Engagement resources in the English news texts of two phases in China’s constructing Greater Bay Area

Abstract This study examined the similarities and differences in the distribution and quantitative use of engagement resources in the English news texts of two phases which are demarcated by an important document issued by the Chinese government in constructing the Greater Bay Area (GBA) by drawing on the engagement framework of the appraisal system in Systemic Functional Linguistics. Sixty English news reports of these two phases, each with a length of 800–1200 words, were collected and all the engagement resources were coded and calculated on the UAM corpus tool (version 6) before comparisons were made in terms of the quantitative use and distribution of each sub-category. The findings show that more dialogic expansion resources are present in the English news texts on GBA than dialogic contraction resources and that the most used and the least used strategies in the news reports are entertain resources and distance resources respectively. Moreover, in terms of the comparison of the sub-category, the distribution of the endorse strategy is significantly different in these two phases.


Introduction
Engagement is concerned with "the ways in which resources [...] position the speakers/writers with respect to the value positions being advanced and with respect to potential responses to that value position [...]" (Martin & White, 2005, p. 36).Z. Yu's (2021) review of the literature demonstrates that it has been under examination from divergent perspectives: the traditional descriptive grammar perspective (Quirk et al., 1985), the cognitive perspective (Muchin, 2001), the Systemic Functional perspective (M. A. K. Halliday, 1994), the dialogistic and heteroglossic perspective (Martin & White, 2005) and so on.Consistent with Bakhtin's (1981) recognition of heteroglossia and the dialogic nature of discourses, Martin and White (2005) put forward their framework of engagement, which is concerned with "sourcing attitudes and the play of voices around opinions in discourse " (p. 35).In this framework, there are two broad categories of engagement resources: dialogic expansive resources comprising disclaim resources and proclaim resources and dialogic ABOUT THE AUTHOR Fan Cao is a Ph.D. student in the Department of Foreign Languages and Literatures at Tsinghua University, Beijing, China.Her research interests include sociolinguistics, discourse analysis, and functional linguistics.contractive resources consisting of entertain resources and attribute resources.In other words, in light of the different functionalities of varying engagement resources to dialogic spaces, the latter can be either expanded or contracted so as to meet divergent communication ends: to align readers/listeners with the given position or to achieve greater solidarity by acknowledging the existence of alternative positions (Martin & White, 2005).As the engagement theory places emphasis on the attitude of voices, it is conspicuously instrumental in analyzing how discourses engage with their readers/listeners.Therefore, it has informed various lines of inquiry such as voices in news studies, genre studies, and academic discourse (e.g., Chang & Schleppegrell, 2011;Huan, 2016;Miller et al., 2014;Y. Yang, 2020;Z. Yu, 2021).Since this study focuses on the ways in which a Chinese official news media platform, namely China Daily, engages with its divergent readers with the purpose of aligning readership or achieving solidarity when reporting a regional development at different times, the framework is specifically conducive to solving the concern of the present study.
The focal region in the current research is Guangdong-Hong Kong-Macao Greater Bay Area, also known as the Greater Bay Area or GBA according to its official website (https://www.cnbayarea.org.cn/english/About%20GBA/content/post_165618.html).It is a national project of China aimed at promoting synergistic development among eleven cities in the region previously known as the Pearl River Delta (PRD) and two Special Administrative Regions (SARs, namely, Hong Kong and Macao).The importance of this project is not only revealed in promoting regional development but is also represented in its strategic position as the geographical node linking countries and regions in other national development strategies such as the Belt and Road Initiative promoted by President Xi in 2013 (Hui et al., 2020).Due to this significance endowed nationally, an increasing body of literature concerning various aspects of this mega-city region has arisen since the initiation of this plan (P.Chen & Li, 2019;Hui et al., 2020;Li et al., 2021;Meulbroek et al., 2022;Park & Song, 2021;H. Yu, 2019) and has aroused a large amount of attention from news media platforms.However, given the multicultural and multilingual context and the coexistence of two social ideological systems (namely, socialism in the PRD cities and capitalism in SARs) in the city cluster, it is evident that the Chinese official publicity platform takes on great challenges of both engaging with as many subjects and audiences as possible and fully narrating this area so as to echo the bigger national ambition of "better narrating China and making China's voice heard" (The State Council of People's Republic of China, 2022).How to better align with the maximum number of national and international readers and provide a fair amount of news information while achieving solidarity in tandem is worth carefully examining, especially under the circumstances of the construction of GBA involving immediate stakeholders with contrasting ideologies.Despite the contribution that various studies have made to the analysis of the news discourse on GBA to the international community (Du & Hou, 2020;Luo, 2020;Meulbroek et al., 2022;Xu, 2020), relatively few studies have focused on how the news texts of GBA engage with the readers both home (including the SARs) and abroad in order to achieve solidarity and alignment.This engagement process or negotiation process, as Martin and White (2005, p. 207) point out, bears the dynamic characteristics during which readers "align and disengage" in reaction to various communities.This observation in tandem with the lack of previous relevant research necessitates the examination of engagement in news texts of GBA in different phases.Therefore, the deployment of engagement resources in different phases of GBA construction is worth adumbrating as such a comparison is bound to provide a closer investigation of the dynamic publicity strategy in divergent times on the part of the official news platform.For this purpose, a pivotal turning point of this national project (the release of the Outline Development Plan of GBA) will be assigned the role of dividing the two phases in the present study.
