Content uploaded by Mohammad Ahmadi
Author content
All content in this area was uploaded by Mohammad Ahmadi on Dec 04, 2024
Content may be subject to copyright.
Applied Research on English Language
V. 10 N. 4 2021
pp: 139-166
http://jare.ui.ac.ir
DOI: 10.22108/ARE.2021.130833.1787
Document Type: Research Paper
___________________________________________
* Corresponding Author.
Authors’ Email Address:
1 Rajab Esfandiari (esfandiari@hum.ikiu.ac.ir), 2 Mohammad Ahmadi (ahmadi.m8362@gmail.com), Edward Schaefer
(schaefered@dokkyo.ac.jp)
ISSN (Online): 2322-5343, ISSN (Print): 2252-0198 © 2021 University of Isfahan. All rights reserved
A Corpus-based Study on the Use and Syntactic Functions of Lexical
Bundles in Applied Linguistics Research Articles in Two Contexts of
Publications
Rajab Esfandiari 1*, Mohammad Ahmadi 2, Edward Schaefer 3
1 Associate Professor of applied linguistics, Department of English Language, Faculty of
Humanities, Imam Khomeini International University, Qazvin, Iran
2 Visiting Professor of TEFL, Department of English Language, Faculty of Humanities, Imam
Khomeini International University, Qazvin, Iran
3 Professor Emeritus, Graduate School of Humanities and Science, Ochanomizu University,
Tokyo, Japan and Adjunct Instructor, Dokkyo University, Saitama, Japan
Received: 2021/10/02 Accepted: 2021/10/27
Abstract: The present study investigated the use of lexical bundles (LBs) in research articles
authored by English L1 and Persian L1 academic writers, with a special focus on the syntactic roles
of LBs in a larger context of sentence level. Four-word bundles were retrieved and classified
structurally. The use of identified LBs was compared in two writer groups. The syntactic roles and
relative complexity of the bundles’ structures were analyzed in relation to Biber, Gray, and Poonpon’s
(2011) hypothesized stages of writing development. The results indicated different patterns of reliance
on LBs, with Persian writers making greater use of LBs at a higher frequency. In addition, Persian
academic writers tended to use high-frequency bundles differently from native-speaker academic
writers. The results of the syntactic analysis of LBs reflected more frequent use of LBs functioning
as compressing lexico-grammatical structures in a native English-speaker corpus, which is indicative
of a more complex academic register compared to that of a Persian L1 corpus. The pedagogical
implications of the findings for the explicit instruction of syntactically complex corpus-driven LBs
for discipline-specific genre writing and suggestions for future research are discussed.
Keywords: Lexical bundles, Syntactic Complexity, Research Articles.
140 Applied Research on English Language, V. 10 N. 4 2021
AREL
Introduction
The study of multiword sequences (MWS) has drawn the attention of researchers over the past few
years. This interest has its roots in the pervasiveness of MWSs and psycholinguistic explanations
which suggest a processing advantage for MWSs compared with the sequences of words that are
processed individually (Conklin & Schmitt, 2008). This processing advantage is attributed to the
“holistic nature of formula” in both L1 and L2 (Jiang & Nekrasova, 2007, p. 433). The
psycholinguistic validity of MWSs has been strengthened in different studies (e.g., Ellis &
Simpson-Vlach, 2009), where formulas have been found to have a processing advantage as well as
clearly defined functions, particularly in English for academic purposes (EAP).
The function of MWSs has been specifically investigated in EAP. The bulk of the studies has
documented that academic writing relies, to a great extent, on formulaic sequences (e.g., Ruan,
2017; Wei & Lei, 2011). This line of research mainly used MWSs as a linguistic means to analyze
different text types produced by native/nonnative or expert/novice academic writers. While the
findings of these studies broaden our knowledge of the construction of MWSs by different writer
groups, they are, by no means, conclusive, as many of them have confounded ‘register/discipline’,
L1, genre, audience, and topic “with the difference between groups of writers (e.g., comparing
general essays written by students to research articles written by professionals)” (Pan, Reppen, &
Biber, 2016, p. 62).
A particular type of formulaic sequence is LBs, which are defined as the combination of
words that recur most commonly in a given register (Biber, Johansson, Leech, Conrad, & Finegan,
1999). They are of special importance in academic writing as they fulfill important discourse
functions and are a hallmark of advanced academic writing (Pan & Liu, 2019). Previous studies
mainly drew on a structural and functional framework of lexical bundles following Biber et al.
(1999), Biber, Conrad, and Cortes (2004), and Hyland (2008). However, the syntactic function of
lexical bundles within the unit of sentence length has received little attention in previous literature.
This is particularly important because lexical units do not stand alone; rather, they are parts of larger
units embedded within a sentence. As Shin (2018) pointed out, previous studies largely analyzed
LBs within phrases and clauses; however, these units might not always be appropriate because “a
bundle’s last word is often the first word of another structure” (p. 116). Shin further calls for the
extension of the scope of the structural unit of LBs to the sentence level in order for researchers to
be able to examine different syntactic roles of bundles within a sentence, as the same LBs which
have been determined on the basis of frequency can occur in different syntactic units which
function differently.
A Corpus-based Study on the Use and Syntactic Functions of Lexical Bundles in Applied Linguistics Research Articles in Two Contexts of Publications 141
AREL
There has been surprisingly little research investigating the syntactic functions of LBs in
academic writing. One of the few existing relevant studies was conducted by Shin (2018), who
explored LBs situated in the texts produced by native and nonnative-speaker freshman university
students. However, the present study is different from that of Shin. Although both studies
investigated LBs in the academic genre, the present study employed published journal articles to
construct the corpus while the study by Shin made use of argumentative essays written by
university freshmen. A research article (RA) is a completely different sub-genre from those
produced by student writers, “with a different purpose, audiences, and repertoire of rhetorical
features” (Hyland, 2008 p. 57). RAs are the most important sub-register of professional academic
writing (Biber & Gray, 2010).
Conventional analysis of LBs within phrasal or clausal units will result in a list of fragmented
bundles which provide very little information with regard to their syntactic properties. Bundles do
not stand alone; rather, they are incorporated into larger structures, so understanding the ways in
which they are used to form larger units can help learners produce texts that read more target-like
(see Garner, Crossley, & Kyle, 2019). Accordingly, the results obtained from the present study may
offer more insights into the way syntactic roles of LBs contribute to the construction of expert
academic registers in native and nonnative contexts. Therefore, the present study aimed at filling
the gap in the literature by extending the structural unit of LBs to sentence level so that their
syntactic properties will be appropriately analyzed.
In addition, previous studies have been inconclusive with regard to native versus non-native
speaker contrast of LBs in academic writing with some studies showing native speakers’ heavier
reliance on bundles for constructing the texts (e.g., Atai & Tabandeh, 2015) while the others
showing the opposite (e.g., Esfandiari & Barbary, 2017; Rahimi Azad & Modarres Khiabani,
2018). As a consequence, more studies are required to investigate the role of native speaker status
in the frequency distribution, overuse, and underuse of formulaic language in advanced academic
writing, as the results could build up a clearer picture of academic formulaicity in the important
sub-register of RAs. Moreover, previous studies did not provide clear evidence as to whether
different distributional patterns of LBs will result in a more/less complex discourse style in relation
to existing taxonomies of academic writing development. Accordingly, the purpose of the present
study is to provide more understanding of the way native and nonnative academic writers employ
LBs in applied linguistics RAs with a special focus on the syntactic roles of the structures in which
the bundles occur.
142 Applied Research on English Language, V. 10 N. 4 2021
AREL
Literature Review
LBs are understood to be semantically transparent combinations of words that are identified as
“simply the most frequently recurring sequences of words” (Biber & Barbieri, 2007, p. 264).
