---
title: "PHI403 Lecture 16 — What is Probabilistic Causation?"
type: source
tags: [philosophy-of-science, probability, frequentism, propensity, ecological-fallacy, dispositionalism]
authors: [Anjum, Rani Lill; Rocca, Elena]
year: 2023
venue: "PHI403 Causation in Science, NMBU"
kind: handout
raw_path: "raw/Philosophy of Science/PHI302 16 What is Probabilistic Causation.pdf"
created: 2026-04-20
updated: 2026-04-20
key_claims:
  - Frequentism (Humean) says frequency of occurrence defines probability; it motivates population studies and evidence-based decision-making.
  - Applying a population frequency to an individual case is the ecological fallacy — each individual is treated as the statistical average.
  - Propensity theory (dispositionalist) holds probabilities are intrinsic tendencies; properties produce frequencies, not vice versa.
  - Propensities allow overdisposing — a power can exceed "probability 1" in strength while manifestation is still not guaranteed, because interferers may be present.
---

# PHI403 Lecture 16 — What is Probabilistic Causation?

On **ontological probability**: what makes probability claims *true*? Two main philosophical answers:

- **Frequentism** — probabilities are defined by long-run frequencies. Fits [[concepts/regularity-theory-of-causation|Humean causation]]: a 60 % chance of *E* given *C* just is the fact that *E* follows *C* in 60 % of trials. Observable from data, in principle.
- **Propensity theory** — probabilities are intrinsic tendencies of individuals or situations. Fits [[concepts/dispositionalism|dispositionalism]]: real-world properties *produce* frequencies.

The course's critical move: applying a population frequency directly to an individual case is the **ecological fallacy**. Frequentism presupposes every individual case is the statistical average of its reference population. Choosing narrower sub-groups (all the way to N=1) is then indicated — which leads to singularism and a propensity view.

Not all propensity theorists are singularists. **Long-run propensity theory** (Donald Gillies) holds propensities reveal themselves through frequencies at scale. Still, probability is intrinsic.

**Overdisposing**: a power may have enough strength to guarantee the effect "twice over" while still not manifesting (because interferers can block it). Classical probabilities are bounded by [0, 1]; propensities are not — which is another argument for their ontological distinctness from frequencies.

## Connections
Back-link: [[sources/2023-anjum-rocca-phi403-causation-in-science]]. Concepts: [[concepts/probabilistic-causation]] · [[concepts/dispositionalism]] · [[concepts/rct-limitations]] · [[concepts/regularity-theory-of-causation]].