1. Introduction

This chapter discussed the centrality of argumentation as a marker of academic literacy and the specific difficulty L2 writers face in balancing their own position with the acknowledgment of alternative stances. It defined voice as the writer’s capacity to project personal views while simultaneously representing and evaluating other voices, and it noted that L2 learners frequently struggle to integrate these demands. Early studies on voice in L2 argumentative texts have relied on Hyland’s interactional metadiscourse model or Zhao’s three-dimensional rubric, yet both were criticized for underplaying the role of context and textual co-construction. To redress these gaps, the chapter introduced a three-stratal analytical framework grounded in Systemic Functional Linguistics, Sydney School genre theory, and Appraisal’s Engagement system. This framework targets global generic stages and local lexicogrammatical choices to trace how authorial voices emerge through schematic structure and how voice negotiations unfold through Engagement resources at paragraph and clause levels. The investigation focused on high-quality argumentative essays produced by Chinese undergraduates in the Uchallenge national contest, observing hybrid schematic stages that resemble “mixed texts.” By analyzing these essays, the study aimed to uncover systematic patterns of voice negotiation and to test whether a robust authorial voice correlates with superior argumentative quality, thereby offering empirical insights for L2 writing pedagogy.

2. Literature Review

This chapter described the Sydney School genre theory’s view of argument genres as staged, goal-oriented social processes and applied Coffin’s (2006) classification into exposition, discussion, and challenge genres, each with distinct schematic stages that our evaluation framework adopts. It noted that Martin and Rose (2008) offer a parallel but less elaborated typology, while Pessoa et al. (2017) extend exposition analysis using SFL resources. The chapter also acknowledged genre hybridity or “mixed texts,” where one genre serves as a stage within another to fulfill overarching communicative aims.

Turning to interpersonal meaning, the chapter introduced Martin and White’s (2005) Engagement system within Appraisal Theory, differentiating monogloss—bare, categorical assertions—from heterogloss, which encompasses dialogic contraction (Disclaim, Proclaim) and expansion (Entertain, Attribute) to position the writer vis-à-vis alternative voices. Empirical studies show that high-graded argumentative essays in both geography and history deploy strategic Engagement moves—Counter, Attribute, Endorse—to negotiate evidence contradictions and project an authoritative yet dialogic stance. The chapter concluded by proposing a three-level (text-paragraph-clause) authorial voice framework that links schematic structures with Engagement realizations to examine how L2 learners negotiate voice across argumentative stages.

3. Research Framework of Authorial Voice Representation in Argumentation

This chapter discussed the construction of a three-tier analytical framework for tracing authorial voice in L2 argumentative texts. It adopted Martin’s stratified SFL model, situating voice negotiation at the levels of full text, paragraph, and clause. Genre is realized through register variables (field, tenor, mode) mapped onto the ideational, interpersonal, and textual metafunctions. At the whole-text level, the obligatory stages of Coffin’s exposition genre—(Background) ^ Thesis ^ Arguments ^ (Counter) ^ (Concession) ^ Reinforcement—form the global scaffolding of authorial stance. Within paragraphs and clauses, Engagement resources (Martin & White, 2005) lexicalize how writers contract or expand dialogic space, thereby negotiating voice moment-by-moment. Empirical samples showed learners consistently opening with Thesis and closing with Reinforcement, aligning with the exposition schema, so this pattern serves as the baseline for detecting stage-specific voice features.

4. The Present Study

This chapter describes the data source and analytical procedures employed to investigate how successful L2 writers negotiate authorial voice in high-quality argumentative texts. Data consist of all 14 national-final argumentative essays publicly released from the Uchallenge contest between 2015 and 2019, totaling 10,717 words. These texts were produced by the top one hundred or so Chinese undergraduates during the three-hour National Final in Beijing and were judged by genre specialists as exemplary models. Each essay is accompanied by a judge’s commentary and the contestant’s own reflection, making the corpus ideal for examining recommended practice. The contest prompt explicitly requires writers to articulate a personal stance toward a provided source passage and to substantiate it with reasons and evidence, thereby foregrounding the interpersonal dimension of argumentation.

Because the corpus is small but rich, it supports intensive qualitative analysis rather than large-scale quantification. No additional software was used; instead, the texts were pasted into a single Word document and labeled S1-Y15 through S14-Y19 for easy reference and manual annotation. Two researchers conducted two complete rounds of coding, resolving disagreements through discussion until 100 % consensus was achieved; inter-coder divergence never exceeded 10 %.

Analytical work proceeded in three integrated steps. First, each essay was segmented into functional stages following Coffin’s (2006) genre-based descriptions: Context, Thesis, Argument for, Counter Argument, Concession, Rebuttal, Reinforcement, and Conclusion. These stages were mapped to show how information is organized to achieve the social purpose of persuasion, yielding schematic structures that specify both sequential progression and simultaneous configurations within paragraphs. Second, the texts were coded for heteroglossic Engagement resources drawn from Appraisal theory, focusing on how writers expand or contract dialogic space by entertaining, acknowledging, countering, or proclaiming other voices. Annotation attended to lexicogrammatical realizations such as modality, attribution, concession, and concession-plus-rebuttal clusters, always interpreted in context. Third, the interplay between stage structure and Engagement choices was examined to reveal how voice is constructed across the unfolding argument. Particular attention was paid to how different stages deploy Engagement resources to negotiate solidarity, anticipate opposition, and maintain authorial credibility. Tables summarizing stage frequencies and schematic structures in openings and body sections are provided to visualize these patterns.

