Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 210 Online version available in : http://arbitrer.fib.unand.ac.id | 2339-1162 (Print) | 2550-1011 (Online) | JURNAL ARBITRER Submission Track A B S T R A C T Received: August 31, 2024 Final Revision: June 02, 2025 Accepted: June 05, 2025 Available Online: June 25, 2025 This study aims to shed light on how the combination of cohesive devices (henceforth CDs) and narrative elements (NEs) contributes to developing the coherence and clarity of administrative texts (ATs). While CDs and NEs were constantly treated as separate elements in previous studies, this research delves into their interaction as complementary strategies within ATs. Quantitative content analysis was employed to analyze 20 Vietnamese ATs, the results show that when CDs (reference, conjunction, ellipsis, and substitution) are combined with NEs (processes, participants, and spatial-situational settings), they occupy a crucial role in constructing textual cohesion, thematic continuity, and systematic transparency. These compound linguistic elements improve reading competence. As a result, readers can retain information more effectively, and the formal tone of administrative texts becomes more reinforced. Moreover, the results suggest that merging structural and experiential meanings in ATs significantly enhances both clarity and reader engagement. Drawing upon systemic functional linguistics as its theoretical framework, this study improves both theoretical interpretation within discourse studies and practical implications for developing ATs. Building on these findings, this study develops an analytical framework for examining how the combination of NEs and CDs operates to achieve structural objectives. Future research can apply this model to other text types, which are characterized by clarity, integrity, and formal structure. In short, this research demonstrates the roles of CDs and NEs in the construction of coherent and accessible ATs. Keywords Cohesive devices, narrative elements, administrative texts, textual coherence Correspondence *E-mail: lntoan@hcmulaw.edu.vn Article Enhancing Administrative Texts Through Cohesive Devices and Narrative Elements Ly Ngoc Toan Department of legal languages - Ho Chi Minh City University of Law, Vietnam Under License of Creative Commons Attribution-Non Commercial 4.0 International.DOI: https://doi.org/10.25077/ar.12.2.210-226.2025 I. INTRODUCTION ATs vary in style because they communicate factual and objective information that readers will access and refer to (Bauman & Briggs, 1990; Espeland, 1993; Zinner, 2011). These texts inherently represent an objective and formal style, which give priority clarity over personal engagement. However, this traditional style has increasingly been reconsidered in light of linguistic research. Recent studies demonstrate that integrating NEs with CDs can significantly improve the quality of texts (Halliday and Hasan, 1976; Taufan, 2021; De Silva & Henderson, 2007; Boyd, Blackburn, & Pennebaker, 2020). In particular, NEs comprise temporal sequences, participant roles, and spatial- situational settings, which serve to reframe abstract propositions into textual forms to reflect standard procedures and communicative goal. Manoharan and Rangarajan (2022) highlight that effective use of NEs would create three positive outcomes: it makes information more relevant to readers; allows readers to engages with information on a deeper level during reading; and boost readers’ ability to remember information from long and complex texts. Moreover, besides improving referential clarity, this approach promotes interpersonal engagement and simplifies complex content to reduce misunderstanding (Chen & Meurers, 2016; Martin, Quiroz, & Wang, 2023; Kumar, 2022) Making sense of the importance of this approach, this study was conducted with two specific purposes. First, it addresses limitations in previous research, in Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 211 which CDs and NEs have typically been studied as separate, independent elements with little attention paid to how they work together. Halliday & Hasan (1976) emphasized that CDs play a crucial role in constructing textual coherence and logical progression. Meanwhile, NEs provide effective reading strategies for readers (Boyd et al., 2020; Law & Matthiessen, 2023). Therefore, the integration of CDs and NEs holds potential benefits for both writers and readers of ATs. However, in practice, this interaction has received limited scholarly attention. To date, no study has thoroughly examined the combination of these two elements in ATs. Haim (2023) and Megerdoomian et al. (2019) have highlighted this gap by pointing out that, although CDs contribute significantly to the structural organization of texts and enhance reader engagement, most studies overlook their interaction with NEs. This limitation has led to significant constraints in analyzing and understanding ATs, particularly in failing to fully recognize the roles of CDs and NEs in textual construction. As a result, readers often encounter difficulties in interpreting and accessing complex and dense ATs, while writers struggle to produce documents that are as clear and coherent as intended. More specifically, the lack of attention to the interplay between CDs and NEs has prevented a comprehensive grasp of how these linguistic resources jointly contribute to reader comprehension and the projection of institutional authority (Herzog & Claunch, 1997; Zinner, 2011; Hermawan, Rohman, & Saputra, 2022). As institutional communication increasingly demands clarity, transparency, and engagement, it becomes essential to investigate how these elements operate in tandem rather than in isolation. A second issue lies in the limited recognition of the benefits that may arise from integrating these two elements in institutional writing. First, it is crucial to examine the role of CDs, which function as essential tools to maintain cohesive properties among the sentences. In discourse analysis, these cohesive properties are constructed by the means of grammatical and lexical cohesion. The grammatical cohesion comprises reference, conjunction, substitution, and ellipsis, which are used to maintain textual continuity and ensure coherent progression across discourse (Halliday & Hasan, 1976; Alyousef, 2016; Jahja, 2021). In contrast, lexical cohesion is associated with repetition, synonymy, antonymy, and collocation, which contribute to semantic consistency and reinforce thematic unity throughout the text. Thus, CDs have been extensively examined in recent studies, particularly in discourse analysis and critical discourse analysis. Rahman, Firdaus, and Ardiansyah (2023) argue that CDs enhance readers’ comprehensibility by improving textual coherence and reducing the mental effort required to decode meaning. Extending this perspective, Kadhim (2024) emphasized that CDs play an important role in shaping thematic progression because they guide the reader’s attention through coherent idea development and ensure continuity across discourse segments. Second, given the contributions of CDs to enhancing the effectiveness of ATs, as discussed above, NEs also play a crucial role in guiding readers through the text. Jopek-Bosiacka (2023) argues that NEs, together with narrative strategies, contribute to clarifying complex texts, and enabling readers to grasp the internal logic more easily and effectively. Building on this foundation, Chen and Meurers (2016) assert that when texts exhibit clear structural organization, they facilitate cognitive processing and enhance memory retention, thereby improving readers’ reading comprehension. Extending the communicative value of NEs, Martin, Quiroz, and Wang (2023) emphasize that NEs offer a contextual foundation that readers from the public domain can identify with. This foundation helps shift abstract policy-related information away from inaccessible reasoning and toward practical understanding (Hermawan, Rohman, & Saputra, 2022; Kumar, 2022). The theoretical orientation of this study draws upon systemic functional linguistics (Halliday & Hasan, 1976; Martin et al., 2023), which argue that language has two main functions: organizing text and representing experience. Building on this theoretical distinction, CDs contribute structural meaning by organizing the logical and referring relationships between textual units to ensure coherence and textual integrity. In contrast, NEs contribute experiential meaning by encoding participants, processes, and contexts that represent actual actions and situations. These linguistic functions have broader implications in policy discourse. Espeland (1993) analyzed how policy texts obscure the social consequences of policy decisions, while Fobbe (2021) explored how language in policy texts shapes