1. Approaches to Explanation
I begin with a survey of various approaches to explanation. Special emphasis is placed on the pattern subsumption approach (a more general version of the unification approach), which will be useful later. Famous cases of asymmetry (flagpole/shadow and so on) suggest a causal account. I specify the minimal properties required of causality if it is to provide a solution to the asymmetry problems, and examine some accounts of a
causal influence relation that have these properties. Any such relation will do – I remain agnostic about foundational metaphysics throughout the book.
2. Causal and Explanatory Relevance
A minimal causal account of explanation is presented – an account that makes the minimal possible use of causality to solve the asymmetry problems. It has its own difficulties: it is far too liberal in what it counts as explanatorily relevant. An additional criterion for relevance is required. Some familiar candidates are considered: probabilistic relevance, counterfactual dependence, the criteria for actual causation favored by manipulationists. All fall short.
3. The Kairetic Account of Difference-Making
A new criterion for explanatory relevance is developed. In conjunction with a (still very minimal) notion of causality, here given some further structure but no particular metaphysics, it constitutes the kairetic notion of difference-making, the causal relation in virtue of which one state of affairs bears causal-explanatory relevance to another.
4. The Kairetic Account of Explanation
A recipe is developed for putting together assorted explanatorily relevant states of affairs into what I call a standalone explanation, the sort of thing that science aims to construct, and in virtue of which, once constructed, it considers itself to understand a phenomenon. Many different dimensions are explanatory goodness are then considered. Some are illusory; some are not. Stepping back, I consider the significance of causal explanation for human existence.
5. Extending The Kairetic Account
Many enhancements to the kairetic account.
First, a notion of explanatory weight – that is, a quantification of explanatory relevance.
Second, an exploration of
rough explanation, explanation that does not quite get the explanandum right.
Third, the idea of an explanatory framework, a background against which explanations are constructed.
Black boxes in explanatory models are to be valid only within a framework – though when valid, are useful for a number of reasons.
Fourth, from black boxes to multiply realizable explanatory constructs in general, such as the explanatory use of functionally defined predicates, which are handled in the same way as black boxes.
Fifth, aggregative explanation, that is, the explanation of a relational fact holding between the outcomes of two or more independent (or partially independent) causal processes.
Sixth, contrastive explanation, and an argument that not all explanation is contrastive.
Seventh, a grand extension of the kairetic account of causal explanation in science to non-causal explanation in science, to mathematical explanation, and then to philosophical, moral, literary explanation … though much work remains to be done!
6. Event Explanation and Causal Claims
To say that one event is a cause of another event, I propose, is just to say that one is a part of the causal explanation of the other. The question of the truth conditions for causal claims – claims of the form c is a cause of e – is more or less identical, then, to the question of the criteria for a good event explanation. If this is correct, then many problem cases in the causation literature ought to be resolvable by a good account of event explanation. I show that the kairetic account can handle a number of such cases: preemption, symmetric overdetermination, absences as causes, the status of prevention as a causal relation, and the question of the transitivity of event causation (which turns out to be transitive after all).
7. Regularity Explanation
The causal explanation of laws, robust regularities,
effects, and so on (I use the term
regularity to encompass all such explananda). A simple causal approach is sketched that I take to be relatively uncontroversial. Big problem: it works only for a small minority of regularities. It fails, in particular, to handle contingent regularities. But almost all interesting laws are contingent, from the various manifestations of the second law of thermodynamics on up. Something new is needed. It is here that I introduce a new, non-causal relation of explanatory relevance: entanglement, a kind of
two-way modal robustness (characterized further in this paper). It is this relation that holds between the antecedent properties in a contingent law (the if part of an if/then) and the properties that do the causal work. Much of the chapter is given over to understanding the properties of entanglement and its ensuing explanatory role. After some applications to particular cases, I go metaphysical, extracting an account of the nature of special science laws from the kairetic account of their explanation – applying the principle that a law's constituents are identical to its explainers.
8. Abstraction in Regularity Explanation
How does idealization work in explanation? What to make of scientific explanations that seem to turn on mathematical, not physical, facts? (You'll find the idealization part of the story here.)
9. Approaches to Probabilistic Explanation
The various roles of probabilistic explanation. The various theories of probabilistic explanation. Related problems: Does the size of a probability cited in a probabilistic explanation matter? Can there be a satisfactory probabilistic explanation of a deterministically produced phenomenon? (The kairetic account will say yes to both questions.)
10. Kairetic Explanation of Frequencies
To explain certain deterministically-produced frequencies, a probabilistic model is preferable to any deterministic model. Both major elements of the kairetic account come into play in reaching this conclusion: the account of difference-making (as opposed to merely causally influencing) developed in chapter three, and the relation of entanglement developed in chapter seven. Then, a new kind of probabilistic entity and corresponding explanation: quasiprobability, a kind of
approximate probability. The chapter ends with more metaphysics: an account of statistical laws and of single case probability (and quasiprobability).
11. Kairetic Explanation of Single Outcomes
The same for single outcomes: they are in some cases better explained probabilistically than deterministically, even where a deterministic explanation is available. These ideas are applied to further topics. First, the significance of robustness in the explanatory enterprise. Second, an account of why not every probability-raising causal influence is a causal explainer. Third, an account of why not every difference-making fact (in any sense of difference-making: kairetic, counterfactual, manipulationist…) is a causal explainer.
12. Further Applications of the Kairetic Account
To explanation in biology, psychology, economics. The probabilistic framework of explanation invoking natural selection. Why wide content is explanatory. Why the assumption of perfect rationality is (sometimes) explanatory. And a few words on the autonomy of explanation in the special sciences.
13. Beauty in Explanation
Why we say that certain especially satisfying explanations are
beautiful. Finally, an interesting question!