Hume on Causation
Hume on Causation is Hume’s skeptical analysis of what we mean, and what we can know, when we say A causes B.
Why is causation problematic?
If “causation” is a real feature of the world, we should be able to point to what we observe when we see one. Hume argues we cannot.
When a billiard ball strikes another and the second moves, we observe only:
- Spatial contiguity: A and B were in contact
- Temporal succession: A preceded B
- Constant conjunction: events like A are reliably followed by events like B
We never observe a “causal connection” directly. The necessity we feel is a habit of mind, not a feature of the world.
Implications
- No number of observations proves causation; a gap remains between “always followed by” and “necessarily produces”
- Inductive inference cannot be deductively justified
- This grounds the ampliative character of all causal reasoning
Since causation is not observed directly, we need methods (Mill’s Methods, controlled experiments) to infer it from patterns of correlation plus experimental manipulation.