Causation Flashcards
What are the arguments for the question: “‘An event A is a cause of distinct event B just in case bringing about the occurrence of A would be an effective means by which a free agent could bring about the occurrence of B.’ Is this a sound analysis of causation?
- M&P’s Agency Theory
- Objections
What does Cartwright claim about the Russelian analysis of causation?
- We can’t get causation from association but we still need causation to explain why A, and not B, is an effective strategy to produce C.
What is Cartwright’s theory?
Probablistic Theory of Causation.
- C causes E iff C raises the probablity of E in every background context B.
- Realist about causation.
How can we formulise Cartwrights example of correlation not being sufficient?
Smoking = SM
Exercise = E
Heart Disease = HD
If P(HD|SM) < P(HD|¬SM)
- Then it seems SM prevents HD.
But this is because SM and E are highly correlated, therefore we must hold fixed E as it is another causal factor of HD:
P(HD|SM ⋀ E) > P(HD|¬SM ⋀ E)
and vis versa.
Therefore, SM does not prevent HD.
What example shows Cartwright’s point?
Smoking and Exercise leading to a longer life.
What is M&P’s Agency Theory?
- C causes E iff bringing about C is an effective means for a free agent to bring about E.
- Effective means is defined in terms of raising probabilities from the perspective of the agent.
- Chooses Anthropocentric over Circularity.
- Anti-Realist
Benefits of M&P’s Agency Theory?
Not circular because our experience of ‘bringing about’ (agency) precedes that of causation.
Objection to M&P’s Agency Theory
Unmanipulable causes:
- It does not account for causes we cannot manipulate
- Such as an Earthquake.
Rebuttal to the Objection against M&P’s Agency Theory?
Price claims that we understand the causal processes of unmanipulable causes with models.
Objection to the Rebuttal for M&P’s Agency Theory?
Models must resemble reality.
- Models can get the causal processes wrong. (How?)
- Therefore, models must resemble the causal processes of reality.
- This makes Agency Theory non-reductive.
Rebuttal to the Objection to the Rebuttal for M&P’s Agency Theory?
Price claims that we can infer the causal processes of reality from non-causal features of reality that are resembled in models.
What example does Price use to show that we must be able to infer from non-causal features the causal features of reality?
Different temporal realities in which we would have no notion of how to manipulate things.
What does Price’s example mean for error and our notion of causality?
We are able to be wrong about what things cause other things because our notion of causation for unmanipulable causes is provisional.
Our notion of causation is inherently antirealist as it is based on what we know.
What is the effect of the unmanipulable causes objection to Woodward according to Price?
Our understanding of intevention is also dependent on how much we know and our principles of extension, just as much as Agency Theory.
If not he is left arguing for an implausible objectivity in these cases.
What does Price mean by principles of extension?
The rules that govern how we are able to extend features from models to reality.
Conclusion to “‘An event A is a cause of a distinct event B just in case bringing about the occurrence of A would be an effective means by which a free agent could bring about the occurrence of B.’ Is this a sound analysis of causation?
If we are willing to accept that causation is a secondary quality, which Price’s example of temporally different realities suggests, then Agency Theory is compelling analysis of causation.
What are the arguments for the question: Does anything cause anything else?
- Russell’s Causal Scepticism
- Cartwright’s Effective Strategies
What is the law of causation proposed by Russell?
“…whenever E1 occurs, E2 follows after interval T.”
What is Russell’s Dilemma?
Either E1 takes time and doesnt cause E2 or E1 doesn’t take time and doesn’t cause E2.
Explain Russell’s Dilemma.
If E1 takes time, then it is only the last part that cause E2.
- E1 is infinitely divisible, so there cannot be a last part.
If E1 doesn’t take time, this doesn’t exist and is too weird.
