Jiayu Chen

As synthetic visuals produced by Generative Artificial Intelligence (GenAI) proliferate across media and research contexts, they raise urgent epistemological, ethical, and methodological questions for ethnography. This commentary extends debates on GenAI's role in communication and knowledge production by focusing on ethnographic concerns of representation, reflexivity, and visual ethics. GenAI imagery, often photorealistic but nonindexical, unsettles long-standing conventions of co-presence, negotiated authorship, and situated seeing. This commentary proposes the 3C Framework of Contextuality, Consent, and Criticality, intended to guide future engagements with GenAI synthetic visual media in ethnographic research. Grounded in foundational ethnographic values, the framework is heuristic rather than prescriptive, offering a starting point for ethical experimentation and positional reflection.

Epistemicism, as proposed by Timothy Williamson (1994), claims that vague predicates possess precise but epistemically inaccessible boundaries. This paper critically evaluates Epistemicism using the original conceptual framework of the Killing Case Paradox (KCP). Through a detailed metaphysical, semantic test, KCP reveals fundamental contradictions within Epistemicism. The paradox demonstrates that the supposed precise boundaries cannot be coherently integrated into causal structures and linguistic practice. Additionally, the paper introduces the Causation–Detectability Principle (CDP), asserting that real-world causal boundaries must, in principle, be detectable or inferable. Ultimately, KCP exposes Epistemicism's reliance on undetectable, metaphysically speculative boundaries, leading to an infinite regress and conceptual collapse. The paper concludes that Epistemicism fails as a tenable philosophical theory of vagueness. Keywords: Vagueness, Epistemicism, Sorites paradox.

This paper defends an evolutionary gradualist account of moral agency against Christine M. Korsgaard’s saltatory moral evolution hypothesis. Korsgaard, drawing on Kantian notions of normative self-government, posits that moral agency emerged fully formed at a singular point, a unique capacity that decisively separated humans from our non-moral animal ancestors. I challenge this claim by invoking the Sorites paradox and a novel “Killing Case” paradox (KCP), highlighting the problem of vagueness inherent in drawing any precise boundary within a continuous evolutionary process. Specifically, KCP shows that any attempt to identify a precise boundary within a continuous evolutionary process inevitably leads to logical contradictions or self-defeating conclusions, thus providing a novel logical tool beyond the traditional Sorites paradox. Korsgaard’s model effectively demands identification of an exact threshold for the onset of moral agency, which leads to logical contradictions and epistemic indeterminacy. Furthermore, I marshal empirical evidence from evolutionary biology, primatology, and moral psychology indicates that complex social behaviours such as empathy, fairness, and norm-sensitivity appear in nonhuman animals to varying degrees. These behaviours serve as transitional forms and provide no evidence for any sudden or singular biologically unexplained cognitive mutation. I also consider potential counterarguments, such as proposals of specific cognitive thresholds or binary divisions in moral status and find them unpersuasive. I conclude that a gradualist model of moral evolution is both philosophically more coherent and empirically better supported. Human moral agency likely arose through numerous small evolutionary steps rather than a single leap, preserving continuity while accounting for human normativity.

Social epistemic instrumentalism (SEI) argues that epistemic norms derive their authority from serving collective epistemic goals, not individual aims. This paper critically examines whether SEI can explain the categorical (unconditional) authority of epistemic norms. While SEI effectively illustrates how social enforcement of truth-conducive norms constrains agents practically, it faces three vulnerabilities needing explicit attention. First, SEI ties epistemic obligations to communal participation, potentially allowing agents to evade these norms by disconnecting from their community. Second, it permits epistemic free riders—those benefiting from others' truth-seeking without personally adhering to epistemic norms. Third, anchoring normativity solely in social utility risks epistemic relativism, lacking an objective standard to evaluate differing communities' rational inquiry practices. To address these issues, I propose a novel theoretical refinement called "Truth-augmented SEI," integrating truth as an intrinsic epistemic goal. This preserves SEI’s core insight into social mechanisms supporting epistemic norms, yet recognizes truth itself as the ultimate epistemic aim, thus maintaining normative authority even for isolated individuals or truth-indifferent communities.