Using Computer-Assisted Approach to Media Discourse Analysis on the
Construction of City Image: Chengdu Park City as an Example
Jing Liu
College of International Education, Chengdu Polytechnic,
No.83, Tianyi Street, Hi-tech Zone District, Chengdu, Sichuan, China
Keywords: Computer-Assisted, Corpus Linguistic Approach, News Media Coverage, International Communication
Reports, Discursive Construction, City Image.
Abstract: This study uses a computer-assisted corpus linguistic approach to analyzing news media coverage of Chengdu
Park City for international communication. With a dataset of 60,080 word tokens of foreign-oriented news
articles on the Park City of Chengdu, the study demonstrates how embedded computational techniques in
corpus linguistic software can aid the quantitative analysis with reference to collocations, frequency and
keyness, as well as qualitative analysis based on concordance lines and lexical networks. Taking Chengdu as
an example, this study suggests a framework for adopting computer-powered statistical measures in analyzing
the discursive construction of a city’s image by pinpointing linguistic patterns of datasets, mainly in terms of
collocations, frequencies, key clusters as well as their diachronic development.
1 INTRODUCTION
City image is the sum of beliefs, ideas, and
impressions that people have of a city (Kotler,
Gertner, 2022). In reality, except for a small number
of people who can directly see and feel a city, most
of the international public has to imagine the city
through mass media. The narratives of media thus
play a significant role in agenda setting and framing
the public’s cognition of city image (Ke, Chen, Yu,
2012). The city image, under the scope of
communication studies, has transcended such
physical objects as geography, scenery, landscape to
capture the mental conception held in common by
those who are exposed to the reality constructed by
mass media. The city of Chengdu boasts an affluent
agricultural region and has long been known as the
Land of Abundance. Such an impression is so
ingrained all around the world that drastic changes in
the city image of Chengdu are not discernible to the
general public. In 2018, Chengdu became the first
place where the Park City concept was officially
mentioned in China, and after more than three years
of exploration and practice, the Park City has become
the most distinctive cultural label of the city of
Chengdu. Despite Chengdu’s great strides in building
itself into a Park City, the image of which receives
little recognition, if not at all, from the international
community. Against such a backdrop, the presence of
Park City-related content in Chinese international
communication reports for foreign-oriented publicity
has increased over the past three years. According to
Fairclough, discourses not only represent the world as
it is (or rather is seen to be), they are also projective,
imaginaries, representing possible worlds which are
different from the actual world, and tied into projects
to change the world in particular directions
(Fairclough 2003). In this sense, international
communication reports on Park City not only reflect
but also construct the image of Chengdu, making it
visible to the international community. No studies so
far have probed into the discursive construction of
Chengdu Park City in media reports. Given that, this
study aims to analyze how Chengdu has attempted to
discursively construct itself as a Park City and the
effects of which on the construction of Chengdu Park
City per se.
208
Liu, J.
Using Computer-Assisted Approach to Media Discourse Analysis on the Construction of City Image: Chengdu Park City as an Example.
DOI: 10.5220/0011909200003613
In Proceedings of the 2nd International Conference on New Media Development and Modernized Education (NMDME 2022), pages 208-214
ISBN: 978-989-758-630-9
Copyright
c
2023 by SCITEPRESS Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)
2 METHODOLOGY
Lexical items ranging from word tokens to multi-
word phrases are prone to carry discoursal traces.
