LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

Noun Phrase Complexity in Academic Writing: A Comparison of

Argumentative English Essays Written by Thai and Native English
University Students
Narisa Jitpraneechai
Chulalongkorn University Language Institute, Bangkok, Thailand
[email protected]

Abstract
Focusing on noun phrase complexity in writing, this study adopted Biber, Gray and
Poonpon’s (2011) hypothesized developmental stages to investigate the academic writing of Thai
and native English university students by comparing their argumentative English essays as
concerns their usage of noun modification. Prenominal modifiers and postnominal modifiers
were identified and coded manually. It was found that both groups of writers heavily relied on
attributive adjectives, nouns as premodifiers and prepositional phrases as postmodifiers, and
there were no significant differences in the use of prenominal modifiers between both groups of
students for the most part. The most significant differences between both datasets were in the use
of prepositional phrases with abstract meanings and multiple prepositional phrases as
postmodifiers. These are hypothesized to be acquired at later developmental stages and were
more frequently used by native English university students than Thai university students. The
findings of this study may contribute to greater insights into the nature of noun phrase
complexity used by Thai undergraduates. Pedagogical implications based on the findings are also
provided.

Keywords: noun modification, premodifiers, postmodifiers, phrasal complexity, academic

writing

Introduction
Effective written communication is among the skills essential for students in aiding them to
advance within their careers in the 21st-century workplace (Wagner, 2010). A survey by the
Association of American Colleges and Universities (2008) also revealed that writing is one of the
most addressed skills in college and university study. For learners who learn English as a foreign
language (EFL), including a number of Thai learners, writing is regarded as one of the most
problematic skills. Writing in English is not easy, and writing in an academic context or within
any professional communities is even more challenging.
Several studies have shown that the more learners’ English proficiency develops, they
tend to rely more on economical and concise but complex features at the phrasal level. Biber and
Gray (2011) analyzed several corpora and suggested that nominal structures with phrasal
modification became much more prevalent in academic research writing compared to
conversation or other written registers such as fiction and news reports. Later, Parkinson and
Musgrave (2014) found that the writing of a more proficient group of writers was greater in
similarity to the published academic articles in the use of higher-level types of noun modifiers,
whereas such modifiers were not prevalent in the writing of less proficient writers. Recently, a
comparative study of Wang and Slater (2016) suggested that one key difference which

71
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

distinguished the writing of EFL Chinese students from that of more proficient writers was the
use of complex nominal structures.
In Thailand, very few studies have explored the issues concerning the grammatical
complexity of Thai learners of English. Chuenchaichon (2014) reviewed EFL writing research
conducted in Thailand from 2004 to 2013 and found that various writing research (e.g. L2
writing errors, writing assessment, writing feedback, coherence in writing, online writing/new
technology and writing, genre-based writing instruction, approaches to teaching writing, written
discourse analysis and learning strategies) in different settings (e.g. schools and universities) was
conducted. Of the 48 studies reviewed, L2 writing errors and written discourse analysis and more
recently, online writing/new technology and writing were studied the most. However, only one
study (Biber, Gray & Poonpon, 2011) addressed complexity in L2 writing, with the purpose of
challenging the use of T-units and clausal subordination as writing development measures and
proposing their own hypothesized developmental stages for measuring writing complexity.
Nevertheless, complexity in the writing development of Thai learners, which this present study
attempts to investigate, is still evidently under-researched.
Staples, Egbert, Biber, and Gray (2016) asserted that the university years are when
development of phrasal complexity becomes most obvious, even for native English writers. A
body of research has observed L2 academic writing development (see Ansarifar, Shahriari &
Pishghadam, 2018; Lu, 2011; Parkinson & Musgrave, 2014), but none has focused on EFL Thai
students. Ai and Lu (2013) investigated the differences between syntactic sophistication in the
writing of Chinese learners and that of native speakers of English. They suggested that future
studies examine students of other L1 backgrounds to better understand the influence of L1 on L2
syntactic development as an L1 background can potentially affect L2 syntactic development.
Adopting Biber et al.’s (2011) hypothesized developmental stages (see Figure 1), this
study aims to investigate noun phrase complexity in the argumentative essay writing of Thai
university students compared with that of native speakers of English (NSs). It also aims to
provide insights into the patterns of noun modifiers used by Thai undergraduates compared with
those used by NSs. The fact that NSs are used as a benchmark for higher proficiency, is another
point that sets this study apart from some recent studies in which the researchers compared the
written work of writers of clearly different levels of study such as comparing EAP students’
argumentative essays to MA students’ assignment papers (Parkinson & Musgrave, 2014) or
comparing MA students’ abstracts to those of PhD students (Ansarifar et al., 2018). Unlike
previous research, this current study compares the writing of second-year Thai undergraduates
studying in Thailand, whose letter grades of the EAP writing course range from B, B+ to A, with
the writing of NSs whose grades and proficiency are unknown. Based on Biber et al.’s (2011)
hypothesized developmental stages, writers who are deemed good or excellent are expected to
exhibit prevalent use of phrasal features of the later stages.
Considering the aforementioned purpose, this study attempts to answer the following
research questions:

1. What are the most and the least prevalent noun modifiers used in the argumentative essays written
by Thai university students and native English university students?
2. Are there any significant differences in terms of noun modifiers between the argumentative
essays written by Thai university students and those written by native English university
students?