5. Research Findings

This chapter described how the 14 award-winning L2 argumentative essays construct authorial voice through distinctive schematic structures. Hybrid generic stages are fused with the expected argument stages, yet only in the opening and body sections. In the opening, every essay situates its Thesis within a multi-voiced context created by optional Issue, Position Challenged and Perspective stages. Issue (present in 11 essays) foregrounds the debatable topic and alternative readings instead of neutral background, signalling an awareness of competing positions. Position Challenged explicitly introduces a stance to be refuted, sharpening the authorial counter-voice, while Perspective presents a partial alternative that is nuanced rather than rejected outright, allowing negotiation without direct negation.

In the body, Position Challenged reappears at paragraph level to voice potential objections; it prefaces counter-evidence that ultimately functions as Argument For Thesis, turning opposition into support. Rebuttal Argument immediately follows, exposing flaws in the counter-position and tightening alignment with the writer’s stance. Proposal stages, borrowed from hortatory exposition, surface twice to recommend concrete action after Thesis elaboration, satisfying anticipated reader questions and enhancing persuasion. Embedded genres (story recount, observation, explanation with exemplification and quotation) are woven into Issue and Proposal stages: narrative observations make the issue personally resonant, and authoritative voices and vivid examples animate recommendations, increasing reader engagement.

Three recurrent patterns organise these stages strategically. The opening section of 11 essays employs “Issue + Thesis” or “Issue + Position Challenged/Perspective + Thesis,” embedding the authorial claim within a contested dialogue from the outset. A second pattern, “Concession + Counter Argument,” follows Argument For in seven essays: Concession anticipates and acknowledges limitations or alternative views, while Counter Argument then diminishes their force, refining the authorial position without denying other voices. These configurations jointly demonstrate how high-quality L2 argumentation orchestrates hybrid stages and their sequencing to project a dialogic yet authoritative voice across the entire text.

6. Discussion

This chapter discussed how high-proficiency L2 undergraduates orchestrated schematic structure, hybrid stages, and engagement resources to negotiate an authoritative yet dialogic voice in award-winning argumentative essays. Text-level analysis revealed a composite macro-genre assembled from Issue, Position Challenged, Perspective, Rebuttal Argument, and Proposal stages, supplemented by embedded story genres within Issue/Background to create reader resonance and explanatory moves within Proposal to operationalize recommendations. Such patterned interweaving of genres echoes the rhetorical framing observed in research articles and foregrounds the pedagogical value of explicit instruction in multi-generic composition.

Paragraph-level configuration displayed deliberate hybrid staging that foregrounded the writer’s stance while acknowledging counter-positions. Opening sections frequently fused Issue with Thesis or combined Issue, Position Challenged/Perspective, and Thesis, whereas body paragraphs coupled Concession with Counter-Argument. Even essays that began with Background+Thesis or Thesis alone embedded alternative voices within Background or later as Concession/Position Challenged, ensuring that authorial voice emerged through negotiated tension rather than monologic assertion.

Clause-level engagement resources manifested a consistent “Expand+Contract” rhythm. In introductions, students alternated Attribute (Distance/Acknowledge) with Counter and Pronounce, first enlarging the dialogic space by invoking external voices from the prompt passage and then contracting it to assert their own claim, thereby positioning the thesis as the governing voice for the entire text. This mirrors the rhetorical pattern documented in MA literature reviews, demonstrating cross-generic socialization into academic argumentation. In body sections, the “Concede+Counter” pairing qualified claims against anticipated objections, while an additional “Entertain” move invited readers holding similar positions into the argument, balancing critique with solidarity. These finely tuned engagement choices sustained coherence across clauses and fostered persuasive alignment.

Nevertheless, two shortcomings in source integration were identified. First, students often left external voices implicit, replacing specific citations such as “Sir David Attenborough” or “Charles Darwin” with vague references like “some people,” thereby forfeiting the rhetorical capital of high-status sources. Thirty-four out of forty-six unattributed instances involved this evasive strategy, suggesting gaps in topical knowledge or citation skill under contest conditions. Second, endorsement—the resource for overtly validating a source—was markedly underused; only twenty-four instances appeared across the corpus. When quoting figures such as Don Murray, writers typically resorted to attribution followed by “entertain,” offering no interpretive framing of the quotation’s significance. Effective endorsement would require contextual tags that specify the speaker’s identity, the original setting of the utterance, and its relevance, thus strengthening intertextual alignment and reinforcing the overall argumentative force.

7. Conclusion

This chapter discussed how authorial voice is negotiated in prize-winning argumentative essays produced by fourteen Chinese undergraduates, applying a three-level stratified SFL framework that tracks meaning from full text through paragraph to clause. It revealed that the writers effectively deploy hybrid generic stages and Engagement resources to stage their arguments, yet often mask the origin of external voices through implicit attribution and endorsement. Rather than claiming broad generalizability, the study emphasized its methodological contribution: the “full text-paragraph-clause” lens offers a holistic picture of how voice is constructed and negotiated, guiding teachers to make generic and interpersonal expectations explicit. Pedagogically, instructors should present both the fixed conventions and the permissible variation within argumentative genres, exemplified by the productive “Issue + Position Challenged/Perspective + Thesis” opening pattern, and train students to use Engagement resources such as “Concede + Counter” to manage dialogic space. To counter limited topical and citation knowledge, the chapter recommended integrating “arguing to learn” with “learning to argue,” urging students to interrogate source texts for structure and voice use, thereby deepening contextual understanding and critical thinking. Finally, it suggested future research refine the analytical model by incorporating gradability within Engagement values to explore how intensity of stance, realized lexicogrammatically, shapes authorial voice across cultural contexts.