Due to their pervasive nature, a frequency threshold has been chosen for the identification of
LBs, which has the great advantage of being methodologically straightforward and having face
validity (Ellis, 2012). Previous studies normally used the frequency threshold of 10 times per
million words (e.g., Ellis &Simpson-Vlach, 2009), 20 times per million words (e.g., Csomay,
2013), 25 times per million words (Chen & Baker, 2010), or 40 times per million words (Biber
& Barbieri, 2007). In order to get round the problem of idiosyncrasies from individual writers,
the criterion of dispersion is also used, which determines the number of texts in which a
linguistic feature occurs (Gries & Ellis, 2015). This is to ensure that the identified bundles are
typical of the entire corpus (Pan et al., 2016). Frequency distribution of LBs provides evidence
for the description of register variation such that frequent language features that typify a
particular register are prioritized (Grabowski, 2015).
An important register for the investigation of variations in LBs is academic writing. LBs
are important building blocks of coherent discourse in academic writing because they serve as
an effective discriminator of the register which employs distinct sets of LBs that are tailored to
its communicative purposes (Wang & Zhang, 2021). Hyland (2008) holds that the investigation
of LBs is of particular importance in EAP, as there is mounting evidence that LBs have
important functions in academic writing (Staples, Egbert, Biber, & McClair, 2013). Similarly,
Cortes (2004) argues that the appropriate use of formulaic expressions is the marker of
proficiency in a register, including academic writing.
Recently, there has been a growing number of studies exploring fixed expressions within
academic writing by L2 writers, compared with native-English speaking writers (e.g., Adel &
Erman, 2012; Pan et al., 2016; Salazar, 2014; Esfandiari & Barbary, 2017). For example, Chen
and Baker (2010) investigated LBs in L1 and L2 academic writing. Two corpora of published
academic writing and student writing were used to be explored in terms of types and tokens of
LBs both qualitatively and quantitatively. The results indicated that published academic texts
used the widest range of LBs, whereas L2 Chinese student writing exhibited the smallest range.
Another finding of their study was that L2 students overused certain LBs which native-speaker
academics rarely used. Similarly, Adel and Erman (2012) compared the use of LBs by L1
speakers of Swedish advanced learners and their English native-speaker counterparts who were
A Corpus-based Study on the Use and Syntactic Functions of Lexical Bundles in Applied Linguistics Research Articles in Two Contexts of Publications 143
AREL
all undergraduate students in the discipline of applied linguistics. Four-word lexical bundles
were extracted from the corpora, and they were analyzed both quantitatively and qualitatively
in terms of the functions they served. The results of their study showed that native speakers
used more varied and a larger number of lexical bundles in comparison to L2 writers. Their
findings supported previous native/non-native research traditions focusing on MWSs in
general and LBs in particular. Recently, Lu and Deng (2019) explored the use of lexical bundles
in dissertation abstracts by Chinese and L1 English doctoral students. Four-word bundles were
extracted from 13,596 and 4,755 abstracts of doctoral dissertations. The identified bundles
were categorized according to their functional and structural attributes. The results of his study
revealed that Chinese students used lexical bundles in a fundamentally different way with
regard to functional and structural features of LBs. They also exhibited incomplete knowledge
of LBs, indicating L1 transfer. The other finding of their study was that LBs that were used by
Chinese learners did not meet the conventions of academic writing in hard sciences.
While the results from previous studies on LBs produced by native versus non-native
language speakers are valuable in revealing the role of nativeness in academic writing
proficiency (See Romer & Arbor, 2009), what is less clear is the effect of methodological
issues, such as comparability of corpora and frequency/distribution thresholds, on the extracted
bundles from the corpora to be compared. In a study on methodological issues in contrastive
lexical bundle research, Pan, Reppen, and Biber (2020) revealed that “the difference in the
number of words and number of texts across sub-corpora can have a strong effect on claimed
differences in bundles across groups even when the corpora are closely matched for their
register and topic” (p. 215). Pan et al. (2020) conducted a similar study on the effect of
identification threshold on lexical bundle research, and it was found that “different
identification thresholds applied to the same pair of corpora may yield conflicting results” (p.
336). Accordingly, it is suggested that researchers base their bundle analysis on structural and
functional characteristics, rather than comparing lists of specific bundles (Pan et al., 2016).
In order to arrive at a clearer picture of the pattern of LBs associated with certain groups,
and to get round the problem of long lists of produced LBs by native/non-native groups, which
were of little pedagogical value, some scholars have categorized LBs through structural and
functional taxonomy. Two commonly cited classifications are those of Biber et al. (1999) and
Hyland (2008). The former classifies LBs based on their structural attributes, which include
verb phrase (VP) bundles, noun phrase (NP) bundles, and prepositional (PP) bundles. The
latter, however, takes a functional perspective on LBs, which fall into three categories:
144 Applied Research on English Language, V. 10 N. 4 2021
AREL
research-oriented bundles, text-oriented bundles, and participant-oriented bundles.
Although structural and functional classifications of LBs act “as alternative formulas
[which] emerged as a matter of inquiry in the language teaching field” (Güngör & Uysal, 2016,
p.177), identified LBs do not reflect the developmental path to use discourse conventions
appropriately (Shin, 2018). The same bundles may occur in different syntactic positions for
which structural and functional classifications do not capture the complexity of the language
unit within which the LBs occur. For example, the bundle one of the most can be used in
different syntactic roles such as subject (e.g., One of the most notable findings of the present
research is…), subject predicative (e.g., …balance of power as being one of the most crucial
elements…), or direct object (e.g., The software identified one of the most…).
In a series of studies, Biber and Gray (2010, 2013, 2016), and Biber et al. (2011) have
documented that academic prose is structurally more compact than conversation. This
argument ran counter to previous assumptions that academic writing is maximally explicit in
meaning. These researchers have shown that a compressed discourse style in academic writing
is at odds with explicitness, arguing that traditional clausal measures of syntactic complexity
cannot gauge the grammatical complexity of academic texts because of their poor theoretical
foundations. In order to characterize the development in academic writing, Biber et al. (2011)
hypothesized the developmental sequences of grammatical complexity along two grammatical
parameters: grammatical form and syntactic function. Accordingly, three grammatical types
were identified: finite dependent clauses, non-finite dependent clauses, and dependent phrases.
These grammatical stages progress from finite dependent clauses through intermediate stages
of non-finite dependent clauses and finally to the last stages of dependent phrases (Biber et al.,
2011). Although the hypothesized stages of writing development did not specifically
investigate lexical bundles, they “paved the way for the exploratory use of this approach in the
production of other linguistic features such as lexical bundles” (Shin, 2018, pp. 119-120).
Different studies have tried to provide empirical evidence to support the hypothesized
stages of writing development proposed by Biber et al. (2011). For instance, Parkinson and
Musgrave (2014) explored the syntactic complexity of academic texts produced by MA and
undergraduate students. With a special focus on noun phrase modifiers, the authors confirmed
the developmental stages of writing complexity in the sense that undergraduate writers relied
heavily on premodifiers, which are supposed to be acquired at earlier stages of writing
development. On the other hand, noun modifiers employed by MA writers better approximated
those of published academic prose. Similarly, Lan and Sun (2019) examined the quality of
A Corpus-based Study on the Use and Syntactic Functions of Lexical Bundles in Applied Linguistics Research Articles in Two Contexts of Publications 145
AREL
student papers across three tiers of first-year L2 students. The results revealed that low-rated
papers demonstrated lower complex nominal densities, lower mean length of clauses, and
lower mean length of T-units, providing further evidence that development in academic writing
moves from clausal embedding to phrasal embedding.
The current study intends to extend the structural analysis of LBs in the existing literature
by analyzing the identified bundles within the framework of Biber et al.’s (2011) hypothesized
stages of academic writing. To this end, we identified and examined LBs in two corpora of the
RAs authored by L1-Persian and L1-English academic writers. Specifically, the study is guided
by the following two research questions:
1. What are the patterns of use of lexical bundles in the writing of L1-Persian and L1-
English academic writers?
