서강대학교

목차

Acknowledgements i
Table of Contents iv
List of Tables ix
List of Figures x
Abstract xi

Chapter 1. Introduction 1
1.1 Theoretical orientation of AI-assisted revision 3
1.2 Methodological orientation and research context 6
1.3 The present study 10
1.4 Definition of terms 12
1.5 Significance of the study 16
Chapter 2. Review of the literature 19
2.1 Writing, revision, and mediation in L2 contexts 22
2.2 Cognitive models of revision and diagnostic decision-making 26
2.2.1 The cognitive process model of writing and revision 27
2.2.2 Diagnostic strategies in revision: Expert and novice writers 28
2.2.3 Cognitive models as foundations for AI-mediated revision analysis 31
2.3 L2 composing as constrained cognition. 33
2.3.1 L2 writing as constrained cognitive activity 34
2.3.2 Revision as hypothesis testing in interlanguage development 38
2.3.3 Learning to write and writing to learn 41
2.3.4 Revision, proceduralization, and attentional control 43
2.3.5 Implications for analyzing AI-mediated revision 46
2.4 Textual quality as evidence of development 47
2.4.1 From product accuracy to multidimensional quality 48
2.4.2 Linguistic dimensions of textual quality 50
2.4.3 Discourse-level quality: Coherence, cohesion, and rhetorical control 53
2.4.4 Errors, restructuring, and developmental progress 55
2.4.5 Limitations of automated quality measures 57
2.4.6 Textual quality as a revision-sensitive construct 58
2.5 Mediated feedback and the role of technology 60
2.5.1 Feedback as mediation in L2 writing 61
2.5.2 Teacher and peer feedback: Insights and limitations 64
2.5.3 Automated writing evaluation (AWE): From correction to mediation 65
2.5.4 Cognitive constraints and feedback overload 67
2.5.5 Large language models as dialogic feedback tools 68
2.5.6 Feedback literacy, learner agency, and analytic implications 70
2.6 An integrated framework for analyzing AI-mediated revision 73
2.6.1 Conceptual foundations of the framework 74
2.6.2 Overview of the three analytic dimensions 77
2.6.3 Revision type: What changes occur 78
2.6.4 Revision effectiveness: What revisions accomplish 79
2.6.5 Learner engagment: How and why writers revise 81
2.6.6 Integrating the three dimensions 83
2.7 Research gaps and research questions 85
Chapter 3. Methodology 90
3.1 Research context and participants 91
3.2 Research instruments and instructional tools 95
3.3 Research procedures 97
3.3.1 Course structure and writing tasks 97
3.3.2 AI-assisted feedback and revision process 100
3.3.3 Data collection 104
3.4 Analytic framework 105
3.5 Reliability and validation of coding and scoring 110
Chapter 4. Results and Discussion 113
4.1 Descriptive overview of revision patterns 114
4.1.1 Distribution of revision operations 115
4.1.2 Meaning- versus surface-level orientation 118
4.1.3 Quality dimensions targeted 121
4.1.4 Interpretive summary and implications for textual analysis 124
4.2 Key findings from textual analysis 127
4.2.1 Proficiency shapes revision orientation 130
4.2.1.1 Additive vs. integrative trajectories 133
4.2.1.2 Surface vs. meaning level balance 142
4.2.1.3 From local to global cohesion 150
4.2.2 Revision effectiveness and developmental imbalances 159
4.2.2.1 Expansion gains, limited structural integration 160
4.2.2.2 Additive growth, local cohesion limitations 163
4.2.2.3 Rhetorical density and integrative overload 166
4.3 Integrating reflection and textual evidence: Patterns of engagement 171
4.3.1 Low-proficiency writers: Mechanical uptake with growing confidence 173
4.3.2 Mid-proficiency writers: Selective adaptation and strategic uptake 176
4.3.3 High-proficiency writers: Critical ownership and dialogic reflection 179
4.4 Discussion 185
4.4.1 Cognitive and rhetorical mechanisms of revision 186
4.4.2 Quality gains and the limits of complexity 189
4.4.3 Engagement, mediation, and emerging autonomy 193
4.4.4 Extending revision theory to AI-mediated contexts 197
4.5 Summary of chapter 202
Chapter 5. Conclusion 205
5.1 Summary of the findings. 205
5.2 Pedagogical implications 208
5.2.1 AI as mediational scaffold 208
5.2.2 Differentiated integration by proficiency 212
5.2.3 Reflection and prompt literacy 216
5.2.4 Integrating AI within the writing curriculum 219
5.3 Theoretical contributions 222
5.4 Limitations and directions for future research 225
5.5 Concluding remarks 229

References 232

Appendices 247
A. Student drafts: First and revised versions 247
B. AI Feedback Prompts 275
C. Scoring Rubric 278
D. Student Consent Form 282