In Causation: A Defense of a Non-Reductionist Approach, Michael Tooley offers detailed criticism of various approaches to understanding causation and makes an argument for the superiority of a theoretical-term, non-reductionist analysis of causation.He begins by offering detailed criticisms of alternative approaches, including the competing non-reductionist view that no analysis of the concept of causation is needed, since the relation of causation is directly observable, thereby entailing that the concept of the relation of causation is analytically basic. In response, Tooley argues that the relation of causation is not directly observable. His argument then considers reductionist approaches to causation, which can be divided into those that accept David Hume's thesis that there can never be logical connections between distinct existents, and those that reject that thesis. In the case of the former, Tooley outlines and criticizes at length accounts that attempt to analyze causation in terms of laws of nature, counterfactual approaches, a variety of probabilistic accounts, analyses in terms of agency, and conserved quantity accounts. Here Tooley offers both specific, detailed objections to each approach, and powerful general arguments that warn against any Humean-style reductionist analysis. Finally, the book discusses non-Humean-style approaches that attempt to analyze both causation and laws of nature in terms of dispositional properties. Tooley argues that the idea of intrinsic, irreducible dispositional properties leads to a contradiction.Clearly outlining the faults in other approaches, the book concludes that a very simple and sound analysis of causation can be given if the relation of causation is viewed as a theoretical relation between events.