72
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

Literature Review
Writing quality is one of the indices that differentiates students of different proficiency, and
syntactic complexity is one of the components used in determining language development (Wang
& Stater, 2016). Syntactic complexity refers to “the range of forms that surface in language
production and the degree of sophistication of such forms” (Ortega, 2003, p.492), and is one of
the essential indices to measure learner proficiency and linguistic development. At university,
academic writing skill development is regarded as a major concern as students should be able to
elaborate their ideas in forms of grammatically-complex discourse (Staples et al., 2016).
Issues about syntactic complexity and L2 writing have been widely discussed in
literature. Silva (1993) highlighted that a stark contrast was evident between native speakers’ and
non-native speakers’ written work regarding their fluency, accuracy and syntactic complexity.
Hinkel (2003) investigated academic essays written by non-native and native English speakers
studying in universities in the U.S. and discovered that the non-native group tended to over-rely
on simple sentence structures. More recent studies (e.g. Biber & Gray, 2011; Biber et al., 2011;
Rimmer 2006; Taguchi, Crawford, & Wetzel, 2013) have pointed out that good quality writing is
characterized by more complex language at the phrasal level such as noun modifiers, especially
prepositional phrases as postmodifiers.

Noun modification and its role in writing complexity

Grammatically speaking, a simple noun phrase is a phrase with a determiner and head noun, and
additions to this simple structure make them become more grammatically complex (Biber &
Gray, 2016). Biber (1988) claimed that when compared to speech, which is typically
characterized by clausal subordination, phrases are more common features in formal writing.
Also, attributive adjectives and prepositional phrases which pre- or postmodify the head noun are
typically found in written registers.
Below is how Biber, Johansson, Leech, Conrad, and Finegan (1999, p. 574) presented the
basic noun phrase structure. The structure can potentially be more complex when multiple pre-
or postmodifiers are added.

Determiner + (premodification) + head noun + (postmodification and

complementation)
e.g. a (big golden) box (that I bought)
The fact (that I haven’t thought of)

There are various types of noun pre- and postmodifiers. Adjectives (e.g. intense
marketing), participles (e.g. no fixed hours), nouns (e.g. a web page), and possessive nouns (e.g.
people’s attention) are categorized as noun premodifiers while relative clauses (e.g. the one that
reaches customer first), ing-clauses (e.g. the majority of people living in big cities), ed-clauses
(e.g. the aspects presented above), prepositional phrases (e.g. commercials on social media),
noun phrases in apposition (e.g. All Seasons – a clothing retailer), and complement clauses (e.g.
the fact that our sales keep going down) are classified as noun postmodifiers.
The fact that information in academic prose is usually required to be presented with high
density has led writers to greatly rely more on structures of noun phrases (Ni, 2003) with
different levels of complexity depending on the number of modifiers those noun phrases carry.
Biber and Gray (2010) note that greater reliance on phrasal expressions results from the need for

73
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

denser textual information that helps texts to be more economical and facilitates faster reading.
This supports Halliday’s (1993) claim that text becomes more compressed when nominal
expressions are used rather than clausal structures. He further commented that when compared to
speaking, writing is phrasally more complex with embedded nominal structures, but clausally
simpler. Wolfe-Quintero, Inagaki, and Kim (1998) also found that the writing of those with
higher English proficiency is more likely to display more conciseness with greater use of more
reduced forms such as the use of prepositional phrases instead of a relative clause or the use of
infinitive or a gerund instead of a nominal clause. The fact that students encounter academic
texts written in compressed discourse styles, their success, to a certain extent, depends on the
ability to comprehend and produce the language that conforms to the norm of the academic or
professional world (Biber & Gray, 2016). Therefore, it can be useful for students to harness the
knowledge of such grammatical structures so as to properly incorporate grammatical complexity
into their writing.

Writing complexity measurement

Research into the syntactic complexity of the language produced by native and non-native
English writers has been conducted. Several studies (e.g. Henry, 1996; Lu, 2011; Ortega, 2003)
adopted different complexity measurements such as sentence complexity, length of production
unit, and varieties of sentence structures and their frequency. Among these measurements, the
analysis of T-unit has been extensively used for measuring the syntactic maturity of written and
spoken language and has been well-received as a reliable measure of overall syntactic
complexity (Gaies, 1980). Coined by Hunt (1965), T-unit refers to a main clause including all
subordinate clauses and other structures that might be attached to it. The growth of syntactic
maturity is shown through the length of the T-units one formulates. T-unit-based measures have
been applied in subsequent studies (e.g. Ortega, 2003; Wolfe-Quintero et al., 1998) and have
been regarded as well-established measures of first and second language development and
acquisition.
Wolfe-Quintero et al. (1998) examined 39 studies on L2 writing, and also highlighted the
importance of clauses and T-unit to language development. They asserted that T-unit complexity
ratio (clauses per T-unit, or C/T) generally grew as proficiency level increased. Nevertheless, the
use of T-unit-based measures has been challenged by many researchers. Several studies (e.g.
Bardovi-Harlig, 1992; Rimmer 2006) showed that learners with high proficiency did not
necessarily produce longer T-units and clauses. Rimmer (2006) suggested that noun
postmodifiers and some other phrasal features be taken into consideration when measuring
syntactic complexity.
A linguistic description of academic prose is comprehensively provided in the Longman
Grammar of Spoken and Written English (LGSWE; Biber et al, 1999). With a large corpus,
around 5 million words from both spoken and written English, the LGWSE can identify the
patterns of grammatical features commonly used in conversations, newspapers, fictions, and
academic texts. According to the corpus analysis reported in the LGSWE, while approximately
150,000 nouns per million words are used in conversation, about twice more (approximately
300,000 nouns per million words) nouns are used in academic prose. The findings from the
LGSWE also illustrate that verbs and related word classes such as adverbs and adverbials are
more prevalent in conversation, making clauses more prevalent than phrases in this register
accordingly. In contrast, nouns and their associated word classes such as adjectives and
prepositions prevail in academic prose.