Nevertheless, only when linguistic elements are
categorized into patterns can discoursal phenomena
be identified and interpreted. Patterns come with
relatively large datasets which require computational
methods to process for efficiency’s sake. Against the
tide of digital Social Science and Digital Humanities,
digital methods are at disposal of communication
scholars to undertake computer-powered research
into the media discourse. However, the bar is high for
those researchers without relevant technical
expertise. Therefore, this study attempts at an
alternative: combining corpus linguistics and
discourse analysis, i.e., quantitative corpus-based
research is employed to supplement the qualitative
analysis of patterns of the language across texts as
well as the social and cultural contexts. The
advantages of merging corpus-based approaches and
discourse analysis lie in varietal aspects. First, it
reduces the cognitive biases of the researcher; second,
it identifies the incremental effect of discourse
presented in a large number of discourses; third,
corpus linguistics advocates triangulation, in other
words, utilization of multiple research methods and
different forms of data to help check the validity of
hypotheses and provide more robust interpretations
and explanations to support the research results
(Baker 2004). Corpus linguistics software tools have
in-built statistical measures for automatic
computation and thus are easy to use from a technical
perspective.
3 RESEARCH DESIGN
3.1 Research Data
We retrieve our articles from the digital database
Nexis Uni with “Chengdu Park City” or “Chengdu
Garden City” as search terms, resulting in 185 texts
spanning February 2018 to August 2021 and the
former was the date when President Xi Jinping
introduced the term Park City on a visit to Chengdu.
Having been manually screened to remove duplicate
texts, foreign media reports, and the news that merely
mentioned Chengdu Park City without relevant
contents, the corpus is reduced to 69 texts (60,080
word tokens), including reports in English published
by China’s media outlets, as well as the English
version of press releases issued by Chengdu’s
companies and organizations via foreign media
channels.
3.2 Corpus Linguistic Software
Albeit the benefits discourse analysis can reap from
computational assistance, the barrier remains high for
scholars who lack programming expertise. In this
study, we use corpus linguistic software called
#LancsBox and it is a new-generation software
package for the analysis of language data and corpora
developed at Lancaster University. There are five
modules in #LancsBox: KWIC, Whelk, GraphColl,
Words and Text. First, KWIC generates concordance
lines sorted alphabetically according to the
surrounding text on the left or right. Second, Whelk
shows the distribution of the search term across
corpus files. Then, GraphColl identifies and builds
visual networks of linguistic collocations. It is worth
pointing out that GraphColl provides varietal
statistical choices to identify collocations since there
are no one definite sets of collocates. Next, The
Words tool compiles frequency and/or keyword lists
in terms of types, lemmas and POS categories. Last
but not least, the Text tool enables in-depth analysis
on the context of a search term. Given the above
computer-aided functionalities,
#LancsBox can be
used for a combination of automated quantitative
research (Whelk, GraphColl, Words) and qualitative
research (KWIC, Text). The remainder of this article
will demonstrate how communications researchers
can harness the power of #LancsBox to conduct
media discourse analysis and a potential workflow in
this regard.
3.3 Workflow
3.3.1 Collocation Analysis
In this corpus-based study, Park or Garden is set as
the node, and the span is set as 5 words to the left and
5 to the right. MI score is a measure of collocational
strength, and a MI score of 3 or higher is proposed to
be taken as evidence that two items are collocates
(Hunston 2002). However, MI Score is affected
strongly by frequency, and low-frequency words tend
to reach a high MI score. To avoid such limitations, I
also consider words with T-score of 2 or higher since
which is normally considered to be statistically
significant (McEnery, Xiao, Tono, 2006). This study
draws on both cut-off points above mentioned as the
benchmark for identifying statistically significant
collocates.