2. How do L1-Persian writers and L1-English writers in applied linguistics use lexical
bundles in academic writing in terms of syntactic functions?
Methodology
Corpus Construction
The present study drew on native and nonnative corpora of RAs in applied linguistics from
leading journals in the field. We chose applied linguistics based on the following
considerations: First, “it is an interdisciplinary field of study which represents a wide landscape
of academic territories” (Shirazizadeh & Amirfazlian, 2021, p. 2). Second, the study of LBs in
applied linguistics has become an increasingly important area in recent years (Wang & Zhang,
2021). Accordingly, the present study intended to extend the existing literature on the use of
LBs in applied linguistics by approaching the issue from a different perspective.
The native corpus (NC) was composed of 103 texts extracted from published RAs in
national English-medium journals in Iran. The nonnative corpus (NNC) was comprised of 106
texts from highly prestigious international English-medium journals. Descriptive statistics of
the corpora are presented in Table 1.
146 Applied Research on English Language, V. 10 N. 4 2021
AREL
Table 1. Description of the Corpora
Corpora
Number of Texts
Mean Length of Texts (Words)
Total Corpus Size (Words)
NC
103
9929.04
1,022,692
NNC
106
9660.80
1,024,999
The inclusion of the journals in this study was based on the two criteria of publication
history and h index, which is defined as the number of publications of a certain author (h) with
a citation number of at least h times (Hirsch, 2005). In other words, a researcher who has
published 15 research papers, each with at least 15 citations, would have an h index of 15. The
advantage of the h index over the traditional journal impact factor (JIF) is that it is less affected
by over-citation because it is not based on mean scores (Harzing & Van der Wal, 2008).
Journals with a higher h-index (more citations in more articles) represent a model of empirical
research articles in the field of applied linguistics and language education because they impact
the field through a high number of highly cited articles. Table 2 presents descriptive
information of the journals from which the articles have been extracted.
Table 2. Overview of Journals Included in Native and Nonnative Corpora
Journal
Years of Publication
H factor
Language Learning
1948-1953, 1955-1956, 1958-ongoing
38
Applied Linguistics
1980-ongoing
38
TESOL Quarterly
1981-ongoing
36
Modern Language Journal
1916-1996, 1998-2001, 2005-ongoing
36
English for Specific Purposes
1980-1981, 1986-ongoing
25
Iranian Journal of Applied Language
Studies
2009- ongoing
—
Journal of Teaching Language Skills
2009- ongoing
—
Journal of English Language Teaching
and Learning
2010- ongoing
—
Journal of Language and Translation
2010- ongoing
—
Journal of Research in Applied
Linguistics
2010- ongoing
—
Issues in Language Teaching
2012-ongoing
—
Applied Research on English
Language
2012-ongoing
—
Iranian Journal of Language Teaching
Research
2013-ongoing
—
Iranian Journal of English for
Academic Purposes
2015-ongoing
—
A Corpus-based Study on the Use and Syntactic Functions of Lexical Bundles in Applied Linguistics Research Articles in Two Contexts of Publications 147
AREL
In order to identify native and nonnative English academic writers, we followed the
identification method suggested by Wood (2001), who took into account the names and
affiliations of authors. To determine the L1 status of the authors in NNC, we simply deduced
the names and affiliations were indicative of Persian writers. As for native English writers in
NC, after checking the Anglophone origin of the names, we made sure if the authors were
affiliated with any institution in Engish-L1 speaking countries. Texts authored by multiple
authors were excluded from the study if the authors had differing native and nonnative English
status.
All research articles followed the IMRD format and were published between 2018 and
2020. The collection of recently published research articles characterizes ‘the present day’
trends in academic writing (Biber & Gray, 2016). Special issues were excluded, as special
issues varied both in article type (in having synthesis or review articles) and in communicative
functions. Only research studies representing empirical studies were included so that rhetorical
and linguistic variations could be controlled for. “Non-empirical and theoretical review articles
often have varied rhetorical organization, which may result in writers’ divergence in making
linguistic choices” (Ruan, 2018, p. 6). Accordingly, articles were excluded if their functions
and organizational structures differed from those of empirical research articles, which included
meta-analyses, position papers, forum discussions, and book reviews. All tables, appendices,
diagrams, graphs, titles, captions, and footnotes were removed from the papers so as to ensure
the reliability of the data.
Identification of Lexical Bundles
In order to identify LBs, the authors needed to decide on the length of word sequences as the
first step in the analysis. It was an important decision because different identification thresholds
may result in different lists of bundles (Pan et al., 2016). Biber et al. (1999) argued that three-
word bundles are extremely common, while “four-word, five-word, and six-word bundles are
more phrasal in nature and correspondingly less common” (p. 992). Given that the retrieved
bundles in this study have been manually checked through concordance lines for determining
the syntactic functions of each bundle, the frequency threshold of three-word bundles would
generate a long list of word sequences whose analysis would be very labor-intensive. On the
other hand, four-word bundles “are far more common than 5-word strings and offer a clearer
range of structures and functions than 3-word bundles” (Hyland, 2008, p. 8). As a result, we
investigated four-word bundles in this study. Frequency and dispersion are two main criteria
148 Applied Research on English Language, V. 10 N. 4 2021
AREL
for the selection of LBs in literature. However, there seems to be little consensus among
researchers regarding the determination of the cut-off point. In this study, we followed Cortes
(2008) and set the frequency criterion of 20 times per million words across at least five or more
texts.
Data Analysis
The bundles were identified using a concordance tool called AntConc version 3.5.9 (Anthony,
2020). Discipline-specific bundles (those which are more frequently found in a given discipline
e.g., students of other languages) and overlapping bundles (those that are part of larger
bundles) were excluded so as not to inflate the number of bundles (See Chen & Baker, 2010).
Following Biber and Barbieri (2007), we normalized identified bundles to 1,000,000 words.
This practice has at least two advantages: first, it allows for the comparability of the results
obtained from the current study to those of others (Biber & Barbieri, 2007), and second, it
allows for employing parametric tests which could otherwise be wasteful of data (Biber et al.,
2011). In order to check for the significance of the differences with regard to the frequency
distribution of the LBs between the two corpora, log-likelihood tests were performed. The next
step for the researchers was to categorize the retrieved bundles based on Biber et al.’s structural
taxonomy, which involved identifying the type of internal structural unit (verb phrase bundles,
noun phrase bundles, and prepositional bundles). Drawing on Biber et al.’s (2011)
hypothesized stages of writing development, and syntactic classification of phrasal bundles
(Cortes, 2015; Shin, 2018), we subsequently analyzed the retrieved bundles in terms of the
syntactic roles they played in the sentence. Concordances surrounding the occurrences of LBs
were examined qualitatively to determine their discursive and rhetorical functions within a
broader context. This allowed us to analyze the construction of LBs produced by Persian
writers and compare them to those of native-speaker writers from the perspective of L1
transfer, overuse, or misuse.
Results
The analysis of the lexical bundles revealed that L2 academic writers employed more types
and tokens of LBs than L1 academic writers. This suggests that L2 writers relied more heavily
on LBs than L1 writers. The final lists of four-word bundles produced by L1 and L2 academic
writers are presented in the Appendix. These bundles have been identified after excluding
topic-dependent and discipline-dependent bundles. Table 3 presents the number of types and
tokens of LBs in the two writer groups.
A Corpus-based Study on the Use and Syntactic Functions of Lexical Bundles in Applied Linguistics Research Articles in Two Contexts of Publications 149
AREL
Table 3. Number of Types and Tokens of Lexical Bundles in Two Pairs of Corpora
Writer groups
Types
Tokens
Native-speaker academic writers
54
2004
Nonnative academic writers
103
4079
Total
157
6083
Closer analysis of bundles revealed that 27 bundles were found to have occurred in both
corpora. Table 4 shows the bundles with the normalized token frequency of occurrences in NC
and NNC. As Table 4 illustrates, nearly 55% of the retrieved LBs are PP-based bundles, 39%
are NP-based bundles, and only 6% of shared LBs are VP-based bundles. These bundles were
used with different frequencies in the two corpora.