74
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

A subsequent study by Biber and Gray (2010) also complements the above-mentioned
findings revealing that any types of written academic texts, even faculty web pages, show more
reliance on condensed, less elaborated structures, and that means university students are exposed
to this style of discourse typically everywhere, not necessarily limited to formal academic or
research writing. Their findings have challenged the stereotype of grammatical complexity in
writing which usually places emphasis on more elaborate and explicit clausal structures. They
also encouraged further studies to shift their focus to grammatical features that have proven to be
ubiquitous in academic prose but overlooked, for instance, noun phrases and their pre-
/postmodifiers such as attributive adjectives, prepositional phrases, and several others. These
features are uniquely common in written register and deserve greater attention.
Biber and Gray (2011) claimed that academic writing, which is regarded as an advanced
written register, shows a prevalence of non-clausal phrases and the highly complex structure of
noun phrases. Findings from corpus-based research and the belief that T-unit analysis alone does
not comprehensively reflect writing complexity led Biber et al. (2011) to question the validity of
T-unit based indices to measure syntactic complexity and propose hypothesized developmental
stages that cast new light on the study of complexity in writing development. They conducted a
comparison between the use of 28 features found in conversation and academic writing. This
study confirmed that academic writing is clausally less complex, but that it showed a higher
degree of noun phrase complexity.
More recent studies (e.g. Ansarifar et al., 2018; Parkinson & Musgrave, 2014; Taguchi et
al., 2013) found that noun phrase modifiers including attributive adjectives as premodifiers and
prepositional phrases as postmodifiers of noun phrases were indicators of writing quality. In their
corpus-based research findings, Biber and Gray (2016) emphasized again that some grammatical
structures, such as dependent clauses, were not prevalent in written registers, whereas phrasal
structures that previous research did not pay much attention to such as attributive adjectives,
noun as nominal premodifiers, prepositional phrases as nominal postmodifiers, and appositive
noun phrases were more frequently used. Staples et al. (2016) examined the writing development
of L1 English university students, from first-year undergraduate to graduate level, and found an
increase in the use of phrasal features but a decrease in that of clausal features as the students’
academic level grew.
Figure 1 shows the developmental stages for noun modification. This index starts from
stage 2, which is regarded as an intermediate stage (Biber et al., 2011, p. 30), and progresses to
stage 5, which is the most advanced stage.

75
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

Figure 1. Hypothesized developmental stages for noun phrase modification proposed by Biber et
al. (2011)

Methodology
Data sources
Thirty-nine argumentative essays (totaling 12,310 words) on social media marketing written in
English by Thai second-year students at a university in Thailand were obtained from students
who received the letter grades of B, B+, and A from their English for Academic Purposes (EAP)
writing course. Informed consent forms permitting the researcher to use their essays were
collected. As a required task of the course, each student wrote an argumentative essay to respond
to the question of whether a company should apply social media marketing to its business. This

76
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

genre of academic writing was chosen as it is believed to require high order thinking. Also, in
order to make their logical thinking evident through argumentation, a wide variety of language
structures, including noun phrases and a variety of noun modifiers, are expected to be used.
Before completing this task, the students had learnt about the topic in class through reading
exercises and group discussion, and had about one week to research the topic. Essay writing was
carried out in class and timed for one hour.
Another set of data was derived from the Louvain Corpus of Native English Essays
(LOCNESS), consisting of British pupils’ A level compositions, and British and American
university students’ essays. This corpus contains 434 essays (totaling 324,304 words). Only
argumentative essays of British and American university students were used to increase the
variety of Englishes. Crawford and Csomay (2015) suggested that two corpora of a similar
number of words, rather than similar number of texts, are preferable when comparing the
frequency of features of interest. Following their suggestion, this study used 28 essays (totaling
12,694 words) randomly selected from LOCNESS. The British essays were not rigidly timed
while the American ones were. The topics of the chosen essays range from international politics
and social issues to science and technology.
Table 1 shows the total number of essays, the mean length of essays, and the total number
of words. It can be seen that the essays of NS students are generally longer, but the total numbers
of words in both corpora are similar.

Table 1
Details of the corpora
Number of essays Mean length of essays Total number of words

Thai 39 315.64 12,310

NS 28 453.36 12,694

It is necessary to note that despite the attempt to make the data between both groups
homogeneous, not all variables could be controlled. First of all, while one set of essays were
produced by Thai students in the same faculty and university, the other set of essays were written
by British and American students from different academic disciplines and universities.
Additionally, information regarding the proficiency and letter grades of British and American
students is not available. However, since the LOCNESS subcorpora contain numerous essays of
the same genre and similar nature, that is timed argumentative essays, 28 essays were randomly
selected and used to compare with those of the Thai group in this study.