Using Computer-Assisted Approach to Media Discourse Analysis on the Construction of City Image: Chengdu Park City as an Example
209
Table 1. Statistically Significant Collocates of Park or Garden
Node word Collocate Freq. Joint freq. MI T-score
park
OR
garden
(total freq.441)
sino-french 32 20 6.42 4.42
jincheng 17 10 6.34 3.12
livable 41 24 6.33 4.84
forest 28 16 6.29 3.95
science 46 21 5.97 4.51
mountain 32 14 5.91 3.68
beautiful 60 26 5.90 5.01
high-tech 53 22 5.83 4.61
agriculture 53 19 5.62 4.27
concept 59 21 5.61 4.49
agricultural 74 21 5.28 4.47
bureau 47 12 5.13 3.37
central 52 12 4.99 3.35
technology 94 20 4.87 4.32
revitalization 49 10 4.81 3.05
airport 63 11 4.58 3.18
cultural 145 24 4.51 4.68
industry 154 25 4.48 4.78
rural 81 12 4.35 3.29
first 124 18 4.32 4.03
tianfu 333 48 4.31 6.58
urban 200 26 4.16 4.81
lake 85 11 4.15 3.13
industrial 141 17 4.05 3.87
ecological 136 16 4.01 3.75
innovation 100 11 3.92 3.10
greenway 113 12 3.87 3.23
area 464 48 3.83 6.44
development 436 39 3.62 5.74
green 137 12 3.59 3.18
new 659 54 3.49 6.70
international 282 19 3.21 3.89
As is shown in Table 1, among the collocates,
tianfu, new, area stand out and the three of them
reveal in partnership the strategic position of Tianfu
New Area. In February 2018, General Secretary Xi
Jinping stated during his visit to Tianfu New Area that
it is an important node for the Belt and Road initiative
and the development of the Yangtze River Economic
Belt, and it must be planned and built well, in
particular, special attention should be paid to
highlight the characteristics of Park City and
ecological value should be taken into consideration so
as to create new growth poles and build inland open
economic highlands. And then, the word first
indicates that Tianfu New Area is the starting point of
the Park City discourse.
Also on the list are development and concept
which point to the New Development Philosophy the
Park City initiative adheres to; innovation, green, and
ecological are indicative of Chengdu’s being
innovative and green while seeking Park City
development; as for the aboutness of international, I
explore its collocational network for further evidence.
As can be seen from Figure 1, the construction of Park
City in the new era focuses on creating an open
international environment where Chengdu spreads to
the world the development concept of opening up via
cooperation and exchanges in domains such as trade,
art, and sports.
NMDME 2022 - The International Conference on New Media Development and Modernized Education
210
Figure 1 The collocational network of international
In addition, rural and urban are ranked similarly
in Table 1 and such similarity in collocational
strength indicates that Chengdu emphasizes
coordinated urban-rural development while
attempting to build a Park City. Also noteworthy is
that beautiful and livable are ranked third and seventh
respectively in Table 1, and the reason why they are
both highly ranked is that Chengdu has pinpointed its
strategic positioning as a beautiful and livable Park
City.
To fully interpret the computed collocates above
mentioned, the context relevant to Park City
discourse has to be taken into consideration.
Retrospectively, at the Fifth Plenary Session of the
18th CPC Central Committee, the New Development
Philosophy featuring innovation, coordination, green
shift, open and shared development, was put forward
for the first time. Then, at the Fifth Plenary Session
of the 19th CPC Central Committee, it was explicitly
required that the New Development Philosophy be
implemented on all fronts.
Based on the above collocational analysis, it can
be noted that the New Development Philosophy has
permeated the Park City discourse and such
philosophy has been collocationally realized in
textual data with different representations.
3.3.2 Key Clusters Analysis
In order to investigate the diachronic changes in Park
City discourse, I divide the corpus into three
subcorpora by time: 2018, 2019, and 2020-2021 and
merge the data of 2020 and 2021 into one subcorpus
since there are only two relevant reports from 2021 at
the time of data collection. Statistically, the sizes of
the three subcorpora are 17,277 tokens, 26,216 tokens
and 14,477 tokens respectively.
Keywords are items of unusual frequency in
comparison with a reference of some suitable kind
(Scott, Tribble, 2006). There is a significant
difference in the frequency of occurrence of
keywords in the study corpus compared with the
reference corpus, i.e., the frequency in the former is
significantly higher or lower than that in the latter.
