---
title: "PHI403 Causation in Science — NMBU PhD Course (Anjum & Rocca, 2023)"
type: source
tags: [philosophy-of-science, causation, course, nmbu, hub, dispositionalism, pluralism]
authors: [Anjum, Rani Lill; Rocca, Elena]
year: 2023
venue: "PHI403 Causation in Science, NMBU"
kind: course
raw_path: "raw/Philosophy of Science/"
created: 2026-04-20
updated: 2026-04-20
key_claims:
  - Philosophical bias is unavoidable in science — scientific methods carry implicit ontological, epistemological and normative commitments about causation (Andersen, Anjum, Rocca 2019, eLife).
  - Causation is real and indispensable to applied science; Russell-style causal scepticism collapses once we consider intervention, observation, and application.
  - No single theory of causation (regularity, difference-making, manipulability, mechanism, probability, disposition) captures all cases; each picks out a "symptom" of causation.
  - The course defends ontological causal monism (causation is one thing, primitive and unanalysable) combined with methodological pluralism (many methods are needed to triangulate its symptoms).
  - The EBM evidence hierarchy that privileges RCTs over mechanisms is itself a philosophical bias; RCTs systematically exclude causally relevant information (severe effects, risk groups, individual variation, negative results).
  - Dispositionalism — causes as tendencies/powers — is the course's preferred positive ontology; it explains context-sensitivity, singular causation, non-transitivity, and stability (equilibrium, homeostasis).
---

# PHI403 Causation in Science — NMBU PhD Course (Anjum & Rocca 2023)

PhD course taught at [[entities/nmbu|NMBU]] (Norwegian University of Life Sciences) by [[entities/rani-lill-anjum]] and [[entities/elena-rocca]]. The raw material is a series of 20 two-page lecture handouts (dated 2023), filed under `raw/Philosophy of Science/`. Filenames carry the course code `PHI302`, but the actual course code is **PHI403** (user confirmation).

*Note on duplicates*: lectures 12 and 13 have `-1` variants in the raw folder; these are whitespace-only duplicates and are ignored. Content is drawn from the non-`-1` files.

## Course arc

The course moves through four connected stages:

1. **Metascience and the role of philosophy in science** (L1–2). Science carries implicit non-empirical norms; [[concepts/philosophical-bias]] is unavoidable. Russell-style dismissal of causation is rebutted: experiment, observation, and application all presuppose causation.

2. **Theories of causation** (L3–9). Ontology vs epistemology ([[concepts/causation]]); the [[concepts/regularity-theory-of-causation|regularity theory]] (Hume) and its rescue strategies for exceptions, outliers, and ideal conditions; monocausality vs multi-causality (INUS, mutual manifestation partners); additive vs subtractive interference ([[concepts/interventionist-theory-of-causation]]); the shift from regularities to tendencies ([[concepts/dispositionalism]]).

3. **Data, mechanisms, and levels** (L10–14). Theory-ladenness of data (Hanson, Leonelli); Humean universalism vs dispositionalist singularism and the problem of [[concepts/rct-limitations|external validity]] (Cartwright & Hardie on India vs Bangladesh); explanatory power of [[concepts/mechanisms-in-science|mechanisms]] and the Russo-Williamson thesis; reductionism, emergence, holism, demergence; causal chains, determinism, non-transitivity.

4. **Probability, RCTs, and methodological pluralism** (L15–20). Credence vs chance; frequentism vs propensity ([[concepts/probabilistic-causation]]); philosophical pathologies of conditional probability (ratio formula, material conditional); external validity and overfitting in prediction; [[concepts/rct-limitations|RCT limitations]] (ecological fallacy, publication bias, policy non-derivability); the positive programme of ontological monism + [[concepts/methodological-pluralism|methodological pluralism]] ([[concepts/causal-pluralism]]).

The underlying argument is that **scientific method is never philosophically neutral** — every method assumes a theory of causation, and every theory of causation has cases it handles well and cases it handles badly. The only way to avoid being misled is to make the commitments explicit and to use plural methods attuned to different symptoms of causation.