Table 4. Shared Bundles with Normalized Frequency per 1,000,000 Words
Rank (NC)
Token (NC)
Rank (NNC)
Token (NNC)
on the other hand
1
86.93
2
155.8
the extent to which
3
71.59
18
54.32
as well as the
4
61.36
13
57.4
in the context of
5
60.34
7
78.92
at the same time
6
59.32
64
26.65
in the present study
7
59.32
9
72.77
on the basis of
8
59.32
60
29.72
the results of the
9
59.32
1
218.32
in the current study
11
54.2
17
55.35
in the case of
12
53.18
21
52.27
at the time of
15
49.09
72
24.6
on the role of
16
42.95
53
31.77
in the field of
17
41.93
19
53.3
in the form of
20
39.88
41
36.9
with respect to the
23
36.82
61
28.7
as a result of
24
35.79
12
57.4
in addition to the
25
34.77
83
23.57
in terms of the
26
34.77
23
49.2
the students in the
28
32.73
57
30.75
the nature of the
30
31.7
97
21.52
a wide range of
31
29.66
100
20.5
the meaning of the
34
28.64
96
21.52
to be able to
36
27.61
67
26.65
on the one hand
37
26.59
77
24.6
in line with the
39
25.57
6
85.07
on the part of
53
20.45
84
23.57
the participants in the
54
20.45
47
34.85
150 Applied Research on English Language, V. 10 N. 4 2021
AREL
LBs in each group were classified structurally using Biber et al.’s (1999) taxonomy.
Accordingly, three are broad categories of VP-based bundles, NP-based bundles, and PP-based
bundles have been distinguished. Table 5 presents the structural distribution of bundle types in
both corpora.
Table 5. Structural Distribution of LBs in NC and NNC
Structural
subcategories
Native-English
writers (%)
Persian writers
(%)
NP-based
bundles
NP with of-phrase fragment
450(0.22)
1016(0.25)
NP with other post-modifier
fragments
117(0.06)
371(0.09)
Other noun phrase
45(0.02)
164(0.04)
Total
612 (0.31)
1551(0.38)
PP-based bundles
PP phrase with embedded of-phrase
fragment
469(0.23)
780(0.19)
Other prepositional phrase fragment
501(0.25)
542(0.13)
Total
970 (0.48)
1322(0.32)
VP-based
bundles
Copular be + NP/Adj. phrase
45(0.02)
216(0.05)
Anticipatory it + VP/Adj. phrase
75(0.04)
162(0.04)
Passive verb + prepositional phrase
fragment
32(0.02)
133(0.03)
VP + that-clause fragment
27(0.01)
140(0.03)
Verb/adjective + to-clause fragment
24(0.01)
49(0.01)
Verb phrase with active verb
23(0.01)
46(0.01)
Adverbial clause fragment
39(0.02)
74(0.02)
Pronoun/noun phrase + be + (…)
22(0.01)
10(0)
Total
287 (0.14)
830(0.20)
Other expressions
135 (0.07)
376(0.09)
Total
2004
4079
A Corpus-based Study on the Use and Syntactic Functions of Lexical Bundles in Applied Linguistics Research Articles in Two Contexts of Publications 151
AREL
VP-based bundles comprised the least proportion of identified bundles in both corpora
in this study (NC: 14%, NNC: 20%). These bundles were subsequently categorized based on
their syntactic roles in relation to a subset of Biber et al.’s (2011) hypothesized stages of writing
development. Table 6 presents the syntactic roles of VP bundles as well as the frequency of the
occurrence of each type, which are compared between two writer groups by means of a log-
likelihood test.
Table 6. Distribution of Syntactic Roles of VP-based Bundles in NC and NNC
Stage
Syntactic Roles
NC
NNC
1
Finite complement clause (CC) controlled by common verbs*
20(0.07)
78(0.09)
2
Finite CC controlled by wider set of verbs
25(0.09)
62(0.07)
Finite adverbial clauses
60(0.21)
185(0.22)
Nonfinite CC, controlled by common verbs
23(0.08)
135(0.16)
3
Finite CC controlled by adjectives
11(0.04)
63(0.08)
Nonfinite CC Controlled by wider set of verbs
45(0.16)
96(0.12)
That relative clauses, especially with animate head nouns
50(0.17)
113(0.14)
4
Nonfinite CC controlled by adjectives
15(0.05)
26(0.03)
Extraposed CC
3(0.01)
13(0.02)
Nonfinite relative clauses
17(0.06)
31(0.04)
5
CC controlled by nouns
4(0.01)
11(0.01)
Other
14(0.05)
17(0.02)
Total
287 (100%)
830 (100%)
Table 6 presents the syntactic functions of VP-based bundles which are compared based
on the number of tokens. The findings revealed that finite adverbial clauses were the most
frequent category of VP-based bundles used in NC. They were followed by that relative
clauses and nonfinite complement clauses. NNC, similarly, showed the heaviest reliance on
finite adverbial clauses which were followed by nonfinite complement clauses controlled by
common verbs, and that relative clauses. The results of log-likelihood showed that none of the
syntactic categories showed a significant difference between the two writer groups.
Persian academic writers demonstrated a greater reliance on NP-phrase bundles than
native academic writers. On the whole, NP-phrase bundles comprised 31% of LBs in NC, while
for NNC the figure is 38%, a substantially, and statistically significant difference. Table 7
152 Applied Research on English Language, V. 10 N. 4 2021
AREL
shows the subcategories of the syntactic roles with the results obtained from the log-likelihood
test for each role.
Table 7. Distribution of Syntactic Roles of Noun-phrase bundles in NC and NNC
Syntactic Role
NC
NNC
Subject**
112(0.18)
482(0.31)
Subject predicative*
97(0.16)
381(0.25)
Direct object*
139(0.23)
202(0.13)
Indirect object
12(0.02)
23(0.01)
Agent in passive voice
6(0.01)
77(0.05)
of-phrase as postmodifier**
195(0.32)
264(0.17)
Relative clause
12(0.02)
35(0.02)
Other
39(0.06)
87(0.06)
Total
612 (100%)
1551 (100%)
Note. **significant at p < 0.001. * = Significant at p < 0.05
As presented in Table 7, both corpora have a different proportion of NP-based bundles,
with NC relying mostly on of-phrase as post-modifiers, and NNC on the subject, which
accounted for 32% and 31% of all NP bundles, respectively. In NC, of-phrase as post-modifiers
was followed by direct object, subject, subject predicative, indirect object, relative clause, and
agent in passive voice. Other bundles accounted for 6% of all NP-based bundles in NC.
However, different patterns of results were observed in NNC, where the second most frequent
bundles were found to be subject predicative, followed by of-phrase as post-modifiers, direct
object, agent in passive voice, indirect object, and relative clause. Other bundles made up 5%
of all NP-based bundles. The results obtained from the log-likelihood test revealed that
significant differences were found in the frequency of the four syntactic roles of subject, subject
predicative, direct object, and of-phrase as postmodifier. NNC made greater use of subject and
subject predicative bundles than NC did, while NC relied more heavily on the direct object,
and of-phrase as postmodifier than NNC.
PP-based bundles constituted the largest proportion of all bundle types in NC (48%),
while for NNC they were the second-largest proportion (32%) after NP-based bundles. As
shown in Table 8, LBs as adverbials were a more frequent type of PP-based bundles in NNC.
