Causality. Models, reasoning, and inference. (English) Zbl 0959.68116

Pearl, Judea

Cambridge: Cambridge University Press. xvi, 384 p. (2000).

This book provides a systematic, mathematically founded account of causality and causal reasoning. Following a description of the conceptual and mathematical advances in causal inference, the book emphasizes practical methods for elucidating potentially causal relationships from data, deriving causal relationships from combinations of partial knowledge and data, predicting the effects of actions and policies, evaluating explanations for observed events and scenarios, and – more generally – identifying and explicating the assumptions needed for substantiating causal claims. The book starts with a summary of the elementary background in probability theory and graph notation needed for the understanding of this book, together with an outline of the developments of the last decade in graphical models and causal diagrams (Bayesian networks, causal Bayesian networks). The author then turns to the outline of a theory of inferred causation, i.e., the question of how one can go about discovering cause-effect relationships in raw data and what guarantees one can give to ensure the validity of the relationships thus discovered. Next, questions of identifiability are discussed, namely, predicting the direct and indirect effects of actions and policies from a combination of data and fragmentary knowledge of where causal relationships might operate. The implications of these findings for the social and health sciences are then discussed, where the author examines the concepts of structural equations and confounding. The subsequent chapter offers a formal theory of counterfactuals and structural models, followed by a discussion and a unification of related approaches in philosophy, statistics, and economics. Applications of counterfactual analysis are then pursued, where methods are developed for bounding causal relationships, and applications to imperfect experiments, legal responsibility, and the probability of necessary, sufficient, and single-event causation are illustrated. The book ends with an epilogue, a transcript of a public lecture, which provides a gentle introduction of the historical and conceptual aspects of causation.

Reviewer: Udo Hahn (Freiburg i.Br.)

MSC:

68T20	Problem solving in the context of artificial intelligence (heuristics, search strategies, etc.)
68-02	Research exposition (monographs, survey articles) pertaining to computer science
03B48	Probability and inductive logic
68T30	Knowledge representation
68T37	Reasoning under uncertainty in the context of artificial intelligence

Keywords:

Bayesian networks; causality; causal reasoning; causal Bayesian networks; counterfactuals

Software:

CausalFX