Compassion, Doubt, and Eternal Damnation: An Analysis of Sincere Belief and Doubt Under Religious Diversity - 31

An Analysis of Sincere Belief and Doubt Under Religious Diversity Thesis. A deity worthy of the predicates “loving” and “just” would not eternally punish agents who proportion belief to evidence amid religious diversity; doctrines that condemn sincere, rational doubters are epistemically and ethically incoherent. Building on a comparative case (“Mariam”) in which Christianity and Islam appear approximately equally supported, this paper unifies three literatures: divine hiddenness (Schellenberg, 1993), proportionality in punishment (Cutter, 2025; Kvanvig, 1993; Talbott, 2017; Walen, 2014), and the social–cognitive formation of theistic belief (Barrett, 2004; Norenzayan, 2013). We formalize the 45–45–10 credence split with a transparent Bayesian skeleton and show that, absent non-circular discriminatory evidence, condemnation conditional on selecting the “wrong” hypothesis violates proportionality and due epistemic process. Recent demographic work on religious switching and the “cultural lottery” (Pew Research Center, 2025a, 2025b) further undermines the claim that a loving God has provided sufficiently unambiguous guidance across cultures. The result is a cumulative case that eternal-damnation doctrines targeting rational uncertainty are inconsistent with a Mariam, born to a Christian mother and a Muslim father, undertakes a careful comparative inquiry. After extended study, she assigns roughly 45% to Christianity ( /u1D43B /u1D436 ), 45% to Islam ( /u1D43B /u1D43C ), and 10% to “neither is true” ( /u1D43B ¬ ). Her credences track available evidence as she sees it; they are defeasible and open to revision. Is it coherent to claim that a compassionate God would consign her to eternal misery for failing to select the “correct” Let a partition of hypotheses be { /u1D43B /u1D436 , /u1D43B /u1D43C , /u1D43B ¬ } with priors /u1D70B /u1D436 = 0.45, /u1D70B /u1D43C = 0.45, and /u1D70B ¬ = 0.10. Let /u1D438 summarize Mariam’s total comparative evidence (texts, experiences, testimony, background data). The odds update is: Pr(/u1D43B /u1D43C | /u1D438 ) = Pr(/u1D438 | /u1D43B /u1D436 ) Pr(/u1D438 | /u1D43B /u1D43C ) · /u1D70B /u1D436 Annotation: Posterior odds equal.