Question and assessment context
Is the evolutionary argument against naturalism successful at showing that naturalism plus evolution gives its believers a defeater for trusting their cognitive faculties?
Based on the complete debate transcript. Summaries are condensed for clarity, and quotations are kept brief and faithful to the source.
Scores are AI-generated estimates of argumentative performance and are intended as transparent, revisable judgments rather than final verdicts.
Slugfester’s assessment reviewers and adjudicators are independent AI agents called by the administrative AI.
Overall assessment
Alvin Plantinga — EAAN defense: 81/100
Published strengths
- Precisely framed the EAAN as a challenge to rational acceptance rather than a direct proof that naturalism and evolution are false.
- Clearly connected recognized faculty unreliability to a self-referential defeater encompassing belief in naturalism and evolution.
- Distinguished adaptive neurological output from semantic truth and later isolated the gap between supervenient determination and correspondence.
- Answered skeptical parity charitably by conceding suffering's evidential relevance while insisting that its force be assessed within total evidence.
Published weaknesses
- The low-reliability route repeatedly moved from direct selection for adaptive behavior to chance-level truth without adequately defending equal odds, independence, or a reliability threshold.
Stephen Law — Reliability challenge: 81/100
Published strengths
- Supplied a concrete resource-seeking mechanism through which accurate environmental beliefs can enhance survival and become indirectly selected.
- Challenged arbitrary content reassignment by invoking conceptual, causal, and behavioral constraints across both identity and nonidentity accounts.
- Correctly separated atheism from naturalism, clarifying the EAAN's conceptual scope without conflating distinct worldviews.
- Tested the claimed theistic advantage through a structurally relevant skeptical-parity argument and kept the final evolutionary claim appropriately fallibilist.
Published weaknesses
- The parity argument did not establish that suffering provides net defeating evidence against the reliability-grounding God, and its final restatement left Plantinga's total-evidence reply unanswered.
Section-by-section summary
These are excerpts from the published assessment, not a new evaluation. Each section shows its score and one representative assessed move from each side.
Adaptive Selection and Truth-Conducive Cognition
20:02–34:58
Alvin Plantinga — section score 77/100
Under materialist naturalism, evolution selects adaptive neurology and bodily behavior rather than semantic truth, leaving associated belief content without direct selection for accuracy.
Strongest feature: The move provides the central explanatory mechanism behind the low-reliability premise by sharply separating evolutionary selection for adaptive bodily performance from direct selection for semantically true belief content. Principal limitation: From the defensible observation that truth is not selection's immediate object, it moves too quickly toward chance-level accuracy without excluding indirect truth tracking through flexible, environmentally responsive representational mechanisms. Live burden: Plantinga must show that adaptive success remains substantially indifferent to content accuracy across relevant environments, despite the possibility that generally accurate representation systematically improves resource acquisition, prediction, and behavioral flexibility. Locked score: The strong/competent band reflects direct engagement with a central bridge, a recoverable distinction, and substantial argumentative relevance, tempered by materially overstated probabilistic force and incomplete treatment of Law's strongest indirect-selection alternative.
Stephen Law — section score 87/100
Selection can favor truth-producing mechanisms because organisms seeking water, food, or mates survive when their beliefs about those necessities' locations are accurate.
Strongest feature: Law directly answers the proposed separation between truth and fitness with a concrete mechanism in which needs, desires, location beliefs, and successful resource-seeking behavior make representational accuracy systematically advantageous. Principal limitation: The example establishes a plausible route for ecologically salient beliefs but leaves implicit how that localized advantage generalizes to overall cognitive-faculty reliability, abstract reasoning, or domains less tightly coupled to survival. Live burden: Law must extend the mechanism beyond illustrative water-seeking while preserving his modest claim that evolution can favor generally accurate cognition without requiring every adaptive action or evolved belief to be true. Locked score: The very strong band reflects outstanding responsiveness, clear causal intelligibility, proportionate plausibility, and charitable contact with Plantinga's alternative, while reserving uncertainty about domain generalization and the threshold for comprehensive reliability.