The reference corpus used in this study is the news
subcorpora of Crown and Clob (Xu, Liang, 2013),
which was co-built in 2012 by Professors Xu Jiajin
and Liang Maocheng of The National Research
Centre for Foreign Language Education (NRCFLE)
attached to Beijing Foreign Studies University. An
examination of the keywords that occur when two
corpora are compared together should reveal the most
significant lexical differences between them, in terms
of aboutness and style (Baker 2006). Compared with
individual keywords, key clusters are more
informative as indicators of the aboutness of
discourses. Thus, in this study, I focus on the top 10
five-word key clusters, as shown in Table 2.
Table 2. Key Clusters Related to Chengdu Park City
Key Clusters
2018
Keyness
Frequenc
y
Key Clusters
2019
Keyness
Frequen
cy
Key
Clusters
2020-21
Keynes
s
Freque
ncy
agricultural
science and
technology
park
12.00 19
internet
audio &
video
industry
4.38 9
summer
world
university
games of
Chengdu
5.15 6
sino-french
agricultural
science and
technology
12.00 19
the belt and
road
initiative
4.38 9
2020 sino-
european
new
economic
online
4.45 5
Using Computer-Assisted Approach to Media Discourse Analysis on the Construction of City Image: Chengdu Park City as an Example
211
the sino-
french
agricultural
science and
7.95 12
of wangcong
tianfu
cultural
industry
4.00 8
Chengdu
new
economic
developmen
t
commission
4.45 5
the belt and
road initiative
6.20 9
overall
planning and
core area
4.00 8
issued the
following
news release
4.45 5
the park city
construction
bureau
5.05 7
wangcong
tianfu
cultural
industry park
4.00 8
the Chengdu
new
economic
developmen
t
4.45 5
between
Sichuan
province and
France
4.47 6
Chengdu
douban red
international
carnival
3.63 7
world
university
games
Chengdu
2021
4.45 5
science and
technology
p
ar
k
p
roject
4.47 6
planning and
core area
conceptual
3.63 7
base of giant
panda
b
reeding
3.76 4
a beautiful and
livable park
3.89 5
the overall
planning and
core
3.63 7
Chengdu
research
b
ase of
g
iant
3.76 4
beautiful and
livable park
city
3.89 5
and core area
conceptual
scheme
3.25 6
new
economic
online
cooperation
conference
3.76 4
belt and road
initiative and
3.89 5
chengdu
municipal
bureau of
commerce
3.25 6
research
base of giant
panda
3.76 4
As is shown in Table 2, key clusters in the three
time periods are both independent and related if
taking into account the Park City discourse as a
whole: the belt and road initiative indicates that the
development of Park City in Chengdu has been
deeply integrated into China’s belt and road initiative;
internet audio & video industry and the Chengdu new
economic development imply that the Park City
initiative gives Chengdu momentum to attempting a
new paradigm for the economic development;
agricultural science and technology park, sino-french
agricultural science, sino-european new economic
online, wangcong tianfu cultural industry park, base
of giant panda breeding, summer world university
games, douban red international carnival are all
unarguable evidences that Park City project
coordinates the development in science and
technology, international cooperation, culture,
agriculture, ecology, as well as sports and music, the
joint efforts of which lead Chengdu to a sustainable
and all-round urban development.
3.3.3 Concordance Analysis
Then, while analyzed across time, the cooccurrence
of 5-word key clusters within the same time span
uncovers the sociohistorical contexts and primary
concerns of that specific period. Through the top 10
key clusters in 2018, it can be noted that the lexical
clusters related to Agricultural Science and
Technology Park occur four times, i.e., agricultural
science and technology park, sino-french agricultural
science and technology, the sino-french agricultural
science, science and technology park project.
Background research shows that the Sino-French
Agricultural Science and Technology Park covers an
area of 18 square kilometers with a total planned
investment of about 30 billion yuan. It was listed in
the 2015 Global Climate Summit Plan and also as one
of the 100 key projects in Sichuan Province in 2018.
Below is the expanded version of a concordance line
with Sino-French Agricultural Science and
Technology Park being its node.
