Mathematical tools are increasingly being applied to domains that once seemed un-quantifiable, including conservation policy and international governance. Recent advances in data availability and natural language processing have created new opportunities to examine how institutions make decisions and respond to environmental change. We should expect to see a proliferation of these approaches in the future, and so it’s an opportune time to consider where the field of quantitative policy analysis is heading, and what sorts of interpretive guardrails are needed.

I will begin by situating these recent developments within the longer history of mathematics extending into new domains. Mathematical approaches can bring precision and predictive ambition to problems that are otherwise approached through qualitative methods. However, these benefits come with costs. Quantification necessarily depends on abstraction, on removing nuance and detail. In physical and biological systems, abstractions can reveal underlying and general mechanisms, such as gravity, diffusion, natural selection. In human decision-making, it is less clear that comparable mechanisms exist, or that they can be separated from their political, legal, and institutional contexts.

Using recent examples from Antarctic governance, I will argue that quantitative analysis should not and cannot replace qualitative expertise, but must rather work in tension with it. Mathematical modelling can ask sharp questions, but the proposed answers must be circumscribed by qualitative insights. The central challenge is therefore interdisciplinary: how to use quantitative methods to reveal structure in institutional behaviour, without allowing abstraction to obscure the context that may determine outcomes.