The purpose of this study is twofold.First, by examining engagement resources in the English news texts of two phases in the construction of GBA, this paper not only seeks to reveal the similarities and differences of engagement deployment in news texts of two phases and further explore the potential and deeper reasons for the results, but it also aims to compare the distribution of each engagement resource in two phases.Second, this study is also interested in uncovering possible ways in which official news media platforms engage with readers.More specifically, it is intended to understand the means by which the dialogic spaces are expanded or contracted and the contingent effect on the author-reader communication by contextualizing engagement resources in the news texts.
The remainder of this paper is organized as follows.Studies on engagement framework and its application, research on GBA, especially its Outline Development Plan, and previous research on its news text will be summarized in the next section.This is followed by the research design where data collection, coding, and analysis will be introduced.After that, the paper presents the major findings on the quantitative use and distribution of engagement resources and seeks to explore possible reasons for the results.This paper concludes with implications as well as limitations in the last section.

Engagement system in SFL framework
Systemic Functional Linguistic (SFL) theory (M. A. K. Halliday, 1994) proposes three metafunctions of language, namely, ideational metafunction, interpersonal metafunction, and textual metafunction, each containing corresponding sub-systems to further elaborate on how language works in construing experience, negotiating relationship as well as organizing texts.As an interpersonal sub-system at the level of discourse semantics, appraisal theory was developed by Martin and White (2005) with the inclusion of three sub-systems which are the attitude system, engagement system as well as graduation system.Informed by Bakhtin's dialogism in his investigation of voices in literary texts (Bakhtin, 1981) and intertextuality later proposed by Kristeva (1986), the engagement system is concerned with the ways in which the discourse aligns or disaligns with its audience.Just as Wang and Lu (2010) summarize, engagement involves varying attitudes, with which authors, as Martin and White (2005) further observe, show as well as negotiate their "alignment/disalignment" with readers.In the same vein, Huan (2016) critically summarizes that the engagement framework transcends a static understanding of personal attitudinal meaning and enables a dynamic meaning negotiation between interlocutors (e.g., news authors and readers in the present study).
The engagement theory has informed divergent lines of inquiry, covering a wide range of research topics.For example, in her study regarding export-oriented textbooks, Y. Chen (2013) compares different engagement resource usages in English and Chinese cultural textbooks and concludes that dialogic expansion and contraction strategies are deployed differently.She also develops the engagement framework for the analysis of multimodal pedagogical discourse (Y.Chen, 2010).Meanwhile, with respect to the news text in particular, numerous studies have contributed their part to the application of this framework (see, Huan, 2016;Jiang & Chen, 2017;White, 2003White, , 2012)).For instance, Huan (2016) investigates patterns regarding the engagement of news sources in the hard news both in Australia and China by resorting to three engagement features: acknowledge, endorse, and attest (a newly introduced feature).In the study, he reports that Australian journalism practitioners used different engagement resources to source others compared with Chinese newspapers, especially regarding the official source, on which he explicated that the political and socio-cultural context may exert their influence.Jiang and Chen (2017) analyze the negotiation pattern of dialogic spaces in different parts of each English news text on the Belt and Road Initiative from China Daily based on the engagement framework and the generic structure of the news text.They further confirmed how various engagement resources are deployed by the news platform to expand or contract dialogic space in the heteroglossic news text out of its specific intention.In other words, these studies all manifest the application of the engagement framework which can be reflective of the means and strategies of how authors engage with their readers.It is out of this observation that this research draws on the framework of engagement to explore and reveal how and the extent to which China Daily deploys different engagement resources to align with or achieve greater solidarity with divergent readers through the formation of differentiated dialogic spaces.
