2.4 developed by Laurence
Anthony was used for this corpus analysis. It was chosen
since it was free and it was available on the website www.antlab.
sci.waseda.ac.jp/ antconc_ index.html. Though there are
the two other common concordance programs of WordSmith
Tools and Monoconc Pro, they were not preferred as they are
Qualitative analysis was carried out so as to identify the functions
of discourse markers. Thus, Frasers (1999) taxonomy for
discourse markers was adopted although there are two other
taxonomies of discourse markers suggested by Halliday and
Hasan (1976) and Quirk et al. (1985). While Frasers taxonomy
embodies the three main subclasses of contrastive, elaborative,
and inferential markers, there are also another two subclasses
of reason (causative) markers and topic-relating markers.
While the analyses were examined for the type, frequency
and functions of discourse markers, the inter-rater reliability
was calculated in order to increase the reliability of results and
minimize the subjectivity (Jalilifar, 2008). The first rater was the
researcher herself and the other rater was her colleague. After
debating on some disagreements, a consensus was achieved.
When the discourse markers were analyzed manually, the use of and, but, or and so in an elliptical sentence such as
Jack and Mary rode horses was not considered as a discourse
marker since a discourse marker should present a different
message in the related sentence whereas there is merely one
message in the sentence above (Fraser, 1999: 939). Furthermore,
semantically inappropriate usage was not counted as a
discourse marker. For example:
like pop music and I dont like jazz music.
He has breakfast so he reads newspaper.
He likes eating healthy food for diner. Then, he watches tv.
The words above (too, and, so, then) used by the students in
their writings in order to connect the sentences and ideas were
not credited as discourse markers since they failed to establish
content integrity and cohesion between two sentences. Finally,
because this paper particularly focused on discourse markers,
grammatical mistakes were not corrected (Feng, 2010:302).
Table 1 below was organized by taking the criteria above into
consideration. It demonstrates the types of discourse markers,
as well as the frequency and percentages, used by elementarylevel
students in their writings that make up the corpus investigated
in this study.
The results from Table 1 reveal that eight types of discourse
markers are adopted by elementary-level students as follows:
and, but, because, then, so, also, too and still. Such types of
discourse markers as and, but and because are used more
frequently than the others (then, so, also, too and still). Furthermore,
while and as a discourse marker accounts for 54.4%
of the overall discourse markers, but covers 28.4% and because
covers 10%, then makes up 3.9%, so and also constitute 1.1%
and lastly too and still, as the least frequently used discourse
markers, cover 0.55%. Here is an example for each discourse
marker employed in the corpus.
I write the poem in my free time. Also I always read the
My brother name is Ahmet. He lives in İstanbul with my
parents. My sisters name is Elif. She lives in İstanbul with my parents, too.
He has a shower. He wears clothes. Then he has a breakfast.
I am learning English because I need it for my job.
I am a student. And I am studying computer engineering.
I like playing football. But I sometimes have time to play
I like swimming so I prefer go to beach for holiday.
I am a university student. I lived in Adana with my family
but I am living in Tekirdağ now. My family still live in Adana.
The examples above included discourse markers employed
properly by the students and it reveals that even low-proficiency
level students tend to use cohesive markers to achieve
content integrity and cohesion in their text.
After the types of discourse markers, their frequency and
percentages were identified by using the software program
AntConc 3.2.4 and the functions of discourse markers based
on Frasers taxonomy of discourse markers were analyzed.
Table 2 above illustrates that elaborative markers (56.2%) were
used more frequently than the other categories of discourse
markers. Jalilifar (2008) conducted a study with Iranian students
where they were asked to write descriptive compositions.
These compositions were analyzed using Frasers (1999)
taxonomy of Discourse Markers. His results conform to this
current study in which elaborative markers are more frequent
markers and he argues that the extensive use of elaborative
markers may be explained because descriptive writing in general
requires elaboration of ideas which depends on the use
of elaborative markers (p.116). Consequently, it can be stated
that students writings in this study which consist of the corpus
were descriptive paragraphs and elaborative markers may be
used more commonly in this study.
Contrastive markers embody the second frequent usage
(28.8%). This study shows that while Turkish students tend to
use contrastive markers in their writings, which is even valid for
elementary level students, Altenberg and Tapper (1998) found
by using the taxonomy of Quirk et al. (1985) that contrastive
markers were underused by Swedish learners.
Causative markers account for 10% of all discourse markers.
Since the texts which made up the corpus were descriptive,
the use of causative markers was less. However, the study conducted
by Heidar and Biria (2011) in order to analyze the discourse
markers in International Law texts shows that causative markers such as because and since are used more frequently
than the other discourse markers in the corpus.
The other subcategory of taxonomy of discourse markers is
inferential markers and 5% of the overall discourse markers
need to be explained within the framework of this category.
Finally, it should be noted that any topic-relating discourse
markers were not used by elementary-level students since the
proficiency level of students was low. However, according to
the study of Lahuerta Martinez (2004), high proficiency level
students did not employ topic-relating discourse markers frequently
in their writings either.
This current study concludes that even elementary level learners
tend to use discourse markers to achieve cohesion in their
writings as they used 180 discourse markers out of 8500 words
properly. This result is not discouraging when the students
proficiency level is considered. The types of discourse markers
used by the students are and, but, because, then, so, also,
too, and still. The most frequently used discourse marker is and
which occurred 98 times. Because, then, so, also, too, and still
were employed 51, 18, 7, 2, 2, 1, 1 respectively in the corpus.