NMDME 2022 - The International Conference on New Media Development and Modernized Education
212
Extract (1): Sichuan Province, having been
working in the spirit of General Secretary Xi Jinping,
launched the Sino-French Agricultural Science and
Technology Park Project, in order to carry out the
strategy of park city for practical results.
Saliently, this extract argues that the Sino-French
Agricultural Science and Technology Park is a solid
example of turning the Park City vision into action
and it sets a model for international cooperation in
constructing the Park City.
Moving forward to 2019, it is evident from Table
2 that 6 out of the top 10 key clusters relate to
Wangcong Tianfu Cultural Industry Park, i.e., of
wangcong tianfu cultural industry, wangcong tianfu
cultural industry park, overall planning and core
area, planning and core area conceptual, the overall
planning and core, core area conceptual scheme. The
reason why lexical words related to Wangcong Tianfu
Cultural Industry Park repeatedly occur in the 2019
Park City discourse is that the project solicited
international consultation on the overall planning and
core area conceptual scheme design at its preparation
stage. As such, it spreads around the world the Park
City concept as demonstrated in the following extract:
Extract (2) Wangcong Tianfu Cultural Industry
Park is a key carrier for the district to create a world-
renowned cultural and creative city and a central
urban area of the park city that embodies the new
development concept in an all-round way.
In the same vein, I also note that lexical words
related to new economic development show marked
frequency in the 2020-2021 Park City discourse, i.e.,
new economy online, new economic development, and
new economic online cooperation. This can be
explained by the influence on the city of Chengdu of
the 2020 Sino-European New Economic Online
Cooperation Conference hosted by the Chengdu new
economic development commission. Taking into
account the Dual Circulation development pattern
and the normalization of epidemic prevention and
control in China, this virtual conference reflects
Chengdu’s efforts in seeking alternative economic
approaches to building a Park City. As is shown in the
example below, Chengdu has been exploring novel
ways to organically combine Park City construction
and new economic development in a bid to build a
more harmonious urban space featuring sustained
economic development.
Extract (3) Building a so-called island for
unicorns... is another ambitious initiative of the city
to develop the new economy. To help advance the
cause of the unicorn hub, the Chengdu government is
seeking solutions to planning, building, and
managing models for the balanced development of
industries, people, and nature in the garden city.
A diachronic analysis of key clusters helps map
out the actions Chengdu has taken to implement the
Park City initiative. From the above analysis, I can
construe that Park City is discursively represented as
a project carried out in a multitiered, phased, and
focused manner, and in turn, the discursive function
of Park City has been cumulatively enhanced as time
proceeds.
4 ANALYSIS
In conclusion, this study draws on the methodological
synergy of corpus linguistics and discourse analysis
to explore how the Park City image is discursively
constructed in international communication reports. I
find, in a holistic view, Park City has been
discursively represented as being a beautiful and
livable city with humans at the center and with New
Development Philosophy as guidelines. In sum, the
discursive construction of Park City exhibits the
following three features.
First, the discursive construction of Park City in
international communication reports is a dynamic
process, demonstrably being socially shaped as well
as socially constitutive. Ever since the Park City
initiative began in February 2018, Chengdu has been
implementing the rural vitalization strategy and
exploring the representation of Park City in rural
Chengdu by establishing projects like the Sino-
French Agricultural Science and Technology Park;
with the release of the Three-Year Action Plan for the
Construction of Three Cities and Three Capitals in
Chengdu (Official Website of Chengdu Municipal
People’s Government, 2019), Chengdu has witnessed
the building of Wangcong Tianfu Cultural Industrial
Park imbued with ancient Shu culture; currently,
under the combined impact of worldwide changes of
a scale unseen in a century and the global coronavirus
pandemic, Chengdu proactively develops the new
economy to create positive momentum for Park City
construction. To reiterate, the discursive construction
of Park City at different periods is always amenable
to social contexts and keeps abreast of China’s
development needs.