processes of public opinion formation. The theory of cognitive process (Chen & Meurers, 2016) further supports this view. It suggests that clear and standard texts reduce mental effort and improve readers’ memory retention. Grounded in this theoretical framework, three research objectives are set up in this study. First, this study aims to examine Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 212 how CDs and NEs interact to frame concepts of clarity, coherence, and engagement in ATs. Second, it determines the optimal usage of these elements, or the optimal combination of these elements, for text coherence and reader interest. Last but not least, it establishes a set of guidelines on how to better the drafting of ATs through actual experience. As a result, these goals directly contribute to filling this research gap identified in previous studies that rarely consider the interaction of these two elements. To fulfill these objectives, the study adopts a mixed-methods design, combining quantitative content analysis (De Sordi et al., 2016; Kreines & Kreines, 2020) with qualitative discourse analysis. The data comprise various types of ATs, including institutional reports, guidelines, and legal texts. The quantitative analysis identifies the frequency and forms of CDs and NEs, while the qualitative component explores how these devices enhance reading competence and reader engagement. Rather than operating independently, CDs and NEs function interdependently: CDs maintain referential clarity and logical flow, whereas NEs structure events and assign roles to institutional actors. Such integration of analytical methods not only provides robust findings but supports practical implications as well. This practical application can enhance communication strategies and inform the development of more effective educational materials. Ultimately, this study combines theoretical insights with applied analysis. The aim is to bridge the gap between frameworks and implementation, thereby fostering a deeper understanding of audience engagement. II. METHOD Methodological principles Both quantitative content and qualitative discourse analysis are employed to examine the role of CDs and NEs in ATs. Each method serves a complementary purpose in addressing the multifaceted nature of the research objectives. The quantitative analysis served three goals: (i) locating the recurrence of certain language features, (ii) categorizing them by their functions, such as reference, substitution, or narrative process, and (iii) plotting them in the genres and institutions. The quantitative method was conducted with the view of recovering formal similarities and structural patterns through systematic analysis of the grammatical and narrative aspects of textuality in a language. On the other hand, the quantitative analysis focuses on form and frequency, the qualitative approach aims to: (i) explore how CDs and NEs operate in real communicative contexts, (ii) analyze how they interact to produce coherence, and (iii) evaluate their rhetorical effectiveness in engaging readers and minimizing ambiguity. Data collection The data consist of 20 Vietnamese ATs, which were collected from legal sources through Vietnam’s National Legal Website (https://thuvienphapluat. vn). They were issued by central and local bodies such as the Prime Minister, ministries, and provincial committees, including the Prime Minister (PM), Government Office (GO), Ministry of Finance (MF), Ministry of Health (MH), Ministry of Education and Training (MET), Ministry of Transport (MT), Provincial People’s Committees (PPC), and Vietnam Social Security (VSS). These texts selected on the basis of organizational diversity and representativeness, hold legal validity and exemplify current standards of Vietnamese ATs. Each text adopts a formal tone, follows an expository or structured format, and is composed in a monologic and impersonal style. In terms of linguistic features, structural elements include legal openings, third-person impersonal constructions, and a high degree of noun transformation. Topics of these texts cover financial plans, infrastructure development, policy execution, organization realignment, and public- service regulation. These texts were selected due to their authenticity, diversity in content, diversity of issuing bodies, and overall suitability |for analyzing linguistic coherence. Moreover, these texts provide a consistent and formal genre suitable for linguistic analysis of CDs and NEs in Vietnamese ATs. Table 1. Classification of administrative texts by issuing authority Code Issuing authority Document titles PM Prime Minister VB11, VB12, VB13 GO Government Office VB4, VB10, VB14 MF Ministry of Finance VB1, VB7, VB15 MH Ministry of Health VB2 MET Ministry of Education and Training VB19 MT Ministry of Transport VB6 VSS Vietnam Social Security VB5 PPC Provincial People’s Committees VB3, VB8, VB16, VB17, VB18, VB20 Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 213 allows the study to examine not only the frequency but also the communicative functions of linguistic features. It provides a comprehensive account of how coherence and communicative clarity are achieved in Vietnamese ATs. III. RESULT AND DISCUSSION Cohesive devices in ATs This section analyzes the use of CDs in ATs to understand just how CDs underpin the clarity, continuity, and overall comprehensibility of discourse in institutional documents. More particularly, this study adopts a text-based analysis to examine how CDs are employed to constructed textual cohesion through the aspects of administrative language. CDs are categorized into four types: reference, substitution, ellipsis, and conjunction. Each type will be discussed in terms of its structural form, textual realization pattern, frequency (F), and influence on meaning. Types and functions CDs occupy a central position in improving continuity and achieving textual coherence. As shown in Table 2, the four main types of cohesive devices: reference, substitution, ellipsis, and conjunction are investigated in this section, which appear with varying frequency across the twenty texts. Conjunctions are the most common (22 instances), followed by reference (15), ellipsis (12), and substitution (8). Through this distribution, several noticeable properties and formal style of ATs are presented and explored in turn. An elaborate examination of each categories provides in-depth insights into the textual functions To facilitate systematic analysis, these texts are classified based on the issuing agencies. Each document is assigned a unique code composed of two capital letters representing the name of the issuing institution. This coding system ensures consistency in referencing and facilitates thematic and structural comparisons across the data. Data analysis The data were analyzed on the basis of systemic functional linguistics by Halliday and Hasan (1976). They divided language into three functions: ideational function, interpersonal function, and textual function. Within this theoretical framework, the study concentrates on two major aspects, which are CDs to realize textual meaning and NEs to reflect experiential ones. The research has conducted through content analysis to identify and categorize four kinds of CDs (reference, substitution, ellipsis, and conjunction). Then, NEs are divided into three groups: processes, participants, and adverbial components. From this categorization, discourse analysis is used to discover how CDs and NEs are integrated to gain communicative objectives. This analysis considers how CDs bind with NEs to keep a logical continuity, which helps reduce ambiguity and involve readers emotionally. In addition, the research examines how agents, actions, and temporal or legal conditions connect with CDs to make texts clearer. The connections between CDs and NEs are shed light on through the application of the matrix coding and thematic. Particularly, matrix coding serves to organize linguistic patterns, while thematic mapping is used to identify semantic aspects across discourse. This integrated approach Table 2. Types and functions of CDs in ATs Type Purpose Linguistic expressions Texts F Reference Maintain continuity, avoid repetition • Personal pronouns (he, she, it, they). • Demonstratives (this, that, these, those). • Comparative references (such as, similarly). VB2-MH, VB6-MT VB9-PM 15 Substitution Replace noun/verb/clause elements • Nominal substitution (one, ones). • Verbal substitution (do, does, did). • Clausal substitution (so, not). VB9-PM, VB10-GO 8 Ellipsis Omit recoverable elements • Nominal ellipsis (omission in noun phrases). • Verbal ellipsis (omission in verb phrases). • Clausal ellipsis (omission in clauses). VB11-PM, VB12-PM 12 Conjunction Connect clauses logically • Additive (and, furthermore, also). • Adversative (but, however, on the other hand). • Causal (because, therefore, thus). • Temporal (when, before, after). VB3-PPC, VB4-GO, VB7-MF 22 Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 214 of CDs. To begin with the analysis of reference, the results show that the frequency of CDs stands at the second highest across the texts, with 15 instances found in texts such as VB2-MH, VB6-MT, and VB9-PM. There are three types of reference, which were observed, such as personal, demonstrative, and comparative. First, personal reference is recognized through the use of pronoun it in order to replace bộ (the ministry), and to maintain continuity. Demonstrative reference occurs in the same document with những chỉ thị này – these, referring to các chỉ thị mới – new directives. In VB9-PM, comparative reference appears through the phrase tương tự như chính sách trước đó – similar to the previous one, which connects new regulations to prior ones for coherence. Second, substitution was found with lower frequency, totaling 8 instances, but still present in essential parts of ATs, notably in VB9-PM and VB10-GO. The results demonstrate that there were three types of substitution, commonly used in ATs, including nominal, verbal and clausal substitution. Firstly, nominal substitution is illustrated in the use of noun phrase chính sách trước đó (the previous policy), which is used to replace one of other policies that is mentioned previously. Second, verbal substitution appears in VB10-GO with the verb phrase đã làm (was done), where this verb can replace the whole process of previous action. Next, clausal substitution occurs in VB9-PM with the expression nếu không (if not), which stands in for the conditional clause nếu điều kiện được đáp ứng – if the conditions are met. This use of substitution shows the tendency of ATs to condense lengthy conditional structures while preserving logical sequencing and procedural clarity. Returning to ellipsis which was found 12 times and contributed to benefits in terms of syntax by omitting information recoverable from context. Similar to substitution, ellipsis is also divided into three categories. First, nominal ellipsis is found in VB11-PM where báo cáo (report) is omitted in the second clause of a parallel structure. Second, verbal ellipsis is present in VB10-GO, where đã làm (was done) is omitted and interpreted based on a preceding clause. Finally, clausal ellipsis was not directly observed in the corpus. Instead, documents such as VB16-PPC often repeat phrases like kế hoạch này (this plan) to maintain clarity and avoid ambiguity. Overall, ellipsis demonstrates how Vietnamese ATs maintain brevity without sacrificing clarity. Last but not least, conjunctions were found the most frequent devices, which appear 22 times in VB3-PPC, VB4-GO, and VB7-MF. Common conjunctions include additive và (and), causal vì (because), temporal trước khi (before), and adversative nhưng (but). These forms consistently contribute to organizing information flow and reinforcing coherence in ATs. Textual realizations In administrative texts, CDs appear in identifiable categories and take form through specific linguistic realizations that serve text- level functions. These realizations are not random, and reflect the formal, ordered, and regulated style typical of official writing. Table 3 shows that how CDs are recognized through linguistic elements, which serve to ensure referring clarity, maintain logical relations, and reduce unnecessary repetition. Each type manifests through distinctive grammatical forms to reinforce procedural Table 3. Textual realizations of cohesive devices Cohesive device Function Linguistic expressions F Texts Reference Establish continuity and track entities across clauses and sentences personal pronouns, demonstratives, comparative phrases 15 VB2-MH, VB6-MT, VB9-PM Substitution Maintain cohesion by replacing previously stated elements nominal (one, ones), verbal (do, does), clausal (so, not) 8 VB9-PM, VB10-GO Ellipsis Promote text economy by omitting recoverable elements nominal ellipsis, verbal ellipsis, clausal ellipsis 12 VB11-PM, VB12-PM Conjunction Indicate logical relationships between clauses or sentences additive (and, also), adversative (but, however), causal (because, thus), temporal (before, after) 22 VB3-PPC, VB4-GO, VB7-MF, VB14-GO, VB16-PPC, VB18-PPC Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 215 meaning in a way that aligns with communicative goals of the documents. As shown in Table 3, reference, substitution, ellipsis, and conjunction function within sentence structures to maintain reference clarity, preserve logical connections, and avoid unnecessary repetition. Structurally, reference can be realized through three forms: personal pronouns anh ấy, cô ấy, nó, họ (he, she, it, they), demonstrative pronouns đây, đó, những này, những kia (this, that, these, those), and comparative constructions tương tự, bằng (similar to, as… as). In VB2- MH, personal reference is exemplified by nó (it), which replaces bộ (the ministry). Demonstrative reference is realized through những chỉ thị này (these directives), which refers to các chỉ thị mới (new directives). Comparative reference is found in VB9-PM with the phrase tương tự như chính sách trước đó (similar to the previous one), which connects current policy with an earlier one. These realizations help maintain continuity and reinforce organizational cohesion. Likewise, ellipsis functions as a cohesive device by omitting elements that are recoverable from context, which enhances textual cohesion. The result shows that ellipsis is observed in three forms: nominal, verbal, and clausal, with 12 occurrences found in the data. In VB11-PM, nominal ellipsis is seen in the omission of báo cáo (report), which can appear before the verb nộp (submitted). In VB12-PM, verbal ellipsis is demonstrated in the omission of phê duyệt (approved) following xem xét (reviewed). Clausal ellipsis is cited in VB11-PM through the structure nếu được chấp thuận (if approved), where the main clause is left unstated. However, it is important to note that, this case does not qualify as clausal ellipsis in the strict grammatical sense, as Vietnamese lacks auxiliary verbs and clause-recovering devices. The structure instead represents a reduced passive form dependent on contextual interpretation. In addition, conjunction is the most frequently used cohesive device, with 22 instances across the texts. It signals logical relationships through four subtypes: additive, adversative, causal, and temporal. In VB4-GO, additive cohesion is realized with và (and), which links two coordinated actions. In VB18-PPC, causal connection is expressed through vì (because), connecting the policy outcome to its justification. Temporal cohesion is found in VB16-PPC with trước khi (before), which orders actions according to procedural timelines. In VB14-GO, adversative relation is realized with nhưng (but), which marks a contrast between expected and actual administrative results. These realizations collectively demonstrate how conjunctions contribute to the structural and argumentative clarity of ATs. Narrative elements in ATs This section investigates NEs in Vietnamese ATs, which focus on how actions, participants, and goals are structured so as to intensify coherence and authority. In this section, narrative components such as sequencing, decision-making, and responsibility assignment are analyzed in terms of functional and lexical patterns. They are not fictional elements, but instead exist to arrange real organizational actions in a more logical and rational manner. By analyzing processes, actors, and settings, the study can further demonstrate how forms of narrative help express policy intent, ensure clarity, and keep track of procedural logic straight. Section 3.2.1 and 3.2.2 will explain the structure, frequency, and function of narrative realizations in administrative discourse. Narrative elements NEs function as discursive mechanisms that encode institutional operations into structured forms. These elements contribute to communicative organization and support the procedural transparency required in formal bureaucratic writing. As shown in Table 4, 39 NEs were identified, comprising processes (11 cases), participants (15), and circumstantial elements (13). These figures indicate the pervasive role of narrative structures in administrative discourse. Notably, temporal expressions appeared in 5 of the circumstantial cases, while government entities and departments were mentioned 10 times among participant references. This highlights the focus on time and institutional agency . These distributions indicate a stable narrative configuration that aligns with institutional norms, ensuring clarity, integrity , and sequential development. The processes identified include decision-making (5 