Data coding and analysis

Since the amount of data is relatively small, this study relied on manual coding. Biber and Gray
(2011) conceded that some features such as prepositional phrases and appositive noun phrases
require manual analysis and that the automated taggers may not yield complete accuracy. In this
study, any target features that contained grammatical errors affecting intended meanings or
intelligibility were not counted, whereas those with minor grammatical errors that did not hinder
comprehension such as misspelling, articles or prepositions in some collocations such as
marketing has a impact for the business (for instead of on) were included. Also, some key words
or phrases from the prompt of the Thai undergraduates’ essays such as social media and social

77
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

media marketing when they do not modify another noun were not counted in order to avoid their
influence on the results since they accounted for almost 350 hits among the 39 Thai
undergraduates’ essays analyzed.
The grammatical features of noun modifiers are the primary focus of the current study.
The developmental stages of grammatical features proposed by Biber et al. (2011, p. 30) were
adopted as a measure to identify the complexity of the features that appeared in the essays. The
frequency of noun premodifiers (common/less attributive adjectives, participial adjectives, nouns
as premodifiers, possessive nouns, and multiple premodifiers) and noun postmodifiers (relative
clauses, prepositional phrases with concrete/abstract meanings, participial clauses as
postmodifiers, noun complement clauses, appositive noun phrases, and multiple prepositional
phrases) were identified and recorded.
Biber et al. (1999) analyzed corpus data and distinguished between common and less
common attributive adjectives. Common attributive adjectives refer to adjectives that occur more
than 200 times per million words and are mostly monosyllabic and semantically simple such as
good, nice, whole, social, important, and so forth. Those not meeting these criteria are considered
less common adjectives. Biber et al. (1999) also reported that less common attributive adjectives
occur more frequently in academic writing. However, when Biber et al. (2011) proposed their
developmental stages, they did not treat common and less common attributive adjectives as
different features. In a recent study by Parkinson and Musgrave (2014), these two types of
adjectives were coded separately. This present study followed suit and used the list of common
attributive adjectives suggested by Biber et al. (1999, p. 512) to help distinguish common
attributive adjectives from less common ones and to see whether the findings of this present
study would support their findings.
Additionally, Biber et al. (2011) did not separate participial premodifiers from attributive
adjectives; however, Parkinson and Musgrave (2014) believed that participial premodifiers are
potentially acquired later and presumably placed them in stage 3. Therefore, participial
premodifiers are considered a separate feature in more recent studies (e.g. Ansarifar et al., 2018;
Parkinson & Musgrave, 2014) as well as in this present study. Also, based on Biber and Gray
(2016), prepositional phrases with concrete/locative meanings refer to those representing
materials, physical forms, or places (e.g. functions on electronic devices, employees in the
meeting room, and users from China), while prepositional phrases with abstract meanings
include those representing concepts, relations, or those without concrete existence (e.g.
information for analysis and progress in integration).
After the researcher coded every essay, 10 percent of the essays were randomly selected
and coded manually by the second coder to ensure that the coding was consistent and reliable.
The second coder was a non-native speaker who is a proficient user of English and has taught
EAP courses for undergraduate students for nine years. He was trained by the researcher to code
the data based on Biber et al.’s (2011) developmental stages and the coding guidelines provided
by the researcher. The percentage of agreement and correlation were calculated to deal with the
consensual and consistent aspects of the inter-coder reliability. Then, Fisher's exact test was
computed to identify whether there were significant differences in the use of noun modifiers
between both groups. Fisher's exact test was chosen since it works especially well for small
samples. The Statistical Package for the Social Sciences (SPSS, Version 20) was utilized to
compute the correlation and to run Fisher’s exact test. The result of the percentage of agreement

78
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

was 94.12% whereas that of the correlation for all the features coded was above .90. For both
procedures, a score over .90 is considered excellent (Portney & Watkins, 1993; Salkind, 2011).

Results
Figure 2 illustrates the frequency of each type of noun modifier and the percentage it accounts
for of all noun modifiers from each corpus, together with the result of a two-sided Fisher’s exact
test. When the p-value is < .05, the difference is significant. Regarding descriptive statistics, see
Appendix A for details.

Figure 2. Noun modifiers in Thai and NS argumentative essays

As can be seen in Figure 2, it was found in the Thai dataset that the most common
grammatical feature was less common attributive adjectives (20.10%), followed by nouns as
premodifiers (18.56%). Those used in moderation were of phrases as postmodifiers with abstract

79
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

meanings (10.05%), prepositions other than of as postmodifiers with abstract meanings (8.45%),
common attributive adjectives (8.32%), prepositions other than of as postmodifiers with
concrete/locative meanings (6.78%), and multiple premodifiers (5.67%). The findings also
revealed that relative clauses (4.50%), of phrases as premodifiers with concrete/locative
meanings (4.01%), possessive nouns as premodifiers (3.95%), multiple prepositional phrases as
postmodifiers with levels of embedding (3.45%), of + ing (2.53%), participial premodifiers
(1.54%), that + noun complement clauses (1.29%) were used slightly. Finally, -ed as
postmodifiers (0.31%), -ing as postmodifiers (0.25%), appositive noun phrases as postmodifiers
(0.25%), were the least frequently used features.
In contrast, the NS dataset showed the highest reliance on less common attributive
adjectives (21.88%), of phrases as postmodifiers with abstract meanings (15.23%), and
prepositions other than of as postmodifiers with abstract meanings (11.02%) whereas common
attributive adjectives (8.96%), relative clauses (6.22%), prepositions other than of as
postmodifiers with concrete/locative meanings (6.11%), multiple prepositional phrases as
postmodifiers with levels of embedding (6.01%), and nouns as premodifiers (5.90%) were
moderately used. This group also slightly used multiple premodifiers (4.96%), of phrases as
premodifiers with concrete/locative meanings (4.85%), participial premodifiers (2.21%), -ing as
postmodifiers (1.53%), possessive nouns as premodifiers (1.48%), and -ed as postmodifiers
(1.37%). Lastly, that + noun complement clauses (0.79%), appositive noun phrases as
postmodifiers (0.79%), and of + ing (0.69%) were the least preferred features among the NS
group.