Mental Content, Neurology, and Behavioral Constraints
35:14–66:07
Alvin Plantinga — section score 88/100
Even if neurological properties determine a particular belief content by supervenience, that determination alone does not establish correspondence between content and adaptive action.
Strongest feature: Plantinga identifies with unusual precision the gap between neurological determination of a particular content and correspondence between that content, the organism's action, and the external circumstances making the content true. Principal limitation: The distinction defeats supervenience considered alone but does not exclude richer externalist accounts in which causal history, environmental relations, and behavioral success jointly constrain content and support accuracy. Live burden: Plantinga must demonstrate that Law's proposal relies only on intraneural supervenience, or else address the stronger possibility that environmental and causal constraints bridge determination to truth without making semantic truth directly selectable. Locked score: The very strong band reflects exemplary diagnostic responsiveness, stable terminology, calibrated concession, and exact contact with both conceptual and causal replies, while retaining a limited reservation concerning the unexcluded externalist alternative.
Stephen Law — section score 84/100
Belief content cannot be inserted arbitrarily into neural structures, because conceptual constraints connect attributable content with the associated state's behavioral role.
Strongest feature: Law identifies weakness in unrestricted content reassignment by arguing that a neural state’s behavioral role conceptually constrains which belief can intelligibly be attributed to the organism. Principal limitation: The response leaves its theory of content and the reach of constraints implicit, while excluding arbitrary attribution does not establish that admissible content accurately represents the external world. Live burden: Law must connect behavioral-role constraints not merely to action but to environmental truth conditions, while accommodating Plantinga’s concession that determinate content may supervene without corresponding to adaptive behavior. Locked score: The very strong band reflects diagnostic precision, exceptional responsiveness, stable presentation, and tentative force, moderated by the missing external-world accuracy bridge and underdeveloped account of semantic determination.
Probability of Reliable Cognitive Faculties
17:03–63:05
Alvin Plantinga — section score 72/100
If naturalism and evolution are true, the probability that human cognitive faculties are reliable is low, although Plantinga presents this formulation as approximate.
Strongest feature: The conditional premise identifies the argument's indispensable probabilistic hinge with commendable directness, explicitly restricting its application to the conjunction of naturalism and evolution while acknowledging that the formulation is approximate. Principal limitation: Neither cognitive reliability nor the threshold represented by “low” receives sufficient specification, and the excerpt announces the probability judgment without yet furnishing the evolutionary mechanism or probabilistic model needed to substantiate it. Live burden: Plantinga must establish that selection's orientation toward adaptive behavior makes truth-conducive cognition genuinely improbable or inscrutable, rather than merely showing that semantic truth is not selection's immediate target. Locked score: The strong/competent band reflects excellent relevance and coherent conditional architecture, moderated substantially by underdeveloped warrant, terminological instability, and an asserted probability stronger than this isolated presentation establishes.
Stephen Law — section score 81/100
A naturalist can plausibly explain why unguided evolution tends to favor true beliefs, making the EAAN's reliability challenge substantial but surmountable.
Strongest feature: Law's conclusion is carefully calibrated as a surmountability claim, recalling his earlier truth-fitness mechanism without pretending to prove universal accuracy or deny that Plantinga has presented a substantive evolutionary challenge. Principal limitation: The summary does not separately answer Plantinga's compound even-odds model and compresses the contested relationship among adaptive neurology, conceptual constraints, causal association, and external-world truth into a brief assertion of tendency. Live burden: Law must integrate those mechanisms into a general reliability account and explain why correlated, domain-sensitive selection defeats the assumptions behind Plantinga's aggregation from individual uncertainty. Locked score: The strong/competent band reflects coherent framing, central relevance, charitable qualification, and meaningful contact with the low-reliability premise, while recognizing partial responsiveness and the absence of a distinct answer to the aggregate-probability component.