instances), approval (2), implementation (3), and evaluation (1). This sequence reflects a typical administrative pattern: from decision to validation, implementation, and final assessment. Processes represent the central sequence Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 216 of institutional activities. These linguistic forms describe the key administrative steps, from initial decisions to final evaluations. Decision-making is seen in texts such as VB5-VSS and VB15-MF, where verbs like sẽ xem xét (would review) denote administrative initiation. In the case of approval, VB10-GO uses the verb thông qua (approve), which conveys formal endorsement by an authorized body. Implementation is exemplified in VB17-PPC through phải thực hiện (must implement), conveying obligation and procedural staging. Evaluation is found in VB20- PPC, in which sẽ được đánh giá (will be evaluated) emphasizes objective outcome assessment through passive construction. Participants act as institutional agents within these processes. They represent the entities responsible for enacting, approving, or responding to administrative directives. Government agencies are represented in VB13-PM and VB15-MF through phrases like Theo Nghị định của Bộ Giáo dục (According to the Ministry of Education Decree), which signal legal authority. Departments are featured in VB4-GO through Phòng Công nghệ thông tin (IT Department) and chịu trách nhiệm thực hiện (is responsible for implementation), assigning operational roles. Decision-makers are identified in VB18-PPC with Hội đồng phê duyệt (The Board approves), which represent centralized authority. In VB3-PPC, stakeholders appear through references to các tổ chức xã hội (social organizations), whose feedback sẽ được tổng hợp (will be incorporated). This construction highlights participatory input while preserving the text’s formal tone. Circumstantial elements situate processes in relation to time, location, legal framework, or contextual conditions. They serve to contextualize actions and clarify institutional conditions under which decisions are made. In VB16-PPC, temporal framing appears in the phrase trước khi năm tài chính kết thúc (before the end of the fiscal year), offering a precise timeframe for policy implementation. In VB1-MF with trong phạm vi toàn quốc (nationwide), spatial reference is seen with the view of defining policy scope. Legal foundation occurs in VB8-PPC through theo điều 5 Nghị định 2021/30 (under Article 5 of Regulation 2021/30), which lends normative force. As identified in VB5-VSS with trong trường hợp khẩn cấp (in case of emergency), situational context can frame procedural flexibility within institutional control. Lexical patterns In ATs, the lexical patterns used to express narrative elements reveal institutional priorities such as clarity, authority, and procedural precision. These elements are encoded structurally and patterned lexically through deliberate and recurrent choices. As displayed in Table 5, the lexical patterns of narrative elements are grouped into two categories: process-related verbs and participant- defining nouns. Each category fulfills distinct communicative functions, but they together constitute the narrative structure that is central to ATs. Process-related verbs are composed of four dominant lexical items: thực thi (implement), Table 4. Narrative components in ATs Component Subtype Function F Texts Process Decision-making Direct actions 5 VB5-VSS, VB7-MF, VB15-MF Approval Endorse actions 2 VB10-GO Implementation Carry out decisions 3 VB17-PPC, VB18-PPC Evaluation Assess outcomes 1 VB20-PPC Participants Gov. agencies Represent authority 5 VB13-PM, VB15-MF Departments Implement policies 5 VB4-GO, VB10-GO Decision-makers Make decisions 3 VB5-VSS, VB18-PPC Stakeholders Provide input 2 VB3-PPC, VB17-PPC Circumstances Temporal Situate time 6 VB5-VSS, VB16-PPC Location Indicate scope 3 VB1-MF Legal context Frame by law 2 VB16-PPC, VB8-PPC Situational Define conditions 4 VB5-VSS, VB17-PPC Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 217 quyết định (decide), phê duyệt (approve), and đánh giá (evaluate). These verbs form the core of ATs. They report institutional actions, express institutional intent, assign responsibility, and sequence procedures over time. The verb triển khai (implement) appears most frequently, with 91 instances across texts such as VB5-VSS, VB7-MF, and VB16-PPC. In VB7-MF, the clause Các đơn vị trực thuộc triển khai dự toán ngân sách theo hướng dẫn mới – Subordinate units implement the budget estimates according to the new guidelines contains the verb triển khai, which signals institutional responsibility and procedural command. This shows how implementation verbs act as formal instructions embedded within the broader structure of administrative sentences. The verb quyết định (decide), which is observed 78 times, functions as a text marker for initiating top-down procedures. In VB1-MF, Bộ quyết định phân bổ ngân sách cho các đơn vị dự toán – The ministry decides on budget allocation for dependent units, the verb quyết định establishes institutional authority and opens the procedural chain. Similarly, phê duyệt (approve), found 64 times, formalizes prior decisions and prepares for execution. In VB3- PPC, the expression UBND tỉnh phê duyệt đề án xây dựng nông thôn mới – The provincial People’s Committee approves the rural development plan uses the verb phê duyệt to signal institutional endorsement and legitimize action. The final verb đánh giá (evaluate), seen with 52 times, marks the closing phase of institutional processes. In VB14- GO, Hiệu quả thực hiện sẽ được đánh giá vào cuối năm tài chính – The effectiveness of implementation will be evaluated at the end of the fiscal year, the passive form sẽ được đánh giá emphasizes the outcome and maintains administrative impartiality. Parallel to these verbs, participant-defining nouns also reflect functional roles and institutional positions. These nouns identify who performs or authorizes the actions and how institutional agency is distributed. The term cơ quan (agency) is the most frequent, found 95 times, as seen in expressions such as cơ quan ban hành (issuing agency) and cơ quan quản lý (managing agency), which designate the origin and oversight of policies. In VB1-MF and VB6-MT, the recurrence of cơ quan indicates its centrality in authorizing and supervising administrative action. The noun phòng or đơn vị (department) is also prominent, with 84 instances. In VB12-PM, the clause Phòng kế hoạch tài chính hướng dẫn quy trình giải ngân (the finance planning department provides guidance on the disbursement process) positions the department as a mid-level entity responsible for implementation. The lexical pattern các bên liên quan (stakeholders), recorded 56 times, introduces participatory dimensions. In VB6-MT, Các bên liên quan đóng góp ý kiến vào quy hoạch giao thông đô thị – Stakeholders provide input into the urban transport plan, the phrase các bên liên quan functions as an external contributor, while đóng góp reflects participatory interaction in shaping administrative decisions. Lastly, the noun thẩm quyền (authority) appears 48 times. In VB20-PPC, the expression Cơ quan có thẩm quyền xem xét và xử lý hồ sơ – The competent authority reviews and processes the dossier includes the phrase có thẩm quyền, which denotes institutional legitimacy. The verb pair xem xét và xử lý (review and process) reflects both evaluation and execution. This dual function shows how institutional authority is compactly expressed in a single verbal structure. Functional characteristics of administrative language ATs are characterized by their obligation to express institutional authority, procedural clarity, Table 5. Lexical patterns of narrative elements Category Lexical item Function F Texts Process verbs Implement Execute policies 91 VB5-VSS, VB7-MF, VB16-PPC Decide Initiate action 78 VB1-MF, VB7-MF, VB12-PM Approve Endorse proposals 64 VB3-PPC, VB19-MET Evaluate Measure outcomes 42 VB10-GO, VB14-GO Participants Agency Policy originator 95 VB1-MF, VB4-GO, VB6-MT, VB18-PPC Department Executor 84 VB3-PPC, VB5-VSS, VB7-MF, VB12-PM Stakeholder Input provider 56 VB6-MT, VB10-GO Authority Regulator 48 VB2-MH, VB11-PM, VB20-PPC Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 218 and formal consistency. As shown in Table 6, six key features characterize administrative texts: formal tone, nominalization, modal expressions, sequencing, referring opacity, and information density. These features were identified a total of seventy-eight times, and indicate their central role in structuring how ATs are realized. Each characteristic serves as a stylistic choice and a functional necessity that ensures the effectiveness of institutional communication. To begin with, a formal tone is maintained throughout administrative