1. TIME AND CAUSATION: WHERE (AND HOW) DOES CAUSATION ARISE IN THE ARCHITECTURE OF THE COSMOS
Jenann Ismael
University of Arizona

2. "The aim of physics, consciously or unconsciously, has always been to discover what we may call the causal skeleton of the world." Russell, The Analysis of Matter (1955).

3. NEWTON
Causation didn’t appear in the presentation of his theory at all.
Mathematical equations, usually expressed in the form of differential equations that give the rate of change of a quantity over time
There is no direction or asymmetry in the determination.

4. RUSSELL (1913, “ON THE NOTION OF CAUSE)
“Causation is a folk notion that has no place in mature science.’

5. CARTWRIGHT
Dynamical laws cannot play the role of causal relations in science because specifically causal information is needed to distinguish effective from ineffective strategies for bringing about ends.

6. SCM…
provides a precise formal framework for representing causal relationships that is well suited to causal search and discovery in science.
can be used to define normative solutions to causal inference and judgment problems.
has facilitated new insights into the role of causal thinking and cognition.

7. THE FORMALISM Type causation is basic and relative to networks
Networks are defined by collections of variables.
Individual properties are represented as the values of variables.
Singular causal claims are derived from type causation.

8. DAGS (DIRECTED, ACYCLIC GRAPHS)
DAGs represent the causal relations among the variables in a network.
A variable Xi is a direct cause of another variable Xj, relative to a variable set V, just in case there is an intervention on Xi that will change the value of Xj (or the probability distribution over the values of Xj) when all variables in V except Xi and Xj are held fixed.

9. DAGS

10. HOW DOES CAUSAL INFORMATION GO BEYOND THE INFORMATION CAPTURED IN DYNAMICAL LAWS?
Global dynamical laws give us information about relations between global states at different times.
Causal information is information about the results of hypothetical interventions
The antecedents to intervention counterfactuals are nomologically impossible)

11. HOW IS CAUSAL KNOWLEDGE POSSIBLE?
There are various ways of effectively randomizing the values of a variable, so that we can treat changes to the values of some variables as interventions
Causal knowledge rests on the fact that we can effectively isolate systems in the laboratory, manipulate the external input to them, and observe the effects of our manipulations.

12. THE PHYSICAL SUBSTRUCTURE THAT SUPPORTS CAUSAL BELIEF
In a classical world that decomposes into a collection of rule-governed components and the laws governing configurations are given in terms of the laws governing components and their interactions, causal information can be extracted from laws governing components and interactions.
So, facts about the results of hypothetical interventions can be grounded in the local nomological structure of the world

13. CAUSAL PROCESSES (PHIL DOWE)
Causal processes are chains of events related by causal interactions.
An interaction that involves the local exchange of a conserved quantity.
Causal processes can be defined more directly in terms of microdynamical laws.

14. TEMPORAL ASYMMETRY Causal relationships are temporally asymmetric.
The underlying dynamical laws have no intrinsic direction of determination.
How do temporally asymmetric high-level causal pathways arise from temporally symmetric underlying laws?

15. DAVID ALBERT (TIME AND CHANCE)
The account has three first principles:
1. The Newtonian laws of motion
2. A Statistical Postulate: Pr(S is in microstate m/S is in macrostate m*)
3. The Past- Hypothesis: the universe started in a state of low entropy

16. RESERVATIONS
Albert tells the whole story at the global level, but might be better applied at the level of local adiabatically isolated subsystems of the world
There is some dispute about the proper form of the past hypothesis

17. WHICH PAST HYPOTHESIS?
Mathias Frische
David Wallace
It would do just as well to assume a causal Markov condition, microscopic randomness in the initial, but not final state of a system.
It’s notthe entropy, but the randomness or uniformity of the early universe that really matters.

18. THE ASYMMETRY OF AGENCY
If you start with a thermodynamic gradient, and hold fixed all of the information embodied in the present surveyable macroscopic state of the world, it turns out that
the probabilistic effect of local interventions will propagate asymmetrically
Local interventions on the world’s present state will make a difference to the macroscopic future, but leave the macroscopic past discernibly untouched
So strategic routes to brining about ends will run from past to future

19. CAUSAL CONCEPTS Introduce an agent
Couple it informationally to the macroscopic structure of the world
Look to psychology and cognitive science for an understanding of how causal concepts arise, and the role they play in an internal inferential network that is linked to experience at one end and action at the other

20. STRUCTURE AT DIFFERENT LEVELS
the geometry of space- time
the matter content
the temporally symmetric microscopic laws
the emergent macroscopic asymmetries.
creatures that use information to guide behavior, and develop causal concepts

21. TWO QUESTIONS PHILOSOPHERS ASK
Are causes inside or outside the head? Are they part of the mind-independent fabric of reality, or in the eyes of the observer?
Where does the asymmetry come from?

22. THIS ACCOUNT…
secures the objectivity of causal relations, making them apt objects of scientific study, and
is a good antidote to reifying causal relations as compulsive asymmetric relationships between localized events.

23. WHAT ABOUT THE QUANTUM???
If there are causal processes in quantum mechanics, they have to happen at a more fundamental level, and they don’t occur in space- time.

24. THANK YOU!!!