In NC, 23% of PP-based bundles were adverbials, while for NNC the figure is 77%, a
substantial and statistically significant difference. Native-speaker writers relied more heavily
A Corpus-based Study on the Use and Syntactic Functions of Lexical Bundles in Applied Linguistics Research Articles in Two Contexts of Publications 153
AREL
on LBs such as post-nominal modifier (65%) than nonnative writers (23%). This suggests that
a larger number of PP-based bundles in NC occur in syntactically more complex units (post-
nominal modifiers as opposed to adverbials) compared to those of NNC (see Biber et al.’s
(2011) hypothesized stages of writing development).
Table 8. Distribution of Syntactic Roles of PP-based Bundles in NC and in NNC
Syntactic Role
NC
NNC
Adverbial*
340 (0.35)
1021 (0.77)
Post-nominal modifier*
630 (0.65)
305 (0.23)
Total
970 (100%)
1322 (100%)
Note. * = Significant at p < 0.05
Discussion
The purpose of the present study was to compare lexical bundles used by L1 Persian and L1
English academic writers. The results of the study indicated that Persian academic writers made
greater use of LBs at a higher frequency than English academic writers. Structural analysis of
LBs revealed that PP-based bundles made up the greatest proportion of all bundle types in
NNC, followed by NP-based bundles, and VP-based bundles. However, NC showed different
patterns of use where PP-based bundles constituted the largest proportion, followed by NP-
based bundles, and VP-based bundles. Retrieved bundles were also examined in terms of the
syntactic roles of the units in which they occurred. Significant differences were found for the
syntactic roles of NP-based and PP-based LBs between the two writer groups. The syntactic
roles of VP-based bundles, however, showed no significant differences between the groups.
The finding that VP-based bundles were the least favored bundles in the entire corpus is
not surprising given that clausal bundles are more extensively used in the spoken register than
academic writing. This finding supports that of Biber et al. (1999), who argued that the majority
of the bundles in academic writing are phrasal bundles. Similarly, Hyland (2008) noted that
“most bundles in academic writing are parts of noun or prepositional phrases” (p. 9). The
writers’ reliance on phrasal bundles reveals that both groups are aware of the way information
is densely packed into phrasal groups (see Fang, Schleppegrell, & Cox, 2006; Staples, Egbert,
Biber, & Gray, 2016). However, PP-based bundles were the most frequent bundles in NC,
while NP-based bundles comprised the largest group of bundles in NNC. This finding supports
that of Chen and Baker (2010), who found that expert writers tend to use more NP/PP-based
bundles and fewer VP-based bundles.
154 Applied Research on English Language, V. 10 N. 4 2021
AREL
The fact that Persian L1 writers made greater use of LBs at a higher frequency than L1
English writers is notable, suggesting that the former group drew on their lexicalized
knowledge to construct academic research articles to a greater extent than the latter group did.
“Although greater use of the target bundles may indicate L2 phraseological development,
learners may also develop their competence in RMCs [recurrent multiword combinations] that
do not pass the strict corpus-based distributional criteria for bundles” (Chen, 2019, p. 6). The
findings of the present study are consistent with those of Ahmadi, Esfandiari, and Zarei (2020),
who revealed that Persian writers used significantly more lexical bundles of all types as noun
modifiers compared to native writers. In the same vein, Shahmoradi, Jalali, and Ghadiri (2021)
have revealed that L1 Persian writers used more LBs in RAs in applied linguistics and
information technology than did their native-speaker counterparts. Similarly, Lu and Deng
(2019) found that Chinese doctoral students used LBs more frequently than their native-speaker
counterparts, although they “exhibited incomplete knowledge of some aspects of the English
lexico-grammatical system” (p. 1).
Analysis of shared bundles in our study revealed that they have been used with different
frequencies in both corpora. However, four PP-based bundles (i.e. in the current study, in the
case of, to be able to, for example in the) show a similar pattern of use in NC and NNC.
Previous research has suggested that these LBs are among the most common bundles in the
academic register, and RAs in particular (e.g., Bychkovska & Lee, 2017; Chen & Barker, 2010;
Hyland, 2012; Pan & Liu, 2019). Out of 53 shared bundles, 30 were used more frequently in
NC, and 23 were used more frequently in NNC (See the Appendix).
As noted above, certain bundles were overused in NNC, while the LBs which are
commonly used in academic writing were either underused or were nonexistent in NC. In
addition, a great number of LBs were used differently in terms of syntactic roles or discursive
features in NNC compared to those of NC. The following examples show how two groups of
writers used in the process of. In NC, the bundle was often employed as a subject predicative
after copula be-verb, or as the post-modification of an NP, whereas in NNC the bundle often
occurred in the sentence-initial position functioning as the premodification of an NP.
(1) All participating youth are in the process of learning English. (NC)
(2) The ‘framing’ power of metaphor constitutes this bias in the process of
conceptualization. (NC)
(3) It appears that in the process of EFL teacher recruitment and selection there should
be a variety of selection stages and methods. (NNC)
A Corpus-based Study on the Use and Syntactic Functions of Lexical Bundles in Applied Linguistics Research Articles in Two Contexts of Publications 155
AREL
Similarly, the bundle on the other hand, which was found to have been far more common
in NNC than in NC, was not actually used appropriately by Persian L1 writers. Native writers
generally use the bundle “to introduce a contrary view of the previous sentence” (Pan & Liu,
2019, p. 153). However, a closer investigation of concordance lines revealed that Persian
writers seemed to employ on the other hand as a text-linking bundle for joining any types of
ideas (especially additive markers) irrespective of any contrasting links between them. A
considerable proportion of all the occurrences in NNC were found to be inappropriate.
Examples 4 and 5 show the use of this bundle in NNC and NC, respectively.
(4) Considering native speakers, this paper tries to tentatively develop a PP which
contributes to the way of utilizing metadiscourse units in spoken genres. On the
other hand, the current study aims to apply the PP and its maxims to the analysis
of three spoken genres. (NNC)
(5) Much of the contribution of LP to multiple-documents comprehension is mediated
via impacting single-text comprehension. On the other hand, a smaller share of
the contribution of PK to multiple-texts comprehension is mediated through single-
text comprehension and a larger share of it is unmediated. (NC)
An important finding of the current study is that PP-based bundles were employed
proportionally less frequently in NNC than in NC. The most frequent bundles in both corpora
were the sequences of preposition + NP + of (e.g., in the case of). Such structures are hallmarks
of advanced academic writing because they “are highly productive in sentence framing” (Ruan,
2017, p. 9). L2 writers’ underuse of prepositional phrases in general and overuse of particular
common academic structures (such as in the context of) suggest that they may be familiar with
their functions in academic writing, but they “cling to words or phrases with which they feel
comfortable using” (Appel & Wood, 2016, p. 66).
As for syntactic roles of NP-based bundles, Persian L1 writers were found to have used
significantly more LBs in subject and subject predicative positions than English L1 writers.
On the other hand, English L1 writers relied more heavily on LBs as direct object and of-phrase
as postmodifier than L1 Persian writers. Persian L1 writers’ greater use of LBs in the subjective
position indicates their tendency to overuse sentence-initial bundles. As Grabowski (2015)
pointed out, a great number of high-frequency bundles in the sentence-initial position are
156 Applied Research on English Language, V. 10 N. 4 2021
AREL
typical of non-academic spoken discourse. Similar to the results of the present study, Shin
(2018) and Li, Franken, and Wu (2019) have found that nonnative academic writers tend to use
LBs in the sentence-initial position. In their study of Chinese postgraduate students’ sources of
sentence-initial bundles in their thesis writing, Li and her colleagues found that such reasons
as interlingual transfer, literal transfer, semantic transfer, and transfer of training accounted for
the sources of a major proportion of the LBs used in the subjective position. The following
examples demonstrate how the same LB is used in sentence-medial and sentence-initial
positions in NC and NNC, respectively.