In order for a better understanding of the content of the engagement framework, linguistic realizations of each strategy are exemplified in Figure 1.Here, it is also worth mentioning that besides dialogic texts, there are also monoglossic texts with no engagement resources present, that is, no voices included, hence, this category is not in the scope of this research.In other words, it is in the attitudinal/heteroglossic texts that a wide array of engagement resources is deployed to "negotiate solidarity with a complex readership" (Martin, 2004).These resources differ in their influence on dialogic spaces: disclaim and proclaim categories contract the dialogic spaces by "excluding" or "constraining" the dialogic alternatives (Martin & White, 2005, p. 117) while entertain and attribute categories lead to dialogic expansion in that they invoke and invite dialogic alternatives.In this vein, the writer-reader relationships vary accordingly.
As shown in Figure 1, the sub-categories of disclaim include deny and counter.Deny, also called denial and negation, refers to the rejection of the position after introducing and acknowledging it (Martin & White, 2005, p. 118).It can be realized by expressions like no, not, never, don't, etc. Counter strategy such as still, yet, instead, but, even, and only provides viewpoints that contradict the reader's expectation and it construes the writer as surprised as the reader.As regards proclaim in the framework of Martin and White (2005), concur, pronounce and endorse are subsumed.Concur refers to the strategy by which the author explicitly announces his agreement with the readers.The news reports employ concur resources such as obviously, certainly, and no doubt.Pronounce involves "authorial emphasis or explicit interventions or interpolations" (p.127), which include in fact, the fact of the matter is, and as we all know in the news text.With endorse such as the statistics demonstrate, the data has shown that . . ., the speaker/writer formulates the propositions by means of sourcing "correct, valid, undeniable, or maximally warrantable" (p.126) external sources.Endorsement resources like the report found that, the blueprint shows are present in the texts.These five resources are all utilized to contract the dialogic space, as such, the author aligns the reader to his point of view while sustaining the dialogical nature of the news texts.
On the other hand, dialogic expansive resources are composed of entertain and attribute.According to Martin and White (2005), entertain constructs the voice of the author as one of the possible positions so that readers' voices can be largely accommodated to be dialogic alternatives.It can be realized by mental verbs, for example, I think and modals such as may.Besides entertain resources, attribute, including acknowledge and distance, "represent the proposition as grounded in the subjectivity of an external voice" (p.98) and the textual voice as but one of the possible voices.Acknowledge resources such as reportedly, it is to be hoped that, according to are also used in these two corpora.According to Martin and White (2005), distance is differentiated from acknowledge in that the authors not only attribute the information to external sources but also are explicitly disassociated from these voices, thus creating the maximal dialogic space for alternative views.Distancing strategies deployed in the corpora encompass it is a shame that, criticism of, etc.
To offer a fine-grained understanding of the ways in which engagement resources impinge on the expansive dialogic space and the contractive one, two examples are given to elaborate on these two aspects in a contextualized way.
(1) Of course, the per capita GDP of financialized Hong Kong is still higher than Shenzhen.
(Phase One-text 1) The expression of "Of course" in example (1), being a concur strategy, construes readers and the writer as sharing the same knowledge that Hong Kong's per capita GDP is higher than that of Shenzhen, a PRD city in the GBA area.By taking this to be universally shared and known, the author assumes the consensus of this "background information" from all the readers, thus contracting the dialogic space.
(2) . . .development of the Greater Bay Area may help relieve the land shortage problem to a certain extent by diverting some demand to the neighboring region.(Phase One-text 1) By using the modal "may" in example (2), one of the possible advantages of GBA brought to Hong Kong, that is, alleviating the land shortage problem, is mentioned.Here, the author's voice, eschewing the imposition on readers, is weakened to accommodate more stances on solutions to this problem or more advantages of GBA construction.That is also to say, other voices are entertained in this sentence.Given the context of this text, which was in 2017 when the GBA construction was in its inception, with great challenge and an uncertain future envisioned, it is safe and credible for the author to entertain other latent views.
Following the exemplification of the theoretical framework, the ensuing sections engage with the context of the data, that is, GBA and the Outline Development Plan.