writing to reinforce authority and institutional impersonality. This tone is achieved through the use of third-person references, institutional terminology, and fixed expressions. Phrases such as theo quy định (according to regulations), căn cứ vào nghị định (based on the decree), and cơ quan có thẩm quyền (authorized agency) exemplify this style. With 18 occurrences found in documents such as VB1- MF and VB7-MF, these expressions distance the speaker from the message and construct the text as an objective voice of the institution. By eliminating subjectivity, this tone promotes neutrality, legitimizes policy, and minimizes personal interpretation of administrative directives. By eliminating subjectivity, this tone promotes neutrality, legitimizes policy, and minimizes personal interpretation of administrative directives. Closely related to formality is the use of nominalization, which occurs 14 times in texts such as VB11-PM and VB7-MF. This strategy transforms actions into abstract entities, often via phrases like việc thực hiện (implementation), sự điều chỉnh (adjustment), or quy trình phê duyệt (approval process). In VB11-PM, the phrase việc đánh giá hiệu quả (effective assessment) replaces explicit agents with an abstract process. This allows complex information to be condensed and integrated into larger sentence structures. This technique condenses complex information into compact sentence structures, reinforcing the abstract and dense nature of administrative language. As a result, procedures are framed as impersonal systems rather than individual actions. In addition to form and abstraction, modality plays a key role in expressing institutional force. Modality markers appear in 10 instances, especially in VB2-MH and VB17-PPC. These include verbs such as phải (must), sẽ (will), and cần (need), which express institutional obligation and legal force. For example, the sentence Các đơn vị phải thực hiện đúng tiến độ – Units must follow the timeline employs phải to enforce policy compliance. These modal verbs not only express institutional force but also guide the reader toward expected behavior within the administrative framework. Their recurrence reflects how obligation is embedded into procedural language to secure consistent implementation across hierarchical levels. Another key feature is the explicit structuring of processes through sequencing and hierarchy. In 9 instances, as seen in VB3-PPC and VB7-MF, these patterns help organize content according to institutional procedure. Temporal adverbs and connectors such as trước hết (first of all), sau đó (after that), and cuối cùng (finally) are used to break processes into discrete stages. These markers create a procedural roadmap while reflecting the internal logic of institutional workflows. The hierarchical dimension further clarifies the chain of responsibility by indicating who initiates, who validates, and who executes each step- in policy enactment. This contributes to procedural transparency and administrative integrity . Another stylistic strategy that reinforces impersonality is referring opacity, identified 11 times in documents such as VB10-GO and VB12-PM. This feature is primarily realized through passive constructions such as được ban hành (to be issued) Table 6. Functional characteristics of administrative language Feature Function F Texts Formal tone Convey authority through impersonal expressions 18 VB1-MF, VB4-GO, VB7-MF Nominalization Increase abstraction and density 14 VB1-MF, VB7-MF, VB11-PM Modality markers Express obligation and institutional force 10 VB2-MH, VB4-GO, VB17-PPC Sequencing & hierarchy Indicate procedural order and logic 9 VB3-PPC, VB7-MF Referring opacity Obscure agency via passive structures 11 VB4-GO, VB10-GO, VB12-PM Dense packaging Combine multiple ideas in complex sentences 16 VB5-VSS, VB7-MF, VB14-GO Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 219 and sẽ được thực hiện (will be implemented). By focusing on actions and outcomes rather than actors, these structures obscure agency by focusing attention on the action or outcome rather than on the actor. This strategy is not stylistically incidental but serves to maintain institutional distance and diffuse individual responsibility. By emphasizing the process, itself, the text sustains a formal tone and promotes objectivity in how actions are described and justified. Finally, dense packaging, identified 16 times across documents like VB5-VSS and VB14-GO, refers to the compression of multiple ideas into single, complex sentences. These constructions often include embedded clauses, conditionals, and justifications, as in the combination of nghĩa vụ, thời hạn và trường hợp ngoại lệ (obligations, deadlines, and exceptions) within one sentence. This compactness permits multiple legal and procedural requirements to coexist within a cohesive structure. Although such density may reduce accessibility, it enhances interpretive precision and ensures the completeness of administrative directives. Interaction between CDs and NEs This section examines the interaction between DCs and NEs in Vietnamese administrative texts, which is involved in how this integration contributes to textual cohesion. CDs establish grammatical links and ensure logical consistency. In contrast, NEs introduce time structure, clarify roles, and explain causality, thereby making administrative texts more transparent and easier to understand. This analysis involves two layers of organization that coordinate abstract policies and concrete actions. This section explores three key interactions: how reference maintains thematic continuity across sentence boundaries, how conjunctions convey cause and process, and how these cohesive tools help construct institutional actions step by step. Furthermore, the combination of these features makes ATs easier for laypeople to read in comparison with formal documents, which are designed for government institutionals. Textual pattern and matrix coding The interplay between CDs and NEs in ATs reflects a codified discursive strategy rather than a coincidental occurrence. As shown in Table 7, this interaction manifests through three interdependent textual configurations: narrative construction, cohesive linking, and matrix coding. Each pattern functions at a distinct level of organization. Together, they form a coherent system that allows policy texts to regulate agency, organize procedures, and legitimize institutional decisions. To begin with, narrative construction is oriented toward sequencing administrative acts, articulating rationales, projecting intended outcomes, and assigning institutional roles. Forty-four occurrences were identified in which policy-related procedures are ordered to reflect organizational logic. A representative case is found in VB5-VSS: Trước hết, xem xét đề xuất; sau đó, trình phê duyệt; cuối cùng, thực hiện – First, review the proposal; then, submit for approval; finally, implement. In this case, the use of adverbial markers trước hết, sau đó, and cuối cùng frames the action sequence within a canonical progression. This configuration also facilitates explanatory coherence, with seventeen instances in which decision-making is paired with justification. As seen in VB9-PM: Do ảnh hưởng dịch bệnh, ngân sách bị điều chỉnh – Due to the epidemic, the budget was revised, the clause do ảnh Table 7. Textual pattern and matrix coding Aspect Function F Texts Narrative construction Structure actions and highlight intent 44 VB4-GO, VB5-VSS, VB7-MF, VB15-PPC Explain reasons behind decisions 17 VB7-MF, VB9-PM Express expected outcomes 11 VB2-MH, VB15-MF Assign roles to institutional actors 15 VB4-GO, VB17-PPC Cohesive linking Ensure continuity and coherence 38 VB2-MH, VB4-GO, VB6-MT Avoid repetition 24 VB10-GO, VB12-PM Maintain logical progression 40 VB7-MF, VB10-GO, VB18-PPC Matrix coding Link actions to temporal context 98 VB3-PPC, VB6-MT, VB8-PPC Clarify referents of institutional actors 123 VB1-BTC, VB18-PPC, VB20-PPC Ground actions in legal reasoning 87 VB6-MT, VB10-GO, VB14-PPC Integrate time, actors, and legal conditions 32 VB4-GO, VB7-MF Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 220 hưởng dịch bệnh encodes causality. This pattern also expresses intended goals, as seen in VB15- MF: Nhằm nâng cao hiệu quả quản lý (To improve management efficiency). In this case, the intended outcome is explicitly stated through lexical choice. Finally, fifteen occurrences demonstrate role assignment, as in VB17-PPC: Bộ chỉ đạo, các đơn vị triển khai, địa phương báo cáo – The ministry directs, departments implement, localities report. This mapping of institutional agents ensures that responsibility and authority are textually distributed along administrative hierarchies. This mapping of institutional agents ensures that responsibility and authority are textually