(6) The revised principles informed the design of the second-year ELA curriculum and
enabled us to propose new instructional theories. (NC)
(7) The design of the present study was both quantitative and qualitative; therefore,
mixed method is applied. (NNC)
The more frequent use of of-phrase as postmodifier in NC compared to NNC indicates
that L1 English writers are more attuned to these constructions as important academic writing
conventions. The following examples indicate how LBs are used in syntactic units functioning
as of-phrase as postmodifier in NC (8) and NNC (9).
(8) For example, the plural marker at the end of the verb is redundant because number
is expressed by the subject. (NC)
(9) Both learners and their instructors were asked to provide information on the
content of the courses, particularly as related to pronunciation. (NNC)
In comparison, English native writers often used NP-based bundles within of-phrase
postmodifiers functioning as nominal modifiers, while Persian native writers often employed
them as adverbials. The former contributes to a compressed discourse style, whereas the latter
results in an elaborated discourse style (See Biber & Gray, 2010; Biber et al., 2011; Biber &
Gray, 2016). The following examples from NC and NNC show how NP-based bundles are
used to function as adverbials.
(10) By alternating learning and test trials, we were able to examine how cue use and
relative strength changed over the course of learning. (NC)
A Corpus-based Study on the Use and Syntactic Functions of Lexical Bundles in Applied Linguistics Research Articles in Two Contexts of Publications 157
AREL
(11) They commented on the design of the semi-structured interview, adequacy and
usefulness of the questions, and adjustments were made accordingly. (NNC)
According to Biber et al. (2011), prepositional phrases as adverbials are acquired at
earlier stages of writing development compared to prepositional phrases as post-nominal
modifiers. The more frequent use of these structures in postnominal prepositional phrases in
NC suggests that English L1 academic writers used a greater proportion of NP-based bundles
in more complex syntactic units than Persian L1 academic writers did. This different pattern
of reliance may be due to dissimilar amounts of exposure to these structures. Persian writers
may still need more exposure to compressing lexico-grammatical features required for
academic research writing.
Similar differences could also be observed in PP-based bundles where English L1 writers
used post-nominal modifiers significantly more frequently than Persian L1 writers. As Biber
et al. (1999) put it, postmodifying prepositional phrases are the most common type of
postmodifier in the written register in general and in academic writing in particular. They
further argue that many of the most common frequent LBs in academic writing include of-
phrases prepositional phrases because they mark abstract/logical/physical relations. Examples
12 and 13 demonstrate how two groups of writers used PP-based bundles functioning as
postnominal prepositional phrases to show meaning relationships.
(12) It led participants to form a predictive strategy such that they might have
predicted to produce regulars in the absence of irregulars in the experimental list.
(NC)
(13) In the literature on teacher candidates’ identity, reflection is widely considered
as a critical process in the development of teacher professional identity. (NNC)
Biber and Gray (2010) asserted that the recurrent use of post-modifying prepositional
phrases, and of-phrases inter alia, indicates the less explicit and more complex nature of
academic writing in which a great deal of meaning is embedded in phrasal expressions.
Accordingly, we can safely argue that the more frequent use of LBs in PP-based syntactic units
adds to the complexity of the texts. This finding is in line with that of Shin (2018), who found
that native academic writers used more than four times as many postnominal prepositional
phrases as nonnative academic writers did.
158 Applied Research on English Language, V. 10 N. 4 2021
AREL
Phrasal embedding as postmodifiers has been proposed as the most complicated feature
in Biber et al.’s (1999) hypothesized stages of writing development. Several studies have
documented that advanced academic writing relies heavily on phrasal features, many of which
are postnominal prepositional phrases as opposed to post-modifying prepositional phrases
functioning as adverbials (e.g., Parkinson & Musgrave, 2014; Staples et al., 2016; Taguchi,
Crawford, & Wetzel, 2013). Postnominal prepositional phrases contribute to the complexity of
clauses. Fang et al. (2006) argued that expanded nominal groups (e.g., postnominal
prepositional phrases) can compress information that could otherwise take different clauses to
convey into a single clause. These compressing elements are central features of advanced
academic writing, as they facilitate the flow of information and the development of a complex
discourse style.
Conclusion
The present study has examined the use of LBs in RAs authored by English L1 and Persian L1
academic writers in applied linguistics, compiled from two corpora of RAs from leading
international journals and Persian English-medium journals. Four-word LBs in both corpora
were retrieved and their frequency distribution and syntactic roles in the clause were compared
between writer groups. The findings revealed that Persian writers made greater use of LBs at a
higher frequency than English academic writers.
Identified bundles were subsequently categorized based on Biber et al.’s (1999)
taxonomy. It was found that VP-based bundles were the least frequently used structural
category in both NC and NNC. PP-based bundles constituted the largest proportion of all
bundles in NC, followed by NP-based bundles. NP-based bundles, however, accounted for the
most common structure in NC followed by PP-based bundles. The analysis of syntactic roles
of LBs in the clause indicated that Persian writers tended to use NP-based bundles in the
sentence-initial position, whereas English writers often used the expressions in sentence medial
position. As for PP-based bundles, adverbials made up the greatest proportion of all PP-based
bundles in NNC, while postnominal prepositional phrases were the largest sub-category in NC.
Given that VP-based bundles constituted the smallest proportion of LBs and that no
significant differences were found between L1 Persian and L1 English academic writers in
terms of syntactic functions of VP-based bundles, it seems that Persian writers are already
familiar with the structural/distributional/functional features of VP-based bundles in the
academic register and know how to use them in the same way as expert native English
A Corpus-based Study on the Use and Syntactic Functions of Lexical Bundles in Applied Linguistics Research Articles in Two Contexts of Publications 159
AREL
academic writers do. However, based on Biber et al.’s (1999) hypothesized stages of writing
development where progression starts from clausal features to phrasal features, particularly
multiple prepositional phrases which are the most advanced level of developmental category,
L1 English writers in our study, who predominantly employed LBs as PP-based bundles mostly
functioning as post-modifying prepositional phrases, appeared to rely on syntactically more
complex bundles than did L1 Persian writers.
The findings of the current study have several pedagogical implications. In addition to
structural and functional classifications of LBs, syntactically developmental classifications of
LBs can also be developed, and LBs generated on the basis of these classifications could be
integrated into academic writing courses. The explicit instruction of syntactically complex LBs
seems necessary, as an increasing number of studies have shown that advanced lexico-
grammatical features in writing, particularly LBs, are not naturally acquired in the same way
as complex language features in spoken register (Biber et al., 2011; Cortes, 2004; Staples et
al., 2016; Wei & Lei, 2011). Accordingly, L2 writers need to be explicitly aware of the way
complex ideas are embedded in compressing language features through the use of LBs. This
study has also shown that native academic writers tended to use certain bundles in particular
positions in the sentence which differed from those of nonnative academic writers. Therefore,
it seems that instruction in LB usage may benefit from corpus-based learning approaches for
exploring, comparing, and analyzing the positional distribution of bundles to resolve any
discrepancies in the rhetorical conventions of LBs in advanced academic writing (see Li et al.,
2019).
Although corpus-based studies provide invaluable insight into patterns of L2 writers’
language use and guide researchers in hypothesizing sources of deviations from target norms,
corpus data does not explain why language users opt for particular features while writing
(Hyland, 2012). Accordingly, future contrastive analyses of LBs could carry out qualitative
analysis such as interviews to complement quantitative methods and to elicit L2 writers’
“interpretation of their own bundle choices” (Li et al., 2019, p. 3).
Declaration of Interests
The authors of this study declare that they have no known competing financial interests or
personal relationships that could have appeared to influence the work reported in this paper.
160 Applied Research on English Language, V. 10 N. 4 2021
AREL
References
Adel, A., & Erman, B. (2012). Recurrent word combinations in academic writing by native and
non-native speakers of English: A lexical bundles approach. Journal of English for
Specific Purposes, 31(2), 81-92.
