lecture 11 Flashcards
why bother about causation
understanding causes matters, because it allows us:
To intervene and predict, so that we can avoid mistakes (as much as possible) and delive desired outcomes instead
To explain why or how things happen
Philosophers have typically worried about 2 sets of questions regarding causality
1) how can we make reliable causal inferences
2) What is causality
What is causality
No agreement among philosophers
still different theories on the table
Causality is…
regular association
difference-making and manipulability
energy transference
dispositions / tendencies
Causality is…
Regular association
Difference making and manipulability
energy transference
Dispositions/tendencies
How can we make reliable causal inferences
Scientists use experiments and models to investigate causal relations
How can we make reliable causal inferences (take home message)
Many alternative metaphysical accounts of causation
Still we have a meaningful discourse about how to make reliable causal inferences by using a minimal account of causation
This minimal account of causation includes the notions of causality as difference-making (or counterfactual reasoning) and as manipulability (or intervention)
Association
two variables C and E are associated (or dependent) when one variable provides information about the other
Correlation
two variable display an increasing or decreasing trend
Causation
many possible definitions; here we go with a minimal one: association + manipulability
Temporal succession
C regularly comes before E
Contiguity
C and E happen nearby in space
Conditional dependence
P(EIC) != P(E) and P(CIE) != P(C)
which kind of information we get from association
Causes and effects are discoverable not by reason but by experience
Problems with regular association view
1) some variables are causally related, but not (spatially or temporally) contiguous
E.g. The lockdown in china caused a drop of oil prices in saudi arabia, though the two events are not spatially (or temporally) contiguous
2) causal relationships are asymmetric
C causes E but E does not cause C
E.g. The lockdown in china caused a drop of oil prices in saudi, but the drop of oil prices in saudi did not cause the lockdown in china
3) problems with regular association view
Some variables are conditionally dependent/associated/correlated but not causally related
E.g.Whenever ice cream sales are high, incidence of drowining is also high. Do ice creams cause drownig? no! both are effects of a common cause, namely warm weather
Manipulability
two variables C and E are causally related when, if the value of C changed, the value of E would change too
Basic idea: If C and E are only associated or correlated, then intervening on C will not change the value of E
what is a common cause
An observed dependence between two variables X and Y is indicative of either X causing Y or Y causing X or the existence of a COMMON CAUSE Z
What is a common cause
in the case of ice cream consumption and drowining rates, both have a common cause namely hot weather
However that hot weather is a common cause of these two events is an inference
Hot weather works here as a commmon cause which explains the existing relationship between two events
Common causes and bayesian networks
The inference from the association of x,y to the common cause z will be assigned a probability value
The causal links z -> x and z -> y can also be assigned a probability value
There will be background assumptions justifying x,y -> z
In complex situations, one can use bayesian networks/ causal models to
make the assumptions behind causal inferences explicit
Ask justification for any particular set of assumptions, and investigate what happens if an assumption fails
Develop algorithms for causal discovery