PSA2016: The 25th Biennial Meeting of the Philosophy of Science Association

Full Program »

Information theoretic tools for the philosophy of causation

The interventionist or manipulability analysis of causation has been extensively used in recent years to address questions right across the philosophy of science 1. The formal underpinnings of this account are given using causal graph theory (CGT) 2. CGT is a way to model causal relationships in complex networks. But events have many causes and we need to distinguish amongst different causes of the same effect, for example, when assessing their relative causal explanatory power. CGT does not in itself provide tools to do this.

Information theory (IT) is the obvious tool with which to enrich CGT in these respects. There is a significant body of work in complex systems science on how to add IT to CGT in order to model additional features of causation 3–5. Amongst the many degree properties of causes that it would be desirable to have formal measures of are proportionality, specificity, stability 6.

In this poster we exemplify the power of information theory to clarify philosophical questions about causation, with the aim of encouraging wider uptake of these tools. The poster will contain accessible, graphical presentations of an IT analysis of causal specificity 7 and of two useful results which can be obtained with this measure:

(1) The interventionist criterion of causation is derived as a special case: C causes E when there is at least one background in which C has non-zero specificity for E.
(2) We can precisely distinguish the causal power of one variable with respect to another from the actual effect of one variable on another on a particular occasion (actual causation). These are captured by measuring causal specificity with two different probability distributions over the cause. The measure of actual causation is formally equivalent to an existing measure, ‘information flow’, in the complex systems science literature 4.

Brief indications will be provided of how similar precision can be added to the concepts of proportionality and stability, and to topics such as ‘causal preemption’ 8. The authors will be available to expand on these indications at the poster session.

References
1. Woodward, J. Making things happen: A theory of causal explanation. (Oxford University Press, 2003).
2. Pearl, J. Causality: Models, Reasoning, and Inference. (Cambridge University Press, 2009).
3. Tononi, G., Sporns, O. & Edelman, G. M. Measures of degeneracy and redundancy in biological networks. Proc. Natl. Acad. Sci. 96, 3257–3262 (1999).
4. Ay, N. & Polani, D. Information flows in causal networks. Adv. Complex Syst. 11, 17–41 (2008).
5. Korb, K. B., Hope, L. R. & Nyberg, E. P. in Information Theory and Statistical Learning (eds. Emmert-Streib, F. & Dehmer, M.) 231–265 (Springer US, 2009).
6. Woodward, J. Causation in biology: stability, specificity, and the choice of levels of explanation. Biol. Philos. 25, 287–318 (2010).
7. Griffiths, P. E. et al. Measuring Causal Specificity. Philos. Sci. 82, 529–555 (2015).
8. Glymour, C. et al. Actual causation: a stone soup essay. Synthese 175, 169–192 (2010).

Author Information:

Paul Griffiths    
Dept. of Philosophy and Charles Perkins Centre
University of Sydney

Arnaud Pocheville    
Dept. of Philosophy and Charles Perkins Centre
University of Sydney

 

Powered by OpenConf®
Copyright©2002-2015 Zakon Group LLC