Ahmadi, M., Esfandiari, R., & Zarei, A. A. (2020). A corpus-based study of noun phrase
complexity in applied linguistics research article abstracts in two contexts of
publication. Iranian Journal of English for Academic Purposes, 9(1), 76-94.
Anthony, L. (2020). Antconc: A freeware corpus analysis toolkit for concordancing and text
analysis. Retrieved from: http://www.laurenceanthony.Net/software.html
Atai, M. R., & Tabandeh, F. (2015). Lexical bundles in applied linguistics articles: Exploring
writer, sub-discipline, and sub-genre variations. Journal of ESP across Cultures, 11, 33-
56.
Biber, D., & Barbieri, F. (2007). Lexical bundles in university spoken and written
registers. Journal of English for Specific Purposes, 26(3), 263-286.
Biber, D., & Gray, B. (2010). Challenging stereotypes about academic writing: Complexity,
elaboration, explicitness. Journal of English for Academic Purposes, 9(1), 2-20.
Biber, D., & Gray, B. (2013). Nominalizing the verb phrase in academic science writing. In B.
Aarts, J. Close, Leech, G., & S. Wallis (Eds.), The English verb phrase: Corpus
methodology and current change (pp. 99-132). Cambridge: Cambridge University Press.
Biber, D., & Gray, B. (2016). Grammatical complexity in academic English: Linguistic change
in writing. Cambridge: Cambridge University Press.
Biber, D., Conrad, S., & Cortes, V. (2004). If you look at…: Lexical bundles in university
teaching and textbooks. Applied Linguistics, 25(3), 371-405.
Biber, D., Gray, B., & Poonpon, K. (2011). Should we use characteristics of conversation to
measure grammatical complexity in L2 writing development?. TESOL Quarterly, 45(1),
5-35.
Biber, D., Johansson, S., Leech, G., Conrad, S., & Finegan, E. (1999). The Longman grammar
of spoken and written English. London: Longman.
Bychkovska, T., & Lee, J. J. (2017). At the same time: Lexical bundles in L1 and L2 university
student argumentative writing. Journal of English for Academic Purposes, 30, 38-52.
Chen, A. C. H. (2019). Assessing phraseological development in word sequences of variable
lengths in second language texts using directional association measures. Language
Learning, 69(2), 440-477.
A Corpus-based Study on the Use and Syntactic Functions of Lexical Bundles in Applied Linguistics Research Articles in Two Contexts of Publications 161
AREL
Chen, Y. H., & Baker, P. (2010). Lexical bundles in L1 and L2 academic writing. Journal of
Language Learning & Technology, 14(2), 30-49.
Conklin, K., & Schmitt, N. (2008). Formulaic sequences: Are they processed more quickly
than nonformulaic language by native and nonnative speakers?. Applied
linguistics, 29(1), 72-89.
Cortes, V. (2004). Lexical bundles in published and student disciplinary writing: Examples
from history and biology. Journal of English for Specific Purposes, 23(4), 397-423.
Cortes, V. (2008). A comparative analysis of lexical bundles in academic history writing in
English and Spanish. Corpora, 3(1), 43-57.
Cortes, V. (2015). Situating lexical bundles in the formulaic language spectrum: Origins and
functional analysis developments. In V. Cortes, & E. Csomay (Eds.), Corpus-based
research in applied linguistics: Studies in honor of Doug Biber (pp. 197-218). John
Benjamins.
Csomay, E. (2013). Lexical bundles in discourse structure: A corpus-based study of classroom
discourse. Applied Linguistics, 34(3), 369-388.
Ellis, N. C. (2012). Formulaic language and second language acquisition: Zipf and the phrasal
teddy bear. Annual Review of Applied Linguistics, 32, 17-44.
Ellis, N. C., & Simpson-Vlach, R. (2009). Formulaic language in native speakers:
Triangulating psycholinguistics, corpus linguistics, and education. Journal of Corpus
Linguistics and Linguistic Theory, 5, 61–78.
Esfandiari, R., & Barbary, F. (2017). A contrastive corpus-driven study of lexical bundles
between English writers and Persian writers in psychology research articles. Journal of
English for Academic Purposes, 29, 21-42.
Fang, Z., Schleppegrell, M. J., & Cox, B. E. (2006). Understanding the language demands of
schooling: Nouns in academic registers. Journal of Literacy Research, 38(3), 247-273.
Garner, J., Crossley, S., & Kyle, K. (2019). N-gram measures and L2 writing proficiency.
System, 80, 176-187.
Grabowski, Ł. (2015). Keywords and lexical bundles within English pharmaceutical discourse:
A corpus-driven description. Journal of English for Specific Purposes, 38, 23-33.
Gries, S. T., & Ellis, N. C. (2015). Statistical measures for usage-based linguistics. Language
Learning, 65(S1), 228-255.
Güngör, F., & Uysal, H. H. (2016). A comparative analysis of lexical bundles used by native
and non-native scholars. English Language Teaching, 9(6), 176-188.
162 Applied Research on English Language, V. 10 N. 4 2021
AREL
Harzing, A. W. K., & Van der Wal, R. (2008). Google Scholar as a new source for citation
analysis. Journal of Ethics in Science and Environmental Politics, 8(1), 61-73.
Hirsch, J. E. (2005). An index to quantify an individual's scientific research
output. Proceedings of the National Academy of Sciences, 102(46), 16569-16572.
Hyland, K. (2008). As can be seen: Lexical bundles and disciplinary variation. English for
Specific Purposes, 27(1), 4-21.
Hyland, K. (2012). Bundles in academic discourse. Annual Review of Applied Linguistics, 32,
150-169.
Jiang, N. A., & Nekrasova, T. M. (2007). The processing of formulaic sequences by second
language speakers. The Modern Language Journal, 91(3), 433-445.
Lan, G., & Sun, Y. (2019). A corpus-based investigation of noun phrase complexity in the L2
writings of a first-year composition course. Journal of English for Academic
Purposes, 38, 14-24.
Li, L., Franken, M., & Wu, S. (2019). Chinese postgraduates’ explanation of the sources of
sentence initial bundles in their thesis writing. RELC Journal, 50(1), 37-52.
Lu, X., & Deng, J. (2019). With the rapid development: A contrastive analysis of lexical
bundles in dissertation abstracts by Chinese and L1 English doctoral students. Journal of
English for Academic Purposes, 39, 21-36.
Pan, F., & Liu, C. (2019). Comparing L1-L2 differences in lexical bundles in student and expert
writing. Southern African Linguistics and Applied Language Studies, 37(2), 142-157.
Pan, F., Reppen, R., & Biber, D. (2016). Comparing patterns of L1 versus L2 English academic
professionals: Lexical bundles in Telecommunications research journals. Journal of
English for Academic Purposes, 21, 60-71.
Pan, F., Reppen, R., & Biber, D. (2020). Methodological issues in contrastive lexical bundle
research: The influence of corpus design on bundle identification. International Journal
of Corpus Linguistics, 25(2), 215-229.
Parkinson, J., & Musgrave, J. (2014). Development of noun phrase complexity in the writing
of English for Academic Purposes students. Journal of English for Academic
Purposes, 14, 48-59.
Rahimi Azad, H., & Modarres Khiabani, S. (2018). Lexical bundles in English abstracts of
research articles written by Iranian scholars: Examples from Hhumanities. Iranian
Journal of Applied Language Studies, 10(2), 149-174.
Romer, U., & Arbor, A. (2009). English in academia: Does nativeness matter. Anglistik:
International Journal of English Studies, 20(2), 89-100.
A Corpus-based Study on the Use and Syntactic Functions of Lexical Bundles in Applied Linguistics Research Articles in Two Contexts of Publications 163
AREL
Ruan, Z. (2017). Lexical bundles in Chinese undergraduate academic writing at an English
medium university. RELC Journal, 48(3), 327-340.