distributed along administrative hierarchies. Cohesive linking, as the second pattern, ensures referring continuity, structural economy, and logical sequencing. Thirty-eight instances illustrate how cohesion is achieved through referring maintenance, such as in VB2-MH: Bộ sẽ triển khai kế hoạch. Sau đó, [bộ] sẽ giám sát tiến độ – The ministry will implement the plan. Then, it will monitor the progress. The repetition of bộ functions not as redundancy but as referring anchoring across clauses. Clausal substitution also reinforces this mechanism. For example, VB9-PM states: Nếu điều kiện được đáp ứng thì sẽ duyệt; nếu không thì từ chối (If the conditions are met, approval will be granted; if not, rejection). The phrase nếu không substitutes for the entire conditional clause, preserving syntactic efficiency while ensuring semantic transparency. Logical sequencing is the third subfunction, observed forty times. In VB18-PPC, the construction Kế hoạch bị hoãn [vì] không đủ ngân sách – The plan was delayed because of insufficient funds employs the causal connector vì to explicitly justify institutional outcomes. Such linking ensures that policy texts preserve argumentative cohesion and procedural legitimacy through overt rhetorical coordination. Such linking ensures that policy texts preserve argumentative cohesion and procedural legitimacy through overt rhetorical coordination. Matrix coding constitutes the third and most complex textual configuration, integrating narrative progression and cohesion within regulatory, temporal, and institutional frameworks. It is primarily realized by anchoring actions in time, specifying agents, and encoding legal grounds. Ninety-eight instances involve temporal integration, as in VB8-PPC: sẽ thực hiện trong 30 ngày – will be implemented within 30 days, where temporal expressions define operational timelines. The most frequent function is referent clarification, with 123 cases. In VB1-BTC, referents are reiterated through chains such as Cơ quan ban hành and Cơ quan này chịu trách nhiệm – The issuing agency and This agency is responsible, enabling consistent referent tracking. Matrix coding also performs legal grounding, found in eighty-seven cases, as illustrated in VB14-PPC: Theo điều 5 của Nghị định – According to Article 5 of the Decree. This alignment embeds textual action within legal frameworks, framing institutional decisions as regulatory obligations. Finally, 32 instances show that matrix coding permits the integration of time, actors, and legal justification into unified structures. In VB7-MF, these dimensions co-occur in complex sentence forms that compress institutional assignment, legal reference, and procedural timing, all within a single syntactic unit. This multidimensional encoding achieves high informational density while preserving clarity and interpretive accessibility, which fulfills the communicative demands of formal governance texts. Thematic mapping and communicative function Thematic mapping in administrative discourse reveals how communicative intent is lexically and structurally distributed across recurring semantic configurations. In Vietnamese policy texts, thematic clusters group related propositions and guide readers through institutional priorities, action types, and participation frameworks. Table 8 identifies three dominant clusters: authority and decision-making (58), implementation and coordination (71), and stakeholder engagement (41). Each cluster performs a distinct communicative function. When employed consistently, these thematic structures serve to legitimize institutional power, operationalize procedures, and incorporate consultative input. This reinforces the multidimensional functionality of ATs. The first cluster, authority and decision- making, foregrounds institutional control and the articulation of formal mandates. This cluster is observed fifty-eight times in documents such as VB5-VSS, VB10-GO, and VB15-MF. It is primarily realized through declarative sentence types and performative verbs that convey institutional authority. In VB10-GO, the clause Thủ Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 221 Table 8. Thematic clusters in ATs Thematic cluster Function F Texts Authority and decision- making Express institutional authority and policy decisions 58 VB5-VSS, VB10-GO, VB15-MF Implementation and coordination Direct policy execution and manage inter-departmental actions 71 VB2-MH, VB15-MF, VB18-PPC Stakeholder engagement Incorporate external input and participatory evaluation 41 VB8-PPC, VB13-MP, VB1-MF tướng quyết định giao dự toán ngân sách cho các bộ, ngành (The Prime Minister decides to allocate the budget estimates to ministries and sectors) shows how the verb quyết định (decide) performs and legitimizes authority at the same time. These expressions are typically embedded within tightly structured syntactic constructions, which contribute to the formal tone and institutional assertiveness of the text. Through such configurations, the cluster encodes the hierarchical positioning of the speaker and reflects the central role of state actors in directing and authorizing public policy. Through such configurations, the cluster encodes the hierarchical positioning of the speaker and reflects the central role of state actors in directing and authorizing public policy. Building upon this institutional foundation, implementation and coordination represent the most frequent pattern in the data, with 71 occurrences found in texts such as VB2-MH, VB15-MF, and VB18-PPC. This thematic group is marked by directive sentence structures and modal verbs like phải (must) and cần (need), which express institutional requirements and maintain procedural consistency. In VB18-PPC, the clause Các đơn vị phải thực hiện phân bổ kinh phí đúng thời hạn – Units must carry out the budget allocation on time exemplifies the convergence of command and temporal precision. The combination of modals with institutional actors serves a dual function: it assigns responsibility while simultaneously articulating the expected course of action. This structure prescribes what needs to be done and encodes when it should be executed, by whom, and under what procedural condition. Thus, this cluster enhances the enforceability of directives while preserving syntactic clarity and communicative order. Extending beyond internal directives, stakeholder engagement highlights participatory orientation within administrative discourse. This pattern appears 41 times in the dataset, notably in VB8-PPC, VB13-MP, and VB1-MF. Its primary function is to integrate feedback from non- governmental or external entities into the policy framework. In VB8-PPC, the sentence Ý kiến từ các tổ chức xã hội sẽ được tổng hợp vào báo cáo – Feedback from social organizations will be incorporated into the reports demonstrates this function. The use of the passive construction sẽ được tổng hợp (will be incorporated) allows the text to maintain institutional detachment while still signaling openness to input. This passive form reinforces formality, whereas the inclusion of các tổ chức xã hội (social organizations) indexes external involvement. This balance between centralized authority and consultative legitimacy enables the institution to frame participation as a managed, yet acknowledged, dimension of governance. In doing so, stakeholder input is textually recontextualized into a codified part of administrative operations without undermining the formal decision-making hierarchy. IV. DISCUSSION Linguistic strategies for institutional coherence The analysis of CDs in Vietnamese ATs demonstrates that these devices play an indispensable role in constructing textual unity, procedural order, and institutional authority. These devices, such as referring tracking, conjunctions, and lexical reiteration, do more than serve as surface-level connectors. They also function as systemic strategies that reinforce the hierarchical and procedural nature of policy communication. Since Vietnamese is not an inflectional language, it lacks morphological agreement. This leads writers to depend more on discourse and pragmatic cues to ensure coherence. Halliday and Hasan (1976) point out that cohesion is expressed through grammatical and lexical relations that extend beyond the limits of individual sentences. However, the findings challenge this view because they show that cohesion in ATs is shaped by Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 222 structure and by context, pragmatic flow, and thematic development. Therefore, cohesion needs to be understood as a typological feature and a language-specific phenomenon, which is shaped by functional demands. When examined in light of previous studies, this study both supports and challenges the previous studies. Particularly, it partially corresponds to Rahman et al. (2023) and Kadhim (2024), who underscore the centrality of cohesive devices in enabling comprehension and information retention. Meanwhile these studies presuppose morphological richness as a main means of cohesion, but they are entirely appropriate in Vietnamese ATs. This study demonstrates how analytic languages like Vietnamese use repetition and pragmatic signals to achieve cohesion. These insights expand systemic functional linguistics and support rethinking cohesion theory from a cross-linguistic angle. Based on the findings, the study recommends that administrative training in Vietnam should emphasize awareness of text-level coherence rather than just syntactic accuracy. Coherence should be taught as a communicative goal, not merely a structural requirement. Nonetheless, certain limitations remain. The study focuses on a narrow range of institutional genres. Therefore, it does not account for how cohesive strategies may vary across spoken, informal, or digital forms of communication. As such, future research should examine cohesion in these emerging modes, as well as in cross-linguistic comparisons, in order to assess the theoretical elasticity and practical applicability of cohesion across typologically diverse administrative systems. To bridge these limitations, a comparative approach should be adopted in future research to delve into CDs in cross-linguistics ATs such as Chinese and Vietnamese or French and Vietnamese. It is widely acknowledged that this approach would enable a more comprehensive understanding of how cohesion functions in typological patterns of different languages, not least administrative styles. Furthermore, digital and spoken ATs are also essential to explore, particularly in online government websites, in which informalized styles and multimodal features may influence cohesive aspects. Therefore, these directions are expected to enhance cross-genre and cross-linguistic applicability and test the adaptability of systemic functional models to the evolving modes of public communication. Methodologically, temporal analysis of institutional language change may also shed light on how cohesion evolves in response to policy, technology, and user feedback. Structuring policy through narrative framing Complementing the cohesive perspective, the examinations of NEs in the data reveal that NEs are not merely a stylistic supplement, but a core mechanism for organizing systemic reasoning. To be more specific, ATs tend to reflect policy implementation as activity rather than abstract regulation through sequential processes, agentive structures, and circumstantial contextualization. The results prove that clause connection and discourse coherence of Vietnamese can be expressed through syntactic properties, which give rise to the mechanisms for enabling NEs to support process clarity. In this way, the view that NEs work against formal procedures is called into question by these results. More accurately, the findings reframe NEs as essential linguistic tools to enhance clarity and reinforce regulatory principles. This analysis harmonizes with and expands Jopek-Bosiacka’s (2024) arguments, which asserts that clarity and social engagement in ATs can be bettered by effective use of NEs. However, while Jopek-Bosiacka emphasizes reader relatability, the Vietnamese data suggests that narrative structuring also serves epistemic and institutional needs. Likewise, the findings confirm Martin, Quiroz, and Wang’s (2023) contention that narrative ordering aids reader cognition, particularly through temporal and causal sequencing. By realizing institutional mandates in action-oriented structures, Vietnamese ATs demonstrate how narration operates as a scaffolding device that guides both comprehension and execution. These insights extend the theoretical conversation by demonstrating how narrative logic intersects with administrative clarity in a non-Indo- European linguistic context. Despite its contributions, the study remains bounded by the topical limitation of its data, which concentrates on finance and planning documents. This raises questions about whether similar narrative strategies appear across domains such as education, healthcare, or environmental governance. Future research should investigate whether institutional narratives adapt to different policy logics and examine how these variations influence public reception and implementation efficacy. Further inquiry could deepen our understanding of how Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 223 narrative functions in institutional discourse and extend the relevance of this study’s findings to a broader range of administrative settings. In particular, comparative frameworks should be adopted to examine whether narrative framings differ in urgency, modality, or sequencing across policy areas. These domains might exhibit distinct narrative preferences depending on institutional timelines or audience expectations. A useful approach would be to combine textual analysis with interviews of policy drafters or institutional editors. This would help reveal how narrative strategies are intentionally used to align with policy objectives. A mixed-method approach would help validate the textual findings and clarify whether narrative structuring is intuitive or policy-driven across institutional settings. Integrated textual architecture in administrative discourse By combining both types of analysis, the study shows that the interaction between CDs and NEs in Vietnamese ATs creates a dynamic system. This system makes a considerable contribution to building structural precision and helping construct institutional meaning. Although CDs and NEs do not work in isolation, they operate interdependently to form cohesive texts when united. Particularly, CDs ensure clarity and continuity, while NEs provide time structure and define institutional roles. Vietnamese relies on serial verbs, topic- comment structures, and pragmatic links. These features show how analytic languages can bring cohesion and narrative elements together into a unified structure. This convergence challenges traditional assumptions that treat cohesion and narration as discrete domains and affirms their joint contribution to communicative functionality. This interpretation builds upon and diversifies the theoretical frameworks of Haim (2023) and Megerdoomian et al. (2019), who advocate for the integration of narrative sequencing and cohesive alignment in administrative and legal texts. While their research centers on Indo-European and Persian languages, the Vietnamese data offers a typologically distinct validation of their propositions. CDs and NEs can be merged through textual strategies rather than morphosyntactic mechanisms because Vietnamese does not belong to a derivational language. These strategies create a combination of formal constraint with interpretive accessibility, and enhance both the legitimacy and intelligibility of institutional communication. These findings support a subtler model of textual competence, which includes the ability to coordinate structural coherence with narrative logic. Nevertheless, this analysis is limited to formal written texts and does not consider how such integrations may evolve in spoken interactions, multilingual documents, or digital platforms. Future studies should investigate how these cohesive-narrative dynamics function across genres, technologies, and institutional settings. Future comparative research could examine how administrative cultures differ in their use of cohesion and narrative. Such analysis would clarify how language shapes authority and fosters institutional trust. To go further, future research could compare Vietnamese with unrelated languages such as Korean, German, or Arabic to see whether similar cohesion–narration patterns appear across different grammatical systems. Additionally, ethnographic methods including institutional interviews or protocol analysis could uncover the intentional strategies behind textual coordination. In digital contexts, studying administrative discourse on institutional websites or crisis announcements could show whether cohesion and narrative features change when messages are produced under time pressure or using multimedia tools. This would offer a broader, genre-sensitive model for how textual architecture evolves across modes, languages, and policy environments. V. CONCLUSION This study examined how cohesive devices and narrative elements interact to enhance the coherence, clarity, and communicative effectiveness of Vietnamese ATs. The study finds that cohesive tools like reference, conjunction, and ellipsis help maintain unity in administrative texts. At the same time, narrative elements such as processes, participants, and circumstances ensure logical flow and support institutional responsibility. The integration of these