Guangdong-Hongkong-Macao Greater Bay Area
As the fourth bay area in the world, the other three being San Francisco Bay Area, Greater Tokyo Area, and the New York Metropolitan Area, the Greater Bay Area received profound attention since the idea was first proposed by the State Council of China in 2015 in the 13 th Five-Year Plan for Economic and Social Development of the People's Republic of China.As a major national strategy, building such a megacity region aims to promote synergy development and deepen cooperation among nine mainland cities in the Pearl Delta Region and two SARs, that is, Hong Kong and Macao under the "one country, two systems" policy.Despite the first proposal in 2015, it did not start the stage of formal planning and layout until 2017 when the National Development and Reform Commission and the governments of Guangdong, Hong Kong and Macao signed the framework agreement.The national strategy has much to offer both in terms of its status in the country and its global orientation.Internally, with a total land area of 56 square kilometers and a population of 67 million, the greater bay area is by itself a large economic powerhouse and development engine worth noting in China.With all-around integration, mainland cities and SARs are expected to benefit from the reciprocal exchanges of resources and capital in this most economically developed region.Externally, the construction of GBA is of immense importance in that it resonates with the development of the Belt and Road Initiative as a facilitative connector with the neighboring countries along the Maritime Silkroad (H.Yu, 2019).Given the significance of this strategy and the appeal from the central government to construct China's international discourse and narrative systems, publicity stands out as extremely urgent and important for the greater bay area.There is a growing body of literature on the overall study of this region as well as on its media discourse.By virtue of bibliometrics and CiteSpace, X. Chen (2021) finds that besides taking on interdisciplinary characteristics and attracting rising scholarly attention, studies on GBA mainly focus on six subjects, namely, bay economy, city clusters, Macao, synergy development, scientific innovation as well as relations with Belt and Road Initiative.Its media discourse studies mainly deal with topics such as the identity of cities and the image of the bay area (Du & Hou, 2020;Ju, 2021;Luo, 2020;Meulbroek et al., 2022;Xu, 2020),).For instance, Ju (2021) discusses the extent to which Hong Kong identifies with its national identity by investigating the news texts of the Greater Bay Area collected from the South China Morning Post.Du and Hou (2020) claim the importance of constructing a positive image of China in the media discourse of the Greater Bay Area while Xu (2020) and Luo (2020) examine the relationship between the image of the Greater Bay Area and English news texts about it from the perspective of intertextuality and multimodality respectively.

Outline development plan for the Bay Area
In the wake of the initial development of GBA, a crucial document was released, namely, the Outline Development Plan, in order to further clarify several aspects of the region's development.This section is intended to introduce the landmark document.
On 18 February 2019, the Central Committee of the Communist Party of China and the State Council unveiled the Outline Development Plan for Guangdong, Hong Kong, and Macao Greater Bay Area which specified the main aspects of development and cooperation in the region with its 11 chapters and 41 articles.These aspects include science and technology, infrastructure, modern industrial system, and so on.It served as the role of guidance for coordinating the construction of the Bay Area, signifying the arrival of "the new era of rapid integrative development" (The State Council of People's Republic of China, 2019).Just as L. Yang (2022) notes in her study, the outline manifests itself as the mark of comprehensive progress of the Greater Bay Area construction.Its pivotal role is further illustrated as "the turning point for the formation of a city-cluster" in that the construction turns to the implementation stage after the release of the outline (H.Yu, 2019) and clarification of the "strategic position" of this region (Li et al., 2021).It can be concluded that the Outline Development Plan for the Bay Area takes up a significant role, based on which this research divides the phases of comparison into two: one before the date of release and another after the date of release.

Data
News texts used are all drawn from the official website of China Daily, which is the leading media platform in China.As the primary outlet and means for showcasing China to the world, it enjoys a wide national and international readership of more than 350 million as reported on its website (http://www.chinadaily.com.cn/e/static_e/about).It is representative of China's official efforts to narrate China worldwide, therefore this research collects linguistic data from its website in an attempt to gain more credibility and representativeness in the result.After an initial search, more than 4000 articles and 3000 pictures mentioning Greater Bay Area are accessible on the website, thus ensuring relatively ample official data for the research.
Since the difference in engagement strategies before and after the Outline Development Plan is examined, the time span of the selected data is from the first legitimate release of the article on GBA in March 2017 to the unveiling date of the document on 18 February 2019, and 19 February 2019, to January 2021 respectively, the latter of which is to ensure the equivalence of the time span with the former one.There are three criteria for amassing the data: First, at least one of the following three keywords should be included in the title of the news: Bay Area, Greater Bay Area and GBA, to the extent that the content of such texts is inextricably linked to the topic; Second, the selected news texts are concerned with at least one aspect of GBA construction (also to avoid the report with a mere mention of this topic), for example, the economy of the area and etc; Third, only texts with a word range between 800-1200 are considered to eschew large disparity in the length of news data.On these premises, only about 30 texts in Phase One can meet the requirements, which leads to the formation of two corpora with 30 texts each, with Phase One corpus subsuming reports before the Outline releasing date and Phase Two corpus after that date.By means of the calculation by the software Antconc 4.0.9, the tokens of these two corpora amount to 29,518 and 28,034 while the types are 4309 and 4183 respectively, thus achieving the approximate equivalence in the data volume for subsequent research.

