Modelling in Philosophy- Workshop

Programme

Tuesday 8 December

Albano, Hus 3, 2nd floor

• 12:15–13:45 Lunch
• 13:45–15:00 Nicolas Cote, Models of Virtue
• 15:00–15:30 Break
• 15:30–16:45 Nina Emery, Models in Science and Models in Philosophy: Methodological Naturalism at Work

Wednesday 9 December

Campus Frescati, Gula Villan, room Minerva

• 09:00–09:30 Coffee
• 09:30–10:45 Cailin O’Connor, Many Paths to Polarization
• 10:45–11:00 Break
• 11:00–12:15 Facilitated discussion (Dunja Šešelja)
• 12:15–13:45 Lunch
• 13:45–15:00 Lukas Beck, Policy Optimization IAMs as Normative Models
• 15:00–15:30 Break
• 15:30–16:45 Tarja Knuuttila, Statistical signature or theoretical mechanism? Template pluralism in scale-free network modelling
• 19:00 Dinner

Thursday 10 December

Campus Frescati, Gula Villan, room Minerva

• 09:00–09:30 Coffee
• 09:30–10:45 Catherine Herfeld and Edoardo Peruzzi, Model Transfer as Model Construction: A Taxonomy
• 10:45–11:00 Break
• 11:00–12:15 Facilitated discussion (Finnur Dellsén and James Nguyen)
• 12:15–13:15 Lunch
  

Abstracts

Nicolas Cote, Models of Virtue

It is common in normative ethics to model the point of view of idealized agents and ask oneself: “what would this person do? What would they want?” This exercise can serve different purposes: in virtue ethics it is a user-end instrument for generating case-specific advice; in consequentialist ethics it is a theoretical tool for deriving general moral principles. In this presentation, I trace the use of these “models of virtue” across different ethical traditions, where they are often of foundational importance, and elucidate their varying uses. I argue we can find at least three distinct families of models of virtue in the history of the discipline. In order of increasing abstraction, these are: (1) exemplar simulations, (2) transcendent experts models, and (3) ideal-self models. These families differ in their methodological profile, moral epistemology, and the data they rely upon, and they suggest interestingly different strategies for dealing with the problem of unreliable moral intuitions.

Nina Emery, Models in Science and Models in Philosophy: Methodological Naturalism at Work

Methodological naturalism is the view that philosophers should, whenever possible, use the same methodology that scientists use. In this talk, I review the argument for methodological naturalism from my book Naturalism Beyond the Limits of Science (Oxford University Press, 2023), and discuss several important lessons that follow from that argument. In particular, I note that the argument leaves open the possibility of what I call methodological extension, which would permit philosophers to adopt a methodology that is consistent with, but goes beyond, the methodology of science. I also discuss the way in which methodological pluralism in science might generate methodological pluralism in philosophy. I then explore the idea that the methodological naturalist should extend the use of modeling in science to various philosophical domains.

Cailin O’Connor, Many Paths to Polarization

In this talk I use network models to demonstrate how a wide range of causal mechanisms are sufficient to promote belief polarization. This collection demonstrates a number of things about the use of models in the study of social phenomena. As they show, models can play multiple roles in inference, even in the same project - ranging from demonstration of possibility, to expansion of the space of testable hypotheses, to the isolation of causal variables that are not isolable in real systems, to supporting arguments against over-simplified conclusions in empirical studies. The talk also shows how helpful it can be to consider multiple models in concert, especially when investigating highly complex social dynamics.

Lukas Beck, Policy Optimization IAMs as Normative Models

An increasing number of philosophers regard so-called normative models as crucial tools for pursuing normative questions, underscoring their value in disciplines ranging from normative ethics and formal epistemology to decision theory, computer science, and economics. In this talk, I argue that tools developed in the debate about normative models can be fruitfully applied to controversies around Policy Optimization Integrated Assessment Models (PO-IAMs). In the context of climate change mitigation policy, Integrated Assessment Models combine economic models with bio-geo-physical models. While different IAMs serve different purposes, PO-IAMs utilize the tools of welfare economics to provide us with an explicit ranking of different policy options for tackling climate change. A central theoretical issue related to those models concerns the rate of pure time preference, i.e., how much evaluative weight should be assigned to utility occurring at different points in time. While Stern can be seen as relying on moral theory to set the rate of pure time preference, I will argue that an alternative approach, favored by Arrow, can be reconstructed and defended by drawing on recent work on normative models, specifically the extending account developed by Beck and Jahn.

Tarja Knuuttila, Alexander Gschwendtner, and Andrea Loettgers, Statistical signature or theoretical mechanism? Template pluralism in scale-free network modelling

Paul Humphreys’ notion of a template has become a central unit of analysis for the study of cross-domain model transfer, yet the notion soon diversified (Humphreys 2002, 2004, Herfeld 2024). In Humphreys’ work, his original notion of computational templates gave way to formal and transdomain templates, and the subsequent literature has added customised templates (Houkes 2023) and model templates (Knuuttila and Loettgers 2023), both of which also address the conceptual content a template is taken to carry. Are all these template notions needed? We study this question through scale-free networks. While Humphreys (2019) used the Barabási-Albert model as an example of a formal template with no interpretation beyond the mathematical one, Knuuttila and Loettgers treat it as a model template entangled with concepts from statistical mechanics. We suggest that the question of whether the scale-free networks are formal or model templates cannot be solved at the level of definition. The two arguments Barabási and collaborators gave for scale-free networks, a statistical signature recurring in data and a theoretical grounding in percolation theory, correspond to two modes of use that map onto the formal and model template notions. Classifying approximately 1,500 papers with an LLM-based pipeline, we show that the distinction organises much of the literature, with a grey zone of mixed uses of philosophical interest in its own right. Templates carry conceptual content; what varies is the role that content plays.