Research question 1: What are the most and the least prevalent noun modifiers used in the
argumentative essays written by Thai university students and native English university students?
Overall, the most prevalent features of noun premodifiers used by both groups were
attributive adjectives [a good idea (common), a controversial topic (less common)] and nouns as
premodifiers (sales promotion), whereas the most frequently used postmodifiers were
prepositional phrases with both concrete/locative and abstract meanings [the cost of digital
marketing (abstract), comments about your products (concrete), and the settings on a Facebook
page (locative)].
On the other hand, -ed and -ing participles as postmodifiers (campaigns promoted on
Facebook, customers using social media) were used very slightly by the Thai students, while of
+ ing (the idea of having a single market), and that + noun complement clauses (the fact that it is
an island) were the least prevalent features found in the NS essays. Finally, appositive noun
phrases (All Seasons, a clothing store) were one of the least prevalent features used by both Thai
and NS students. As shown in Figure 2, all of these postmodifiers accounted for less than 1% of
the total noun modifiers found in the essays of each group.

Research question 2: Are there any significant differences in terms of noun modifiers between
the argumentative essays written by Thai university students and those written by native English
university students?
As can be seen from Figure 2, 10 out of 17 features show statistically significant
differences including relative clauses, nouns as premodifiers, of phrases as postmodifiers
(concrete/locative meanings), -ed participle as postmodifiers, -ing participle as postmodifiers, of
phrases as postmodifiers (abstract meanings), prepositions other than of as postmodifiers

80
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

(abstract meanings), of + ing, appositive noun phrases, and multiple prepositional phrases as
postmodifiers with levels of embedding. Interestingly, all of these features, except nouns as
premodifiers, are noun postmodifiers.

Discussion

As regards research question 1, it was found that Thai and NS students most frequently used
attributive adjectives and nouns as premodifiers whereas prepositional phrases were the most
prevalent postmodifiers. This supports Biber et al.’s (1999) claim that attributive adjectives,
nouns as premodifiers and prepositional phrases as postmodifiers characterize academic prose,
and confirms some previous studies (e.g. Ansarifar et al, 2018; Parkinson & Musgrave, 2014)
which found that the most prevalent types of noun premodifiers were attributive adjectives and
nouns while that of noun postmodifiers is prepositional phrases. Staples et al. (2016) also noted
that in both L1 and L2 academic writing, phrasal features such as noun phrases with attributive
adjectives and nouns as premodifiers are considered significant components of academic writing.
As for the least frequently used features, the findings of this study regarding the use of -
ed and -ing participles as postmodifiers (stage 4) support Biber et al.’s (1999) claim that full
relative clauses are a more preferred feature than participles. Additionally, according to Biber et
al.’s (2011) hypothesized developmental stages, relative clauses are placed in stage 3, which
implies that they are acquired earlier than participles. Therefore, both Thai and NS students
might be more inclined to use relative clauses than -ed or -ing participles in this study. In the
case of noun complement clauses (of + ing and that + noun complement clauses), despite being
strongly associated with academic writing, these modifiers are not particularly prevalent by
themselves (Biber & Gray, 2016). Therefore, it is not surprising that they were slightly used by
both groups of students. Moreover, although Biber and Gray (2011) highlighted the growing
importance of appositive noun phrases in academic writing, this present study found these
phrasal devices to be among the least used features. This finding, however, supports Biber et
al.’s (2011) placement of appositive noun phrases in stage 5, the most advanced stage. Overall,
the findings of this current study are in line with those of Parkinson and Musgrave’s (2014) in
that the aforementioned features were least preferred in EAP and MA writing.
To discuss research question 2, noun modifiers were grouped into pre- and postmodifiers.
Regarding noun premodifiers, nouns as premodifiers is the only one among the other
premodifiers that shows statistically significant difference (Thai = 18.56%; NS = 5.90%) in this
study. Unexpectedly, Thai students used these phrasal devices more than the NSs did. This
finding proves inconsistent with the previous studies of Parkinson and Musgrave (2014) and
Ansarifar et al. (2018) which suggested that writers with higher proficiency or more experience
tend to use premodifying nouns more than less proficient writers. A closer look at the data of
Thai writers revealed a possible explanation for this inconsistency. The data suggested that many
Thai students took some phrases directly the reading sources. These recurring phrases included:

social media platforms = 14 counts

social media site(s) = 8 counts
social media users = 5 counts
social media + other nouns = 8 counts
social network + other nouns = 6 counts

81
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

internet marketing = 4 counts

(social media is considered a single entity, so it is regarded as a noun in this study.)

