The Philosophical Intuition and Conceptual Structure Workshop

Abstract: When faced with immoral statutes, people are often drawn to statements such as: “There is one sense in which this statute is a law, but ultimately, it is no law at all”. What best explains this contradictory-sounding linguistic intuition and what are its implications for the study of legal argumentation? According to one influential theory, saying that an unjust statute is not a law reflects a proposal to change the meaning of the word “law” so that it no longer picks out grossly unjust statutes. Thus, although the statement employs descriptive language (about what the law “is” and “isn’t”), what it really expresses is a normative proposal in a negotiation about language. According to an alternative account, the acceptance of the contradictory-sounding statement tells us something important about our shared concept of law. In particular, it tells us that the concept of law has a dual character, meaning that there are two different and relatively independent criteria for something to be a law. One of the criteria is descriptive in nature, while the other is normative. Thus, whenever a statute fulfills one criterion, but not the other, we are drawn to statements that express this conflict. The present talk compares the explanatory merits of the two alternatives.

Abstract: The reasonable person standard is central to ethics and law, particularly in assessing negligence and responsibility. Previous research shows that judgments of what is reasonable typically fall between what is average and what is ideal (Bear & Knobe, 2017; Tobia, 2018; Tobia et al., 2024). We examine whether the same pattern applies to AI, particularly large language models. Given evidence on algorithm aversion and automation bias, we hypothesized that people set stricter standards for AI than for humans. Specifically, we predicted that (H1) reasonable judgments would lie between average and ideal for both agents, and (H2) AI reasonableness would be judged closer to the ideal. In a 2 (agent: AI vs. human) × 3 (average, reasonable, ideal) between-subjects experiment, participants evaluated assertion and non-assertion scenarios. Results supported the first hypothesis: reasonableness consistently fell between average and ideal, but was not significantly closer to the ideal for AI than for humans (only numerically).

Abstract: Ordinary judgments about whether someone is dead may not depend on a fixed set of necessary and sufficient conditions. Across a series of studies, I test whether cases of death vary in typicality and whether judgments reflect the combined influence of consciousness, vital functioning, and bodily integration. These studies aim to clarify whether the ordinary concept of death has a graded, multidimensional structure rather than a simple rule-based one.

Abstract: Truth is frequently assumed to be conceptually straightforward. However, empirical research reveals substantial variation in truth judgments across contexts, experimental tasks, and individuals. In this talk, I examine the factors that pull truth intuitions in different directions and argue that understanding the structure of our concepts of truth is essential for making progress on these questions.

Abstract: The method of cases assumes that philosophical intuitions track classically structured concepts –i.e., categories with necessary and sufficient conditions. Yet many everyday concepts instead show a prototype structure (e.g., Zaki et al., 2003). We ask whether this holds for core philosophical concepts, such as free will, moral responsibility, knowledge or personal identity. In a pre-registered study, we combine three methods: (1) typicality ratings, testing whether paradigmatic instances of a concept are judged more central than atypical ones; (2) reaction time, testing whether reaction times are slower for low-similarity cases, as prototype models predict; and (3) individual-level, non-nested model comparison, to classify participants’ concepts as criterion-based or similarity-based. In this talk, we will present the results and discuss their implications for the method of cases.

Abstract: TBA