Ruan, Z. (2018). Structural compression in academic writing: An English-Chinese comparison
study of complex noun phrases in research article abstracts. Journal of English for
Academic Purposes, 36(1), 37-47.
Salazar, D. (2014). Lexical bundles in native and non-native scientific writing: Applying a
corpus-based study to language teaching. UK: John Benjamins Publishing Company.
Shahmoradi, N., Jalali, H., & Ghadiri, M. (2021). Lexical bundles in the abstract and
conclusion sections: The case of applied linguistics and information technology. Applied
Research on English Language, 10(3), 47-76.
Shin, Y. K. (2018). The construction of English lexical bundles in context by native and
nonnative freshman university students. English Teaching, 73(3), 115-139.
Shirazizadeh, M., & Amirfazlian, R. (2021). Lexical bundles in theses, articles and textbooks
of applied linguistics: Investigating intradisciplinary uniformity and variation. Journal of
English for Academic Purposes, 49, 100946.
Staples, S., Egbert, J., Biber, D., & Gray, B. (2016). Academic writing development at the
university level: Phrasal and clausal complexity across level of study, discipline, and
genre. Journal of Written Communication, 33(2), 149-183.
Staples, S., Egbert, J., Biber, D., & McClair, A. (2013). Formulaic sequences and EAP writing
development: Lexical bundles in the TOEFL iBT writing section. Journal of English for
Academic Purposes, 12(3), 214-225.
Taguchi, N., Crawford, W., & Wetzel, D. Z. (2013). What linguistic features are indicative of
writing quality? A case of argumentative essays in a college composition program. Tesol
Quarterly, 47(2), 420-430.
Wang, M., & Zhang, Y. (2021). ‘According to…’: The impact of language background and
writing expertise on textual priming patterns of multi-word sequences in academic
writing. Journal of English for Specific Purposes, 61, 47-59.
Wei, Y., & Lei, L. (2011). Lexical bundles in the academic writing of advanced Chinese EFL
learners. RELC Journal, 42(2), 155-166.
Wood, A. (2001). International scientific English: The language of research scientists around
the world. In J. Flowerdew, & M. Peacock (Eds.), Research perspectives on English for
academic purposes (pp. 71-83). Cambridge University Press.
164 Applied Research on English Language, V. 10 N. 4 2021
AREL
The Complete List of Lexical Bundles in NC and NNC with Normalized Frequency per
1,000,000 Words
Rank
NC
Token
Type
NNC
Token
Type
1
on the other hand
86.93
44
the results of the
218.32
64
2
the end of the
72.61
34
on the other hand
155.8
64
3
the extent to which
71.59
30
of the present study
124.02
45
4
as well as the
61.36
39
the findings of the
105.57
50
5
in the context of
60.34
34
significant difference
between the
87.12
31
6
at the same time
59.32
35
in line with the
85.07
46
7
in the present study
59.32
22
in the context of
78.92
41
8
on the basis of
59.32
27
at the end of
72.77
39
9
the results of the
59.32
20
in the present study
72.77
39
10
as a function of
54.2
19
the first research question
62.52
39
11
in the current study
54.2
24
as shown in table
60.47
34
12
in the case of
53.18
30
as a result of
57.4
30
13
it is important to
53.18
36
as well as the
57.4
31
14
the ways in which
53.18
23
the results indicated that
57.4
31
15
at the time of
49.09
24
the second research question
57.4
37
16
on the role of
42.95
8
in the process of
56.37
35
17
in the field of
41.93
21
in the current study
55.35
33
18
in relation to the
40.91
28
the extent to which
54.32
28
19
at the beginning of
39.88
20
in the field of
53.3
31
20
in the form of
39.88
29
the participants of the
53.3
29
21
in this study we
38.86
20
in the case of
52.27
23
22
there was a
significant
37.84
17
is one of the
50.22
33
23
with respect to the
36.82
19
in terms of the
49.2
24
24
as a result of
35.79
21
with regard to the
49.2
25
25
in addition to the
34.77
25
it was found that
48.17
29
26
in terms of the
34.77
24
the reliability of the
48.17
30
27
it is possible that
34.77
25
the purpose of the
45.1
32
28
the students in the
32.73
10
as one of the
44.07
28
29
the fact that the
31.7
21
in other words the
44.07
30
30
the nature of the
31.7
21
on the development of
44.07
15
31
a wide range of
29.66
20
the present study was
43.05
28
32
one of the most
29.66
23
to the fact that
43.05
30
A Corpus-based Study on the Use and Syntactic Functions of Lexical Bundles in Applied Linguistics Research Articles in Two Contexts of Publications 165
AREL
33
over the course of
29.66
12
descriptive statistics of the
42.02
20
34
the meaning of the
28.64
18
the analysis of the
42.02
23
35
the use of the
27.61
21
it can be claimed
41
8
36
to be able to
27.61
20
in the use of
39.97
8
37
on the one hand
26.59
20
the following research
questions
39.97
39
38
the onset of the
26.59
6
the results showed that
39.97
22
39
in line with the
25.57
19
of the three groups
38.95
7
40
in the absence of
25.57
15
development and validation
of
36.9
5
41
were more likely to
24.54
12
in the form of
36.9
24
42
a main effect of
23.52
10
the beginning of the
36.9
24
43
as can be seen
23.52
14
the content of the
36.9
19
44
as the dependent
variable
23.52
11
be attributed to the
35.87
21
45
can be used to
22.5
14
can be concluded that
35.87
27
46
the results of this
22.5
14
in this study the
34.85
26
47
as a measure of
21.48
15
the participants in the
34.85
20
48
as part of the
21.48
16
theory and practice in
34.85
27
49
at the level of
21.48
13
they were asked to
33.82
21
50
for each of the
21.48
15
test for equality of
32.8
10
51
than those in the
21.48
8
the mean score of
32.8
17
52
the number of words
21.48
9
on the role of
31.77
64
53
on the part of
20.45
11
can be seen in
30.75
18
54
the participants in the
20.45
11
of the control group
30.75
20
55
the results revealed that
30.75
8
56
the students in the
30.75
21
57
used in this study
30.75
17
58
it should be noted
29.72
25
59
on the basis of
29.72
20
60
with respect to the
28.7
17
61
a systematic review of
27.67
17
62
are presented in table
27.67
6
63
at the same time
26.65
16
64
in the control group
26.65
17
65
it can be argued
26.65
11
66
to be able to
26.65
12
67
a large number of
25.62
17
166 Applied Research on English Language, V. 10 N. 4 2021
AREL
68
experimental and control
groups
25.62
20
69
in the course of
25.62
8
70
as indicated in table
24.6
11
71
at the time of
24.6
13
72
immediate and delayed
posttests
24.6
19
73
in a similar vein
24.6
5
74
of the fact that
24.6
21
75
on the acquisition of
24.6
16
76
on the one hand
24.6
8
77
the descriptive statistics of
24.6
17
78
this study aimed to
24.6
16
79
was an attempt to
24.6
20
80
for the sake of
23.57
18
81
in a way that
23.57
14
82
in addition to the
23.57
18
83
on the part of
23.57
19
84
a comparative study of
22.55
19
85
as far as the
22.55
14
86
as the most important
22.55
12
87
be due to the
22.55
5
88
in the experimental group
22.55
15
89
investigate the effect of
22.55
6
90
items of the questionnaire
22.55
14
91
on the use of
22.55
10
92
to analyze the data
22.55
14
93
to participate in the
22.55
18
94
the majority of the
21.52
20
95
the meaning of the
21.52
14
96
the nature of the
21.52
10
97
the needs of the
21.52
12
98
a case study of
20.5
7
99
a wide range of
20.5
17
100
one of the main
20.5
13
101
so that they can
20.5
16
102
the impact of the
20.5
14
103
was found to be
20.5
14
