Kazuyoshi Kamiyama

Anthropomorphism is often dismissed as a cognitive bias or as a remnant of pre-scientific thinking. This paper argues that such critiques rely on an implicit and problematic assumption of direct, unmediated access to objects. Against this background, I introduce the concept of Model-Mediated Anthropomorphism (MMA), defined as a cognitive operation in which a subject understands an object through the mediation of a human model. MMA is characterized by a triadic structure—self, human model, and object—and should be understood not as a psychological illusion but as a constitutive epistemic interface for meaning formation. Crucially, the human model does not faithfully represent its object. It necessarily distorts, simplifies, and sometimes fails. I argue that this failure is not a defect but an essential feature of MMA: the resistance of the object functions as a negative feedback through which understanding is revised and deepened. This dynamic can be illuminated by Hegel’s notion of mediation and reflection, according to which cognition advances through the externalization and negation of its own models. Reframing anthropomorphism in this way enables a reassessment of contemporary debates on artificial intelligence and ethical anthropomorphism, where rigid human models risk both ethical violence and epistemic breakdown. This text is an extended abstract for an upcoming research presentation.

The concept of common knowledge has played a central role in game theory and mathematical economics, particularly in connection with equilibrium analysis. Aumann’s formal definition of common knowledge, based on information partitions, is widely regarded as providing a precise counterpart to the intuitive notion introduced by Lewis. However, Aumann’s definition appears to rely on an implicit assumption: that the structure of players’ information partitions themselves is common knowledge. Aumann has argued that this assumption is not substantive but follows trivially from the notion of a “state of the world,” whose description is taken to be sufficiently rich. This paper critically examines Aumann’s intuition and the structure of his argument. I argue that the inference from a rich description of states of the world to the conclusion that information partitions are necessarily common knowledge is not adequately justified. In particular, the appeal to fully comprehensive descriptions of states involves problematic assumptions, including implicit self-reference and an overly strong conception of descriptive completeness. I conclude that whether information structures are common knowledge cannot be settled by definitional considerations alone. Rather, common knowledge of information structures should be treated as a substantive assumption whose validity depends on specific epistemic and social conditions.

This paper provides a survey of epistemic approaches to game theory, focusing on the roles of cognitive models and common knowledge in the analysis of game solutions. It reviews how players’ beliefs, higher-order beliefs, and assumptions about rationality are formally represented, and how these epistemic structures underpin standard solution concepts such as Nash equilibrium and its refinements. Particular attention is paid to Aumann’s notion of common knowledge and its significance for coordination, agreement, and equilibrium reasoning. By clarifying the relationships between belief models, informational assumptions, and solution concepts, the paper aims to offer a systematic overview of the epistemic foundations of game-theoretic analysis.

According to Hume’s skeptical argument, our empirical beliefs that go beyond sense impressions and memory lack justification, apparently forcing us into complete skepticism. This is the classical problem of induction. In this paper I propose a new dissolutionist solution, in the tradition of Strawson (1952), Goodman (1955), and Okasha (2001). By criticizing the core structure of Hume’s skeptical argument, I argue that Hume’s skepticism does not succeed in undermining the soundness of induction. Consequently, we are not required to justify induction in general.