Data analysis
Data is analyzed within the engagement framework annotated on UAM Corpus Tool (version 6).Specifically, two coding schemes are established: A phase framework with the inclusion of Phase One and Phase Two and an engagement analysis framework including the sub-categories of both expansive and contractive resources.In the annotation, for example, a selected piece of news text released on NaN Invalid Date NaN is manually assigned to phase 1 in the phase framework as the date is within the time span of the first phase and linguistic expressions in the news texts are also manually assigned a corresponding engagement feature depending on their actual use, for example, the phrase "he said that . . ." will be annotated under the category of acknowledge of the expansive resources since the text attributes the content of the indirect report speech to the subject "he".After three rounds of manual annotation, engagement resources in these 60 texts will be counted.To further examine whether there are significant differences in the sub-categories of engagement resources in these two phases, the Chi-square test will be utilized by the Statistics of UAM corpus tool (version 6) after the standard frequency is calculated.Here, standard frequency means the number of each category of engagement resources per 1000 tokens of the news texts.In short, the standard frequency of each subcategory will be compared so as to find out the differences in each resource between these two phases.After that, examples randomly selected from the data of each subcategory of engagement resources will be analyzed to explore how the dialogic spaces are expanded or contracted.

Results
Tables 1 , 2, and Figure 2 show the findings regarding engagement resources in these two phases.In the first part of the result, similarities and differences in the engagement deployment in these two phases will be presented.First, in general, as can be seen from Table 1, the number of engagement resources in Phase Two is slightly more than those of Phase One: There are 979 engagement resources with a standard frequency of 31.5 per 1000 tokens in Phase Two while 924 items with a standard frequency of 28.8 per 1000 tokens are present in Phase One.This initial finding demonstrates that there are more dialogic negotiation processes in the news report after the release of the Outline Development Plan.
Second, it can also be seen in Table 1 that dialogic expansion resources outweigh dialogic contraction resources in both news corpora: Compared with 44.3% contraction resources, 55.8% expansion resources are utilized in Phase One.When it comes to Phase Two, expansion resources account for 56.3% and contraction resources take up 43.7%.This indicates that more efforts in expanding the dialogic spaces are made by news report authors in the English news on GBA.In Martin and White's (2005) terms, solidarity can be achieved by agreement from the readers and also the recognition of the presence of alternative viewpoints.In this sense, dialogic expansion resources can realize solidarity because dialogic spaces are largely opened up to allow for other views in this dimension.The finding that there are more expansion resources in both corpora can be seen as an effort of the news platform to achieve solidarity with a wider readership.This includes employing entertain resources and attribute resources.Looking at Figure 2, it is apparent that more entertain resources are present than attribute resources, with percentages of 33.2% versus 22.6% in Phase One and 31.8%versus 24.5% in Phase Two.It is also noticeable in Figure 2 that there are manifestly more disclaim resources than proclaim ones in two phases with the inclusion of 37.9% and 36.2%disclaim contrasted with 6.3% and 7.6% proclaim resources.By disclaim resources news writers introduce alternative positions and then directly reject them, which can further stress and reinforce the authors' voices.Also, it is true that by proclaim resources the author can limit the scope of alternatives, but the extent to which the authorial voice is expressed differs.This difference can explain the findings that more disclaim resources are found than proclaim resources since the authorial voice is very important in leading or influencing the readers to take their own stance in both construction phases of GBA, for example, to conceive of GBA as not weakening the basic law (example 6), etc.