The fact that marketing was the topic area of the essay also influenced Thai students’
repeated use of the word as noun premodifier such as:

marketing strategy(ies) = 10 counts

marketing channel(s) = 9 counts
marketing tool(s) = 8 counts
marketing + other nouns = 18 counts

Assumed to be acquired later than attributive adjectives (stage 2), participial premodifiers
were considered a separate feature. Nevertheless, no significant difference in the use of this
feature was identified in this study. Although the result of this study differs from that of
Parkinson and Musgrave (2014) who found significant distinction across the groups they studied,
it is similar to that of Ansarifar et al. (2018). Possessive nouns as premodifiers (stage 3) was
another feature for which no statistically significant difference was found, which was consistent
with Ansarifar et al. (2018) but inconsistent with Parkinson and Musgrave (2014) whose data
showed significant difference in the use of possessive nouns.
Concerning noun postmodifiers, the Thai dataset showed less reliance on relative clauses
than the NS dataset did (4.50% of Thai modifiers compared to 6.22% of NS modifiers). This
supports the findings of Parkinson and Musgrave (2014) and Ansarifar et al. (2018) who also
found that less proficient writers used relative clauses less than more proficient ones. However,
only Parkinson and Musgrave’s (2014) findings revealed a statistical difference in the use of
relative clauses. In this present study, since the NSs’ essay topics (e.g. British sovereignty,
European integration, feminism, and inventions and discoveries) were more specific, the NS
writers may have been more inclined to include more proper nouns or technical terms in their
writing, thus increasing the chance of using relative clauses by way of further explanation of
such terms. In contrast, Thai students were assigned a general topic – marketing; thus, the
chances that they would use relative clauses could have been fewer. Some examples from the
corpora are as follows:

Thai: NS:
people who shop in the mall feminism, which has had a significant impact on…

anyone who visits the website a Federal Union which would ensure decisions
where possible

many competitors who also use this the Women's Suffrage Movement which resulted in
method the women of America being recognized as full
citizens

82
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

Another problem which may be the most This view of the legislative supremacy was expressed
serious by Dicey who wrote that…

Statistically significant differences were also observed across the use of all types of
prepositional phrases as postmodifiers, except for prepositions other than of with
concrete/locative meanings. Both datasets showed frequent use of prepositional phrases, with
higher proportion of abstract meanings (stage 4) than concrete/locative meanings (stage 3). This
corroborates Biber and Gray’s (2011) study which found that about 60 percent of occurrences of
prepositional phrases as postmodifiers in 20th-century written academic prose carry abstract
meanings. Nonetheless, this study contrasts with Ansarifar et al.’s (2018) in that almost every
prepositional phrase feature they analyzed showed no statistically significant difference while
Parkinson and Musgrave (2014) reported such significant distinction only in the use of
propositional phrases with prepositions other than of (abstract meanings). Overall, Thai students
used all types of prepositional phrases significantly less than NSs did.
Placed at stage 4, there were statistically significant differences in the use of -ed and -ing
participles as postmodifiers between Thais (0.31%, 0.25%) and NSs (1.53%). Similar to what
Parkinson and Musgrave (2014) and Ansarifar et al. (2018) found, these postmodifying features
were not the main features of both sets of data considering the small proportions found. This
probably stemmed from higher reliance on stage-3 relative clauses than these stage-4 participles.
Interestingly, Thai students used stage-5 noun complement clauses, including preposition
+ nonfinite complement clauses (of + ing) and complement clauses controlled by nouns (that +
noun complement clauses), more frequently than the NSs did. However, statistically significant
difference was only observed in of + ing (Thai = 2.53%; NS = 0.79%).
The use of that + noun complement clauses is also worth mentioning as despite no
statistically significant difference was found, Thai students unexpectedly used this stage-5
feature more frequently than NSs (Thai = 1.29%; NS = 0.79%). These unexpected results might
result from the influence of the essay prompts and some stock phrases in Thai students’ course
materials, or these students might have been prepared by their teachers to incorporate what they
learned in their writing. However, the findings of this study corroborate the recent studies of
Parkinson and Musgrave (2014) and Ansarifar et al. (2018) in that no statistically significant
difference was found in both. The examples of repeatedly used noun complement clauses found
in the Thai dataset include:

the idea of + -ing = 18 counts

danger of + -ing = 8 counts
I am for/against the idea that…. = 18 counts
It is my belief that … = 5 counts
the fact that … = 3 counts

Appositive noun phrases also deserve attention. Despite being infrequently used across
both datasets, statistically significant difference was found in this study. This is also the case for
Parkinson and Musgrave’s (2014) study but not for Ansarifar et al.’s (2018). The NS students
used this stage-5 feature more than their Thai counterparts (Thai 0.25%; NS = 0.79%). This
possibly resulted from the influence of essay topic areas. As mentioned earlier, the NSs’ essay

83
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

topics were more specific. Thus, the use of proper nouns or technical terms can be anticipated as
appositive noun phrases are sometimes necessary in order to “provide an explanatory gloss to a
technical reference or name of some entity” (Biber et al., 1999, p. 639).
As the last feature of the developmental stages, multiple prepositional phrases as noun
postmodifers are expected to be more prevalent in the NS dataset than in the Thai dataset. A
statistically significant difference between both groups in the use of this feature was found (Thai
= 3.45%; NS = 6.01%). Similarly, Ansarifar et al. (2018) found that expert writers used this
stage-5 feature more than the other groups of writers they studied with statistically significant
differences observed across all groups. The findings of this present study also confirm Biber et
al.’s (2011) hypothesis that more proficient writers would rely more on the final-stage feature
than less proficient ones.
Additionally, a closer look at the essays of both groups enabled the researcher to see that
the strings of noun phrases with multiple prepositional phrases as postmodifiers produced by the
NSs were often longer with more modifiers than those produced by Thai students although the
number of prepositional phrases per case did not differ greatly. In other words, not only did the
NSs show greater use of this stage-5 feature, but the cases found in their essays also exhibited a
higher level of complexity compared to those found in the essays of Thai students as the
prepositional phrases found in their noun phrase strings usually carry fewer modifiers within
them. The following are cases taken from the corpora showing multiple prepositional phrases.

