A Question-Structure Analysis of the Anattā-Saṃsāra Tension in the Nikāyas

This study systematically classifies the Buddha's linguistic strategies in Pāli Nikāya texts where anattā (non-self) and saṃsāra (rebirth) intersect, and presents the resulting patterns. While prior scholarship has sought direct answers to the question 'How is rebirth possible if there is no self?', this study asks how the Buddha himself handled that very question. The central inquiry is not 'What answer did he give?' but 'What question-frame did he demand?' A research corpus of 86 suttas (255 dialogue turns) was compiled from the four Nikāyas (DN, MN, SN, AN) and the early stratum of the Khuddaka, following explicit inclusion and exclusion rules. A codebook comprising seven question-type codes (Q-AGT, Q-IDT, Q-EXS, Q-PST, Q-MEC, Q-OTH, Q-SUI) and seven answer-type codes (A-REJ, A-SIL, A-DEP, A-MID, A-MET, A-COR, A-EXP) was designed, and exhaustive coding was carried out according to a seven-step decision tree. All coding rules were frozen (Freeze-1 and Freeze-2) prior to drawing conclusions, structurally preventing post-hoc rule adjust-ment. The analysis yields three quantitative maps. Map 1 (Q-code × A-code cross-tabulation) re-veals that the Buddha predominantly employed question-rejection and dependent-origination reframing in response to agent-questions and identity-questions, while shifting to process-explanation for mechanism-questions. Map 2 (co-occurrence of marker phrases) confirms that 'no kallo pañho' (the question is not fitting), 'na ca so na ca añño' (neither the same nor another), and the dependent-origination formula co-occur at high frequency. Map 3 (viññāṇa usage-cluster verb-combination patterns) partitions the uses of viññāṇa (con-sciousness) by three criteria — grammatical collocates, position within the dependent-origination chain, and result-predicate linkage — showing that rebirth-related (V-REBIRTH), persistence-related (V-PERSIST), and cognitive (V-COGN) functions are al-ready distinguished within the Nikāyas themselves. For counterexample adjudication, five types of tension-texts (T1: consciousness as rebirth-subject; T2: kamma inheritance; T3: citta/mano/viññāṇa interchange; T4: pabhassara-citta; T5: past-life recollection) were judged as genuine tension, surface tension, or spurious ten-sion, with each verdict traceable to specific rule-item numbers. This study does not propose a single resolution to the anattā-saṃsāra problem. Instead, it provides five deliverables — corpus membership rules, codebook and decision tree, viññāṇa functional map, counterexample adjudication table, and robustness check table — which serve as reusable research infrastructure for subsequent scholars to verify, replicate, or extend.