This result conforms to the general idea that EFL learners have a
tendency to overuse discourse markers in their writings. In this
study, the native language of learners might have influenced
the use of discourse markers since discourse markers such as
and (ve), but (ama) and because (çünkü) are used frequently in
The research of Prommas and Sinwongsuwat (2011) in order
to find out the discourse markers in argumentative compositions
used by Thai undergraduate students and English native
speakers revealed that both groups of students had similar
characteristics as they adopted the same discourse markers of
and, but, because, also.
However, this result is not similar to Altenberg and Tappers
(1998) findings. They carried out a study to compare the adverbial
connectors used in written English by advanced Swedish
learners of English, advanced French learners of English and
native speakers of English. They examined the different corpora
using the taxonomy of Quirk et al. (1985). They reveal that
the markers in their written English are underused by Swedish
Consequently, the discourse markers were also analyzed
within the framework of Frasers (1999) taxonomy. The results
show that elaborative markers embody the largest ratio of the
overall discourse markers (56.2%), followed by contrastive
markers (28.8%), reason (causative) markers (10%) and inferential
markers (5%). This result bears similar findings to the
study of Lahuerta Martinez (2004) who investigated the uses
of discourse markers in the expository compositions of Spanish
As a pedagogical conclusion, it can be suggested that learners
should be taught different discourse markers and asked to
write in different genres on appropriate topics for their level.
Consequently, they will not be confined to the same markers
and will learn the differences among the genres.
IMPLICATIONS FOR FURTHER RESEARCH
The corpus in this study was confined to 104 paragraphs and
8500 tokens, which makes it a small-size corpus. Therefore, the
results cannot be generalized to all elementary level English
language learners since the corpus is small and the number
of students is limited. For further research, students from different
universities and different language institutions could
participate in the study. Furthermore, the lack of comparison
is another drawback of this study. For further research, a longitudinal
research could be designed in order to find out the
changes and differences in the uses of discourse markers of the
students. Finally, further studies could investigate the effects of
the native language (Turkish) on the use of discourse markers
in written discourse by the learners of English.
1) Altenberg, B. & Tapper, M. (1998). The use of adverbial connectors
in advanced Swedish learners written English. In S. Granger
(Ed.), Learner English on computer. (pp. 80-93). New York:
Bell, D. M. (2010). Cancellative discourse markers: A core/
periphery approach. Pragmatics, 8(4), 515-54.
3) Brinton, L. J. (1996). Pragmatic markers in English: grammaticalization
and discourse functions. Berlin: Walter de Gruyter
4) Brown, G. & Yule, G. (1983). Discourse analysis. New York:
Cambridge University Press.
5) Feng, L. (2010). Discourse markers in English writing. The Journal
of International Social Research, 3(11), 299-305.
6) Fraser, B. (1999). What are discourse markers? Journal of
Pragmatics, 31, 931-952.
7) Gardner, D. (2007). Validating the construct of word in applied
corpus-based vocabulary research: a critical survey. Applied
Linguistics, 28(2), 241-265.
8) Grant, L. E. (2010). A corpus comparison of the use of I dont know
by British and New Zealand speakers. Journal of Pragmatics,
9) Halliday, M.A.K., & Hasan, R. (1976).Cohesion in English. London:
10) Heidar, D.M. & Biria, R. (2011). Sociopragmatic functions of
discourse markers in international law texts. Theory and
Practice in Language Studies, 1(11), 1479-1487.
11) Jalilifar, A. (2008). Discourse markers in composition writings:
The case of Iranian learners of English as a foreign language.
English Language Teaching, 1(2), 114-122.
12) Lahuerta Martinez, A.C. (2004). Discourse markers in the
expository writing of Spanish university students. Iberica, 8,
13) McEnery, T. & Wilson, A. (2001). Corpus linguistics: an introduction.
Edinburg: Edinburg University Press.
14) Norrish, J. (1983). Language learners and their errors. London: The
15) Özhan, D.& Zeyrek, D. (2012). A comparative analysis on the
use of but, however and although in the university students
argumentative essays: A corpus-based study on Turkish
learners of English and American native speakers. (Unpublished
Doctoral Dissertation). METU, Ankara.
16) Polat, B. (2011). Investigating acqusition of discourse markers
through a developmental learner corpus. Journal of
Pragmatics, 43, 3745-3756.
17) Prommas, P. & Sinwongsuwat, K. (2011). A comparative study
of discourse connectors used in argumentative composition
produced by Thai EFL learners and English-native speakers. In
Proceedings, The 3rd International Conference On Humanities
and Social Sciences, Prince of Songkla University.
18) Quirk, R., Greenbaum, S., Leech, G. & Svartvik, J. (1985). A
comprehensive grammar of the English language. London:
19) Rahimi, M. (2011). Discourse markers in argumentative and
expository writing of Iranian EFL learners. World Journal of
English Language, 1(2), 68-78.
20) Sadeghi, B. & Kargar, A.( 2014). The effect of explicit instruction of
discourse markers on EFL learners writing ability. International
Journal of Educational Investigations, 1(1), 328-338.
21) Urgelles-Coll, M. (2010). The syntax and semantic of discourse
markers. London: Continuum International Publishing.
22) Wang, Y. & Guo, M. (2014). A short analysis of discourse coherence.
Journal of Language Teaching and Research, 5(2), 460-465.
23) Witte, S. P. & Faigley, L. (1981). Coherence, cohesion, and writing
quality. College Composition and Communication, 32(2), 189-
24) Zarei, F. (2013). Discourse markers in English. International
Research Journal of Applied and Basic Science, 4(1), 107-117.