Thai: NS:
the advertising cost for promoting on offline greatest co-operation and joint action in the
channels areas of foreign, social and environmental
policy

campaigns from the company on social media the expansion of the British market into all the
other markets of Europe and vice versa

rapidness of the flow of information via social more participation in the election of members
media of the European Parliament

customers’ opinions on the products on social favorable trading condition with the USA and
media with the countries of the Commonwealth

high competition among brands on social the dismantling of the Berlin Wall and
media hundreds of other smaller symbols of freedom

By and large, the writing of both datasets was fairly similar in the use of noun
premodifiers (stages 2-3). The only type of noun premodifier showing statistically significant
difference was nouns as premodifiers. Yet, the contrast between both datasets became noticeable
in the use of noun postmodifiers (stages 4-5) in that fewer of these stage-4 and -5 features
appeared in the essays of Thai students. In addition, the use of almost all types of postmodifiers

84
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

by both groups of writers differed significantly, except prepositions other than of

(concrete/locative meanings) and that + noun complement clauses.
Considering the above discussion and the results shown in Figure 2, although
argumentative English essays written by Thai students fairly approximate those written by NSs
in the use of early-stage features, they do not approximate those written by NSs in the use of
higher-stage features (stages 4-5). In other words, the writing of Thai university students was not
as complex as their NS counterparts.

Conclusion
This study compares noun modification in the argumentative English essays of Thai and NS
university students. The developmental stages for noun phrase modification proposed by Biber et
al. (2011) were empirically tested as suggested by Biber et al. (2011) themselves. The findings of
this present study support the idea of the hypothesized developmental stages of Biber et al.
(2011) in that presumably less proficient writers (i.e. Thais) rely more heavily on the features of
earlier stages than more proficient writers (i.e. NSs). This suggests that focusing on each of the
higher-level features less preferred among Thai students can be particularly helpful in developing
writing complexity.
The findings of this study can help English language instructors, especially those who
teach academic writing, gain more insights into noun modification features and encourage them
to identify the complexity levels of their students’ work using Biber et al.’s (2011) hypothesized
developmental stages. These insights can potentially enable the instructors to indicate what needs
to be done in order to improve academic writing skills among Thai students and to fill the void
and increase syntactic complexity in writing. It is hoped that the results of this study will bring
about better material design with an additional focus on grammatical features such as noun
postmodifiers, especially prepositional phrases postmodifying nouns, which have long been
overlooked in academic writing courses. Materials or tasks that help draw students’ attention to
complex phrasal features in academic prose can be incorporated to allow students to produce
written work that meets an academic writing style that is internationally acceptable.
This study also contains some limitations. Firstly, the size of the dataset is relatively
small and the generalizability of the study’s findings might be somewhat limited. Future research
can be conducted with larger data size. Secondly, as mentioned earlier, there are some
differences between the two datasets that the researcher was not able to control, such as the essay
topics or fields of study of the NS group. Although all of the writing samples are of the same
genre, it is undeniable that differences in essay topics and students’ academic discipline may, to
some extent, affect the use of certain grammatical features. Future studies could take the issue of
topic differences into account. Another uncontrollable factor was the nature of the writing
course. The argumentative essay writing of the Thai students was timed due to the design of the
course. In contrast, the NS essays were a combination of timed and not-rigidly-timed essays.
This might not reflect authentic, real-world writing which allows writers sufficient time to gather
and put together information, as well as time to revise and polish their writing. Future studies can
take this issue into consideration to ideate better research design.

About the Author:

Narisa Jitpraneechai is a lecturer in the Division of English for Business at Chulalongkorn
University Language Institute, Bangkok Thailand. She is currently teaching various EAP courses

85
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

for undergraduate students. Her research interests include academic writing, written business
communication, corpus linguistics, and motivation in language teaching and learning.

References:
Ai, H., & Lu, X. (2013). A corpus‐ based comparison of syntactic complexity in NNS and
NS university students' writing. In A. Díaz‐ Negrillo, N. Ballier, & P. Thompson,
(eds.), Automatic treatment and analysis of learner corpus data (pp. 249–264).
Amsterdam: John Benjamins.
Ansarifar, A., Shahriari, H. & Pishghadam, R. (2018). Phrasal complexity in academic
writing: A comparison of abstracts written by graduate students and expert
writers in applied linguistics. Journal of English for Academic Purpose, 31,
58-71.
Association of American Colleges and Universities. (2008). How should colleges assess
and improve student learning? Employers' views on the accountability challenge.
Retrieved from http://www.aacu.org/leap/documents/2008_Business_Leader_Poll.pdf
Bardovi-Harlig, K. (1992). A second look at T-unit analysis: Reconsidering the sentence.
TESOL Quarterly, 26, 390–395.
Biber, D. (1988). Variation across speech and writing. Cambridge, England: Cambridge
University Press.
Biber, D., & Gray, B. (2010). Challenging stereotypes about academic writing:
Complexity, elaboration, explicitness. Journal of English for Academic Purposes,
9, 2–20.
Biber, D., & Gray, B. (2011). Grammatical change in the noun phrase: The influence of
written language use. English Language and Linguistics, 15(2), 223-250.
Biber, D. & Gray, B (2016). Grammatical Complexity in Academic Writing: Linguistic
Change in Writing. Cambridge, England: Cambridge University Press.
Biber, D., Gray, B. & Poonpon, K. (2011). Should we use characteristic of conversation
to measure grammatical complexity in L2 writing development?. TESOL
Quarterly, 45, 5-35.
Biber, D., Johansson, S., Leech, G., Conrad, S., & Finegan, E. (1999). Longman
grammar of spoken and written English. Edinburgh: Pearson Education Limited.
Chuenchaichon, Y. (2014). A review of EFL writing research studies in Thailand in the
past 10 years. Journal of Humanities, 11(1), 13-30.
Crawford, W. J., & Csomay, E. (2016). Doing corpus linguistics. New York: Routledge.
Gaies, S. J. (1980). T-unit Analysis in Second Language English Research: Applications,
Problems, and Limitations. TESOL Quarterly, 14, 53-60.
Halliday, M. A. K. (1993). Some grammatical problem in scientific English. In M. A. K.
Halliday, & J.R. Martin (Eds.), Writing Science (pp. 69-85). London: The Falmer Press.
Henry, K. (1996). Early L2 writing development: A study of autobiographical essays by
university-level students of Russian. The Modern Language Journal, 80 (3),
309-326.
Hinkel, G. (2003). Simplicity without elegance: Features of sentences in L1 and L2
academic texts. TESOL Quarterly, 37, 275-301.