Finally, each sub-category is also compared by using the Chi-square test to examine if there is a statistically significant divergence in these two phases.As shown in Table 2, Phase One and Phase Two exhibit no sharp contrast in the percentage of deploying strategies like deny, counter, concur, pronounce, entertain, and acknowledge whereas in the domain of endorse, there is a significant difference between these two phases with a p-value not exceeding 0.05 (p = 0.0180).This significance can be explained by the concept of register in Systemic Functional Linguistics by M. Halliday and Hasan (1985).In their view, there are three elements essential to register: Field which refers to the content of the happening, tenor which is related to the role and status of the discourse participants, and mode which denotes the particular part language is playing in the discourse.In the registers of Phase One and Phase Two, the modes are both formal written language while the tenors are both discourses between the news writer and readers.However, fields differ as a result of the progress in different stages of development: Compared with Phase One, the construction of the Greater Bay Area in Phase Two proceeds further and enters into a new stage, especially with the release of the Outline Development Plan and more official research and surveys, therefore the news reports have authoritative voices to endorse so as to increase the credibility of the news text.Texts in Phase Two arguably have more opportunity to endorse with a wide array of the official sources, such as the positioning of Macao as a core engine in the greater bay area blueprint demonstrate that. ..(text 32), it (the plan) demonstrates. . .(text 33).
In the second part of the result, each category will be exemplified to elaborate on the ways in which engagement resources impinge on dialogic spaces.Dialogic expansion resources will first be presented.Example (3) is a case of entertain resources.
(3) I suspect that it will not take as long for the world to be impressed by the Greater Bay Area as it witnesses . . .(Phase Two-text 34) Here, the use of the entertain resource I suspect that in example (3) shows that it is in the author's uncertain personal view that the world will be surprised by the construction of GBA in a short time, lowering the assertiveness of the statement on the one hand and opening up the space for dialogue with alternative voices and different views on the other hand.Given the background of this sentence that the Outline was released a few days ago, it was only in the author's conjecture that GBA construction would astonish the world.In this way, readers are participants entitled and encouraged to contribute their own views in the dialogue against this heteroglossic backdrop, therefore, readers can either align with the writer or with other readers holding contrary positions.
Other entertain resources such as if, I propose, could be, and it's highly likely that are also widely used in both corpora.It is worth noting that in both phases entertain resources outweigh all others (with percentages amounting to 33.2% and 31.8% in the two corpora respectively), in which case the heterogeneous aspects of the news report can be largely enhanced with the authorial voice as one of the possible voices.This can reduce the certainty of the proposition so as to make space for other different voices.Just as Martin and White further add, entertain "provides for the possibility of solidarity of those who hold to alternative positions, at least to the extent that those who hold to contrary positions are recognized as potential participants in the ongoing colloquy (p.109) ."It is evident that with the most entertain resources, solidarity is widely achieved by readers either with the author or with the readers holding alternative views in the dialogic texts.Besides the entertain resource, attribute resources comprising acknowledge and distance strategy can be also utilized for expanding dialogic space.Example (4) below was extracted from the news released just after the Outline, as the Outline puts it as an acknowledge resource, attributing the view that an analogy can be made between the environment and human life to the Outline.Here, the author's voice and stance are invisible, potential readers can either hold the same view as the Outline, that is, being aligned with the attributed propositions, or align with those holding dissimilar views.In both cases, solidarity can be achieved and dialogic space is expanded.
(4) As the outline puts it, to "cherish the environment as we cherish our own lives".(Phase Twotext 34) (5) Rumor has it that the absence of the master plan is to avoid further adding to the antagonistic sentiment of the US . . .(Phase One-text 21) Furthermore, Rumor has it that in example ( 5) is a distance strategy where the author holds no responsibility for the proposition of these external voices.The idea that the lack of a master plan is to avoid further aggravation of US hostility is assigned the status of the rumor here by the author as it is considered to be unverifiable and doubtful.Therefore, readers can hold maximally possible positions in terms of the reason for the absence of a master plan, in which case solidarity can be built in the texts that readers are anticipated and tolerated to either agree or disagree with the position attributed.What is worth noting is that this extract was before the releasing of the Outline, thus invoking the thoughts on the reason for the absence of such a plan.Even though distance resources can show the author's recognition of the contrary external voices so as to achieve solidarity, they are also indications of the author's explicit decline and distrust of those voices.It is for this reason that it is used the least of all the resources in the news texts of GBA (the percentage being 0.2% in Phase One and 0.3% in Phase Two in terms of usage).Having said that, the distance strategy is still used in the news on GBA to expand dialogic spaces.
Within the category of dialogic contraction, As we have mentioned in the last section, Disclaim resources include deny and counter while proclaim resources encompass concur, pronounce and endorse.The usage of disclaim is explained below with examples (6) and (7) adduced, after which the possible reason for the finding will be explored.