## Lecture index

| # | Lecture | Page |
|---|---|---|
| 01 | Metascience and Better Science | [[sources/2023-anjum-rocca-phi403-lecture-01-metascience]] |
| 02 | Do We Need Causation in Science? | [[sources/2023-anjum-rocca-phi403-lecture-02-do-we-need-causation]] |
| 03 | Evidence Based on What? | [[sources/2023-anjum-rocca-phi403-lecture-03-evidence-based-on-what]] |
| 04 | What's in a Correlation? | [[sources/2023-anjum-rocca-phi403-lecture-04-whats-in-a-correlation]] |
| 05 | Same Cause, Same Effect? | [[sources/2023-anjum-rocca-phi403-lecture-05-same-cause-same-effect]] |
| 06 | Under Ideal Conditions | [[sources/2023-anjum-rocca-phi403-lecture-06-under-ideal-conditions]] |
| 07 | One Effect, One Cause? | [[sources/2023-anjum-rocca-phi403-lecture-07-one-effect-one-cause]] |
| 08 | Have Your Cause and Beat It | [[sources/2023-anjum-rocca-phi403-lecture-08-have-your-cause-and-beat-it]] |
| 09 | From Regularities to Tendencies | [[sources/2023-anjum-rocca-phi403-lecture-09-from-regularities-to-tendencies]] |
| 10 | Is it the Business of Science to Construct Theories? | [[sources/2023-anjum-rocca-phi403-lecture-10-business-of-science]] |
| 11 | Is More Data Better? | [[sources/2023-anjum-rocca-phi403-lecture-11-is-more-data-better]] |
| 12 | The Explanatory Power of Mechanisms | [[sources/2023-anjum-rocca-phi403-lecture-12-explanatory-power-of-mechanisms]] |
| 13 | Digging Deeper to Find the Real Causes | [[sources/2023-anjum-rocca-phi403-lecture-13-digging-deeper]] |
| 14 | It All Started With a Big Bang | [[sources/2023-anjum-rocca-phi403-lecture-14-big-bang]] |
| 15 | Credence: Probability from Uncertainty | [[sources/2023-anjum-rocca-phi403-lecture-15-credence]] |
| 16 | What is Probabilistic Causation? | [[sources/2023-anjum-rocca-phi403-lecture-16-probabilistic-causation]] |
| 17 | Calculating Conditional Probability? | [[sources/2023-anjum-rocca-phi403-lecture-17-conditional-probability]] |
| 18 | Risky Predictions | [[sources/2023-anjum-rocca-phi403-lecture-18-risky-predictions]] |
| 19 | What RCTs Do Not Show | [[sources/2023-anjum-rocca-phi403-lecture-19-what-rcts-do-not-show]] |
| 20 | Plural Methods, One Causation? | [[sources/2023-anjum-rocca-phi403-lecture-20-plural-methods-one-causation]] |

## Connections

**Concepts (created / updated for this ingest):**
- [[concepts/causation]] — hub concept
- [[concepts/causal-pluralism]] — monism vs pluralism axes
- [[concepts/regularity-theory-of-causation]] — Hume
- [[concepts/interventionist-theory-of-causation]] — Woodward / manipulation
- [[concepts/dispositionalism]] — Anjum / Mumford / Cartwright / tendencies and powers
- [[concepts/mechanisms-in-science]] — Russo-Williamson, Semmelweis, homeostasis
- [[concepts/probabilistic-causation]] — credence vs frequentism vs propensity
- [[concepts/rct-limitations]] — ecological fallacy, external validity, negative results
- [[concepts/methodological-pluralism]] — Feyerabend, Longino
- [[concepts/evidence-hierarchy]] — the EBM pyramid and its critique
- [[concepts/philosophical-bias]] — Andersen, Anjum, Rocca 2019 (eLife)

**Entities:**
- [[entities/rani-lill-anjum]] — course lead, dispositionalism, CauseHealth
- [[entities/elena-rocca]] — course lead, philosophy of pharmacology
- [[entities/nmbu]] — Norwegian University of Life Sciences

**BPM / PPM cross-links:**
- [[concepts/predictive-process-monitoring]] — tendency-framed predictions (L18)
- [[concepts/outcome-prediction]] — probabilities, external validity
- [[concepts/causal-process-discovery]] — mechanism vs regularity in event logs
- [[sources/2020-rama-maneiro-deep-learning-ppm-review]] — external-validity / benchmark critique applies
- [[sources/2021-dumas-process-mining-2-from-insights-to-action]] — prescriptive monitoring as interventionist causal claim