86
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

Hunt, K. (1965). Grammatical Structures Written at Three Grade Levels. National

Council of Teachers of English Research report No. 3. Champaign, IL, USA:
NCTE.
Lu, X. (2011). A corpus-based evaluation of syntactic complexity measures as indices of
college-level ESL writers’ language development. TESOL Quarterly, 45(1),
36–62.
Ni, Y. (2003). Noun phrases in media texts: A quantificational approach. In J. Aitchison,

& D.M. Lewis (eds.), New media language (pp.159-168). New York: Routledge.
Ortega, L. (2003). Syntactic Complexity Measures and their Relationship to L2
Proficiency: A Research Synthesis of College-level L2 Writing. Applied
Linguistics, 24(4), 492-518. doi: 10.1093/applin/24.4.492
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 (2), 48–59.
Portney, L. G., & Watkins, M. P. (1993). Foundations of clinical research: Applications
to practice. Norwalk, CT: Appleton &amp; Lange.
Rimmer, W. (2006). Measuring grammatical complexity: The Gordian knot. Language
Testing, 23, 497–519.
Salkind, N. J. (2011). Statistics for people who (think they) hate statistics (4 th ed.).
Thousand Oaks, CA: Sage.
Silva, T. (1993). Toward an understanding of the distinct nature of L2 writing: The ESL
research and its implications. TESOL Quarterly, 27, 657-677.
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. Written Communication, 33(2), 149-183.
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. doi: 10.1002/tesq.91
Wagner, T. (2010). The global achievement gap: Why even our best schools don’t
teach the new survival skills our children need — and what we can do about
it. New York: Basic Books.
Wang, S., & Slater, T. (2016). Syntactic Complexity of EFL Chinese Students’ Writing.
English Language And Literature Studies, 6(1), 81. doi: 10.5539/ells.v6n1p81
Wolfe-Quintero, K., Inagaki, S., & Kim, H.-Y. (1998). Second language development in
writing: Measures of fluency, accuracy, and complexity. Honolulu, HI: University
of Hawaii Press.

87
 LEARN Journal: Language Education and Acquisition Research Network Journal, Volume 12, Issue 1, January 2019

Appendix A

Descriptive statistics of noun modifiers for Thai and NS groups.

No. of Min Max Mean Std. Dev.
essays

Cmm Adj Thai 39 0 9 3.54 2.73

NS 28 0 16 6.54 4.77
Less Cmm Adj Thai 39 0 20 8.56 4.59
NS 28 3 28 14.93 6.52
Parti Prem Thai 39 0 4 0.95 1.19
NS 28 0 6 2.07 1.96
Rel Cl Thai 39 0 5 1.85 1.53
NS 28 0 11 4.18 2.80
N Prem Thai 39 0 14 7.54 4.25
NS 28 0 13 3.54 3.13
Poss Prem Thai 39 0 9 1.62 2.01
NS 28 0 5 1 1.33
Of Post Conc Thai 39 0 6 1.67 1.46
NS 28 0 8 3.32 2.16
Other Pp Conc Thai 39 0 7 2.82 1.83
NS 28 0 13 4.14 2.77
Ed Part Post Thai 39 0 1 0.13 0.34
NS 28 0 6 0.93 1.22
Ing Part Post Thai 39 0 1 0.1 0.31
NS 28 0 5 1.07 1.12
Multi Prem Thai 39 0 6 2.31 1.76
NS 28 0 8 3.36 2.50
Of Post Abst Thai 39 0 10 4.15 2.24
NS 28 2 22 10.29 4.64
Other Pp Abst Thai 39 0 11 3.51 2.21
NS 28 0 17 7.46 3.91
Of Ing Thai 39 0 3 1.05 0.89
NS 28 0 4 0.46 0.96
That N Comp Thai 39 0 2 0.54 0.68
NS 28 0 2 0.54 0.74
Appos N Post Thai 39 0 1 0.1 0.31
NS 28 0 3 0.5 0.75
Multi Pp Post Thai 39 0 5 1.46 1.10
NS 28 0 9 4.07 2.32

88