(6) It is repeated many times that this plan in no manner weakens "One Country, Two Systems" or the Basic Law.(Phase Two-text 34) (7) . . .international observers projected doom in Shenzhen because the forces that once boosted Guangdong's economic boom . . .were fading.But they missed the big picture.(Phase Twotext 60) In order to employ the denial strategy in example ( 6), the author first introduces the belief of some potential readers that the release of the Outline will weaken the policy of One Country, Two Systems and the Basic Law and then rejects it with a determined denial "in no manner" with the additional emphasis from the expression "(repeated) many times" to indicate the more informed stance of the author's voice.Martin and White (2005, p. 120) have distinguished corrective from confrontational negation, and example (6) is the exact case of the former in that the author seems to have enough knowledge from the multiple repetitions to correct some of the possible ideas among the readers, thus aligning readers to reject the negative effects of the Outline Development Plan on the Greater Bay Area to better reinforce its positive possibility.This rejection and negation is enormously important for the initial stage of the Outline.By means of deny resources, the author attempts to align the readers with their negated views, thereby communicating with the readers in a direct way.
In example (7) with counter resource "but", the news writer first gives the context that some international observers predict the decaying of Shenzhen's economy.This prediction may arouse some readers' same expectations of the fading economy.Then, the author uses "but" here to counter this prediction and expectation by implying it is because of their dismissal of a big picture that they rush into the precedent conclusion.Contextualizing this extract, we can easily find it was released in 2021 when the construction was in its bourgeoning stage after the Outline was released nearly two years ago.With the culmination of the construction, the author is confident to point out that a big picture (GBA) has been missed.In this way, dialogic space is narrowed as readers are assumed to hold the same view as the author otherwise they are "missing the big picture".In other words, the author presents this information as countering everyone's expectation, which contracts the dialogic space as people who attempt to disalign with this "surprising" finding will be alienated.
Moreover, Examples (8), ( 9), and (10) including respectively concur, pronounce, and endorse resources are analyzed here to sketch the map of proclaim resources.
(8) While Hong Kong profits from being "the gateway to China", this role will admittedly diminish as the Chinese mainland keeps opening up its economy and financial system.(Phase Two-text 37) (9) We have to act fast as Guangzhou and Shenzhen have been catching up really fast with increasing international exposure and linking in the last decade . . .(Phase Two-text 35) (10) Data also shows that Guangdong province grew faster than the national average, and reached 6.4 percent in the third quarter.(Phase Two-text 56) Here in example (8), concur resource admittedly is used to show it is universally known that the status of HK as a gateway to the mainland will be weakened by the opening of the latter's economic and financial system.Naturally, the dialogic space is contracted in that readers who fend off such a view will be deemed as not having common sense or uninformed.Situating the example to the text, we can discover that the author attempts to stress the importance of GBA to HK's overall development by first foregrounding a consensus of its diminishing role played in the opening up of China.
With pronounce resource really, example (9) from text 35 of Phase Two underscores the speed of development in Shenzhen and Guangzhou when introducing the advantage of HK in the GBA construction, reflecting the concomitant competitiveness of mainland cities in the Greater Bay Area.This emphasis will largely occasion the contraction of dialogic space in that voices against the textual voice will jeopardize the harmony with the overt stance and intervention from the author.
In example (10), the author uses data to endorse the fact that Guangdong Province is developing at a fast pace, which contracts the dialogic space and provides ample information to the readers.It is also hard to refute the conclusion in example (10) that is drawn from the data.Thus, readers are more inclined to accept the author's stance.
To summarize, both the contraction and expansion of the dialogic space reflect the author's efforts to maintain the dialogical nature of the text: With contraction resources, the author presents information as well as his stance and at the same time seeks to align the readers with his point of view while expansion resources allow for alternative views and can help achieve wider solidarity by giving readers an equal negotiating position to engage in dialogues with the text.The adequate use and switching of these two resources in the coverage of the Greater Bay Area can reflect the official efforts to convey ample news information to the reader with whom the news attempts to align and yet render more dialogues possible.

Discussion
It is found that more dialogic expansive resources (516 in Phase One and 551 in Phase Two) than dialogic contractive resources (408 in Phase One and 428 in Phase Two).This finding echoes Y. Yang's (2020) research in which she demonstrates that in contrast with spoken texts, written texts like news reports employ more dialogic expansive resources so that the author may coordinate various potential voices.Indeed, with more dialogic expansive resources in the news texts, readers are situated in an opened-up dialogic space and able to hold their own stances without challenging the author.