Defeater Generation and Epistemic Self-Reference
17:31–63:55
Alvin Plantinga — section score 88/100
Once naturalism and evolution undermine trust in cognitive reliability, the resulting defeater spreads to every faculty-produced belief, including naturalism and evolution themselves.
Strongest feature: The move supplies the motion-level bridge with exceptional explicitness, connecting recognized doubts about faculty reliability to epistemic defeat and then applying that consequence self-referentially to belief in naturalism and evolution. Principal limitation: Its universal propagation from a defeater for general reliability to defeat for every particular belief remains insufficiently qualified, because compartmentalized trust, independent support, and possible defeat-defeaters are neither distinguished nor excluded. Live burden: Plantinga must explain the epistemic conditions under which global reliability doubt transmits to each faculty-produced belief, especially when some beliefs might possess external corroboration or comparatively resilient warrant. Locked score: The strong/competent band records commanding burden relevance and a lucid conditional structure, while recognizing that the necessity of faculty-wide propagation is slightly overstated and its governing epistemic assumptions remain implicit.
Stephen Law — section score 72/100
Even if naturalism yields global skepticism, theism gains no advantage if supposedly God-given faculties produce strong grounds against God and thereby undercut themselves.
Strongest feature: Law preserves the comparative focus by arguing that conceding naturalistic skepticism would not automatically establish a theistic advantage, and he clearly restates the proposed faculty-undermining loop on the theistic side. Principal limitation: The move does not answer Plantinga's immediately preceding total-evidence distinction, instead presuming that faculty-delivered adverse considerations amount to sufficiently strong grounds against the God invoked to support reliability. Live burden: Law must demonstrate that the relevant evidence is net defeating, not merely significant, and explain why it undercuts the specific theistic reliability basis after all available evidence is considered. Locked score: The mixed band reflects real relevance, cautious possibility framing, and partial contact with the comparative and self-defeat components, moderated by unstable evidential scope, reduced charity, and omission of the decisive latest reply.
Worldview Scope and Theistic Symmetry
23:40–60:34
Alvin Plantinga — section score 84/100
Theism supports an expectation of generally reliable faculties because image-bearing humans possess capacities for knowledge, whereas naturalism provides no comparable purposive expectation.
Strongest feature: The comparative argument is clearly delimited and appropriately qualified, offering creation in the divine image and a purposive capacity for knowledge as an intelligible theistic basis for expecting faculties reliable for the most part. Principal limitation: It assumes without substantial defense that divine purposes and image-bearing favor broad cognitive reliability, while the absence of an analogous purposive guarantee under naturalism does not itself establish a negative reliability expectation. Live burden: Plantinga must justify the transition from theistic anthropology to dependable cognition and explain why naturalistic causal mechanisms cannot generate a comparable, though nonpurposive, expectation of truth-conducive performance. Locked score: The strong/competent band recognizes a precise, relevant worldview contrast and proportionate qualification, while discounting the underargued assumptions connecting divine creation to reliability and lacking purpose to epistemic disadvantage.
Stephen Law — section score 80/100
Atheism does not entail naturalism, because denying gods remains compatible with abstract objects, nonnatural properties, consciousness, and other nonphysical realities.
Strongest feature: Law accurately distinguishes atheism from naturalism through concrete examples of non-theistic commitments to abstract objects, nonnatural properties, or irreducible consciousness, preventing an unwarranted conceptual identification of the positions. Principal limitation: Although important for audience scope, the distinction does not answer Plantinga's actual comparison between naturalism's resources and theism's purposive basis for expecting generally reliable cognitive faculties. Live burden: Law must connect the broadened worldview taxonomy to the reliability dispute by showing either how non-theistic nonnaturalism escapes the EAAN or why naturalism itself retains a positive reliability expectation. Locked score: The strong/competent band reflects precise conceptual analysis, sound calibration, and substantive subsidiary relevance, while registering that the move is a relevant nonanswer to the decisive theistic-expectation claim and therefore contributes less to the central comparative burden.