elements results in a robust textual architecture that conveys policy content clearly and reinforces the authority and procedural logic of administrative discourse. The results of this study are significant for both theory and practice. In terms of theoretical Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 224 perspective, these findings further bring the light to systemic functional linguistics by illustrating how textual coherence is achieved in an analytic language without morphological marking. Also, the results support integrated models of discourse that treat cohesion and narration as mutually reinforcing rather than isolated strategies. Practically, the study suggests that administrative writing training should incorporate structural accuracy and narrative awareness to ensure communicative impact and institutional transparency. Policy designers and public communicators can benefit from these insights to improve both reading competence and public trust. Besides the significant contributions mentioned above, the study still remains potential limitations because it mostly focused on written texts from formal institutional domains without accounting for variation in spoken or digital administrative communication. Moreover, the study was mainly conducted on texts that are inherent in requiring high caution in the process of writing. The application of CDs and NEs should be explored radically in multilingual and multimodal contexts in future research. In addition, more cross- linguistics studies should be across institutional cultures to help evaluate how CDs and NEs vary across administrative traditions. In conclusion, this study offers a new perspective on how linguistic structures support communication in administrative texts. It provides both a theoretical framework and practical guidance to enhance administrative writing. ETHICS STATEMENT I confirm that the current work does not involve human subjects, animal experiments, or any data collected from social media platforms. The research was conducted in accordance with the ethical publication standards of Jurnal Arbitrer. CREDIT AUTHOR STATEMENT Ly Ngoc Toan was solely responsible for the conceptualization, design, data collection, analysis, and writing of the manuscript. All aspects of the research were independently conducted by the author. ACKNOWLEDGEMENTS This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. DECLARATION OF COMPETING INTERESTS I declare that there are no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. REFERENCES Alyousef, H. S. (2016). A multimodal discourse analysis of the textual and logical relations in marketing texts written by international undergraduate students. Functional Linguistics, 3(1), 3. https://doi. org/10.1186/s40554-016-0025-1 Bauman, R., & Briggs, C. L. (1990). Poetics and performance as critical perspectives on language and social life. Annual Review of Anthropology, 19, 59–88. https://doi.org/10.1146/annurev. an.19.100190.000423 Boyd, R. L., Blackburn, K. G., & Pennebaker, J. W. (2020). The narrative arc: Revealing core narrative structures through text analysis. Science Advances, 6(32), eaba2196. https://doi.org/10.1126/sciadv. aba2196 Chen, X., & Meurers, D. (2016). Characterizing text difficulty with word frequencies. Proceedings of the 11th Workshop on Innovative Use of NLP for Building Educational Applications, 84–94. https://doi. org/10.18653/v1/W16-0509 De Silva, N. H., & Henderson, P. (2007). Narrative-based writing for coherent technical documents. Proceedings of the 25th Annual ACM International Conference on Design of Communication, 208– 215. https://doi.org/10.1145/1297144.1297190 De Sordi, J. O., Meireles, M., & De Oliveira, O. L. (2016). The Text Matrix as a tool to increase the cohesion of extensive texts. Journal of the Association for Information Science and Technology, 67(4), 900–914. https://doi.org/10.1002/asi.23434 Ehara, Y. (2019). Analyzing similarities between word frequency distributions of various corpora to find Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 225 effective features for complex word identification. 2019 International Conference on Data Mining Workshops (ICDMW), 1–4. https://doi.org/10.1109/ICDMW48858.2019.9024753 Espeland, W. (1993). Power, policy, and paperwork: The administrative representation of interests. Qualitative Sociology, 16(3), 297–317. https://doi.org/10.1007/BF00990103 Ewell, P. T. (1988). Implementing assessment: Some organizational issues. New Directions for Institutional Research, 1988(59), 15–28. https://doi.org/10.1002/ir.37019885904 Fobbe, E. (2021). Text-linguistic analysis in forensic authorship attribution. International Journal of Language & Law (JLL), Vol 9: Forensic Linguistics: New Procedures and Standards. https://doi. org/10.14762/JLL.2020.093 Haim, A. (2023). Binding language in administrative guidance documents. Proceedings of the Nineteenth International Conference on Artificial Intelligence and Law, 407–411. https://doi. org/10.1145/3594536.3595127 Halliday, M. A. K., & Hasan, R. (1976). Cohesion in English. London: Longman. Hermawan, A., Abd. Rohman, & Riky Welli Saputra. (2022). Analysis of school administration implementation. East Asian Journal of Multidisciplinary Research, 1(9), 1833–1844. https://doi. org/10.55927/eajmr.v1i9.1489 Herzog, R. J., & Claunch, R. G. (1997). Stories citizens tell and how administrators use types of knowledge. Public Administration Review, 57(5), 374. https://doi.org/10.2307/3109983 Hjern, B., & Porter, D. O. (1981). Implementation structures: A new unit of administrative analysis. Organization Studies, 2(3), 211–227. https://doi.org/10.1177/017084068100200301 Jahja, N. (2021). Ellipsis in English and Albanian. Prizren Social Science Journal, 5(1), 118–124. https:// doi.org/10.32936/pssj.v5i1.222 Jopek-Bosiacka, A. (2023). The legislative text as a legal story: A storytelling approach to contemporary legislative drafting. Pázmány Law Review, 10(1), 57–69. https://doi.org/10.55019/plr.2023.1.57-69 Kadhim, H. S. (2024). An in-depth analysis of cohesion utilization in the english language: Implications for effective communication and textual coherence. European Journal of Contemporary Education and E-Learning, 2(5), 77–92. https://doi.org/10.59324/ejceel.2024.2(5).05 Kreines, M. G., & Kreines, E. M. (2020). Matrix models of texts: Models of texts and content similarity of text documents. Mathematical Models and Computer Simulations, 12(5), 696–705. https://doi. org/10.1134/S2070048220050105 Kumar, A. (2022). The quality of references reveal the merit of the research. Journal of Indian Society of Periodontology, 26(6), 519–520. https://doi.org/10.4103/jisp.jisp_410_22 Law, L., & Matthiessen, C. M. I. M. (2023). Revisiting Halliday’s (1990) ‘New ways of meaning: The challenge to applied linguistics’: What has changed and what still needs to be done? Linguistics and the Human Sciences, 15(3), 337–368. https://doi.org/10.1558/lhs.23599 Linde, C. (1999). The Transformation of Narrative Syntax into Institutional Memory. Narrative Inquiry, 9(1), 139–174. https://doi.org/10.1075/ni.9.1.08lin Maheshwari, S. R. (2003). Politics of implementation: The achilles’ heel of public administration. Indian Journal of Public Administration, 49(4), 679–692. https://doi.org/10.1177/0019556120030401 Martin, J. R., Quiroz, B., & Wang, P. (2023). Systemic functional grammar: A text-based description of English, Spanish and Chinese (1st ed.). Cambridge University Press. https://doi. org/10.1017/9781009284950 Megerdoomian, K., Marsh, A. B., Scott, E. O., Branting, K., Modly, N., Wariyar, S. B., Horowitz, C. E., & Petersen, S. J. (2019). Automated narrative extraction from administrative records. Matthew, O. A. et al. (2024). Discourse markers, coherence and its impact on cognitive interpretation of texts. International Journal of Literature, Language and Linguistics, 7(2), 28–37. https://doi. org/10.52589/IJLLL-PUGJFQWZ Ly Ngoc Toan / Jurnal Arbitrer - Vol. 12 No. 2 (2025) 226 Rahman, R. F., Firdaus, D., & Ardiansyah, E. A. (2023). Cohesive devices in Donald Trump’s first speech as the american president. Call, 5(2). https://doi.org/10.15575/call.v5i2.21719 Taufan, G. T. (2021). CDsin English contents written by blog writers and English Department students. Ahmad Dahlan Journal of English Studies, 8(1 (2021)), 1. https://doi.org/10.26555/adjes.v8i1.17446 Sharif, H. N., & Awl, O. M. (2022). CDsof ATs in Kurdish Language. Journal of Kurdistani for Strategic Studies, 2. https://doi.org/10.54809/jkss.vi2.72Jabur Jinam, E. (2023). The Role of Cohesion in Text Organization. Full-Text Book Of Ijher Congress7, 47–64. https://doi.org/10.47832/IjherCongress7-4 Zinner, S. E. (2011). The stories public administrators tell: Finding meaning in the ordinary work life. Public Integrity, 13(4), 385–396. https://doi.org/10.2753/PIN1099-9922130406