Second, the discursive construction of Park City
in international communication reports features
dialectical relations. On the one hand, the Park City
discourses are representations of Park City practices
that adhere to the New Development Philosophy and
ensure harmony between humanity and nature. On the
other hand, such representations enter and shape Park
Using Computer-Assisted Approach to Media Discourse Analysis on the Construction of City Image: Chengdu Park City as an Example
213
City construction practices, leading to the
establishment of such major ecological projects as
Longquanshan Forest Park, Tianfu Greenway
System, Jincheng Park, Sino-French Agricultural
Science, and Technology Park, Chengdu Agricultural
Highland Park, Airport Central Park, Wangcong
Tianfu Cultural Industrial Park, etc.
Third, the discursive construction of Park City in
international communication reports is an enactment
of our country’s mainstream ideology. As stated by
Teun A. van Dijk, discourse plays a prominent role as
the preferential site for the explicit, verbal
formulation and the persuasive communication of
ideological propositions (VanDijk 1995). In Park
City media discourse, the New Development
Philosophy has been expressed and reproduced in the
context of the mass media. Conversely, such
ideologically shaped media discourse may, in turn,
affect the cognitive representations of the Park City
of Chengdu on the part of the recipients.
5 CONCLUSION
In this study, I employ computer-assisted corpus
linguistics analysis to examine the construction of
Chengdu’s city image in media reports for
international communication. With the specific
research question determined a priori, the workflow
proceeds in a top-down manner, i.e., the data research
starts with collocation analysis to frequency of key
clusters for quantitative statistics and then lexical
network graphs and concordance lines follow up to
aid in the manual, qualitative analysis. Such an
approach is an attempt at harnessing the synergy of
computational technology and discourse analysis.
Since communications research usually involves
hundreds of thousands of texts, the methodological
framework of this study can inform other research
projects associated with discursive construction and
communication of a city’s image. At the same time, it
would be desirable to further examine the extent to
which computer-assisted corpus linguistic programs
are able to contribute to media discourse analysis.
REFERENCES
Baker, P. (2004). Querying Keywords: Questions of
difference, frequency, and sense in keywords analysis.
Journal of English Linguistics, 4, 346-359.
Baker, P. (2006). Using Corpora in Discourse Analysis.
London: Continuum, 10-17.
Fairclough, N. (2003). Analysing Discourse: Textual
analysis for social research. London: Routledge,124.
Hunston, S. (2002).Corpora in Applied Linguistics[M].
Cambridge: Cambridge University Press, 71.
Ke, X., Xi Chen.,& Mingyang Yu (2012). Can the World
Expo change a city’s image through foreign media
reports? Public Relations Review 38:746–54.
Kotler, P., and D. Gertner. (2002). Country as brand,
product, and beyond: A place marketing and brand
management perspective. Journal of Brand
Management 9 (4):249–61.
McEnery, T., Xiao, R. and Tono, Y. (2006) Corpus-based
Language Studies: An Advanced Resource Book. New
York: Routledge,57.
Official Website of Chengdu Municipal People’s
Government (2019). The release of the Three-Year
Action Plan for the Construction of Three Cities and
Three Capitals in Chengdu. Retrieved from
http://www.chengdu.gov.cn/chengdu/home/2019-
02/15/content_7559011846374b33ab32ae613a08e03f.
shtml?share_token=b060aee2-ae7b-4ce2-b23d-
e8b9121c43bc
Scott, M. &C. Tribble. (2006). Textual Patterns: Key
Words and Corpus Analysis in Language Education.
Amsterdam/Philadelphia: John Benjamins Publishing
Company, 55.
VanDijk, T.A. (1995). Discourse analysis as ideology
analysis. In C. Schäffner & A. Wenden (Eds.),
Language and Peace . (pp. 17-33). Aldershot:
Dartmouth Publishing. 1995.
Xu, J. & M. Liang. (2013).A tale of two C’s: Comparing
English varieties with Crown and CLOB (The 2009
Brown family corpora). ICAME Journal,1,175-183.
NMDME 2022 - The International Conference on New Media Development and Modernized Education
214