Within the category of dialogic expansive resources, entertain resources and distance resources respectively account for the most and the least percentage, with entertain being 33.2% and 31.8% in phase One and Phase Two respectively, distance being 0.1% and 0.3% in Phase One and Phase Two.On the one hand, in her study related to textbook analysis, Y. Chen (2013) found that entertain resources are one of the two most used engagement resources in English textbooks, as they can increase the negotiation and dialogue between the authors and readers, thereby enabling more space for the readers to think and evaluate.It is evident that solidarity is widely achieved by readers either with the author or with the readers holding alternative views in the dialogic texts.On the other hand, the finding that distance accounts for the fewest part is in accordance with Y. Yang (2020) where she discovers little presence of distance strategy both in oral and written texts centering around the same topic.As Y. Yang (2020) and Jullian (2011) suggest, news authors tend to express their stance in an implicit way for the sake of credibility and objectivity of the news report.This can explain why distance strategy is used the least in both of the news reports on two phases of GBA as it explicitly lays the stance bare of the author's opinion to external voices.
Moreover, in the scope of dialogic contractive resources, there are more disclaim resources (350 in Phase One and 354 in Phase Two) than proclaim resources (58 in Phase One and 74 in Phase Two), which is consistent with Jiang and Chen's (2017) study on the English new texts of the Belt and Road initiative.They critically point out that compared with proclaim, disclaim forcefully reinforces the stance of the news author by presenting the opposite opinion first before denying it.And Y. Yang (2020) also reports in her study that there are more disclaim resources than proclaim resources in the genre of written texts.Hence, news authors of GBA use this strategy for the sake of being more objective while acknowledging the existence of alternative voices, thus achieving greater solidarity.
After the comparison of the sub-category from the two phases, it is discovered that endorse demonstrates a significant difference in deployment.Just as Huan (2016) claims, Chinese news media tend to endorse "implicit hard proof sources released by the government".After the release of the Outline Development Plan and the conduction of more official research and surveys, it is natural that journalism practitioners have more official documents, data, and authoritative voices to endorse so as to increase the credibility of the news text, such as "the positioning of Macao as a core engine in the greater bay area blueprint demonstrate that. .." (text 32), "it (the plan) demonstrates. .." (text 33).

Conclusion
The present study sets out to examine the similarities and differences in the employment of engagement resources in English news texts in two phases of the construction of the Guangdong-Hong Kong-Macao Greater Bay Area by drawing on the engagement framework proposed by Martin and White (2005).In sum, it is found that more dialogic expansion resources (entertain and attribute resources) are present in the English news texts on GBA than dialogic contraction resources (disclaim and proclaim resources), thereby allowing for adequate spaces for alternative views and helping achieve wider solidarity by giving readers an equal negotiating position to engage in dialogue with the text.Meanwhile, the most used and the least used strategies in the news reports are entertain resources and distance resources respectively.Moreover, there is no significant difference in the use of engagement resources in these two phases except in the distribution of the endorse strategy, which enables the news author to lend more validity and credibility to his representation of the news.This paper has revealed the engagement resources usage in different phases of construction of GBA as well as depicted the possible ways official news media platforms intend to engage with readers.By applying a new perspective in studying the news discourse on GBA, this study might inform, to some extent, the subsequent relevant studies on this city cluster, especially with respect to its publicity.As a case study, the findings can provide some inspiration for journalism practitioners such as news authors in their endeavors to engage audiences and in a broader sense, might also shed some light on instructing pedagogical practice concerning the news genre.
Notwithstanding the above findings and implications, a number of limitations need to be noted.As one of the initial attempts to explore news texts on GBA diachronically, the main weakness is that due to a paucity of texts with a single topic resulting from the criteria of data, the news texts selected are not confined to the same specific aspect of GBA construction (for example, economy reports), which leads to a situation of danger that it seems to compare apples with oranges, though the findings are very important and revealing in their own right.Therefore, further research might explore news texts with the same orientation.Meanwhile, this study only involves the news texts from two phases of GBA construction drawing on a single perspective, which is rather inadequate in revealing a full picture of the complexity and dynamism of news discourse.As such, additional perspectives such as the other two domains in the appraisal system (attitude and gradation) can be drawn on in tandem in order to gain a better understanding of the news discourse of the Greater Bay Area as a whole.Another suggestion is that a larger news corpus covering more timespan of this national project be examined to improve the quality of a diachronic study on the English news text on GBA.This study can also be improved by multiple coders and the calculation of inter-rater reliability.

Figure 2 .
Figure 2. Visualization of the percentage of sub-categories of engagement in two phases.