Robert Batterman's new book presents some original and very stimulating views about explanation, reduction, and emergence that will interest both philosophers of physics and philosophers of the higher-level sciences concerned with intertheoretic relations.
Batterman's chief insight is that understanding a phenomenon is as much a matter of exclusion as inclusion. To explain a phenomenon is to see what is relevant to the occurrence of that phenomenon. But by the same token, it is to see what is irrelevant to the occurrence. Thus, an important set of tools in the scientific repertory comprises the tools used to discard irrelevant details.
Consider, for example, a ball released on the lip of a mixing bowl (not Batterman's own example). The ball rolls into the bowl, and after a time, comes to a stop in the deepest part of the bowl. In order to understand this phenomenon, it is necessary to appreciate the positive roles of gravity, friction, and the concavity of the bowl in bringing the ball to rest at the bowl's bottom; but equally, it is necessary to understand that certain elements do not play a role in determining the ball's final position, notably, the point in the bowl at which the ball is released and the exact shape of the bowl.
Batterman rightly contends that philosophers of science have paid far too little attention to scientific techniques for discarding details, and therefore to the implications of the principled neglect of detail for the philosophy of explanation, reduction, and emergence.
The significance for explanation is already clear in the case of the ball and bowl: a good explanation must not only specify what made a difference to the explanandum, but what failed to make a difference. This is an extremely important point, which has been raised before in the literature on explanation (for example, in Alan Garfinkel's Forms of Explanation), but never, in my view, adequately explored. Batterman, however, quickly drops the thread of the argument concerning explanation in favor of points concerning reduction and emergence, which require the introduction of more complex examples.
The first of these is the celebrated explanation of the behavior of materials undergoing phase transitions near the “critical temperature.” What is striking about this behavior is that it is identical in certain respects across a wide variety of substances undergoing a wide variety of transformations. Such “universality”—the occurrence of similar behavior in very different systems—is a consequence of the fact that even quite salient properties of the different systems do not affect their behavior near the critical point. The universality is explained by finding a sophisticated technique for showing that these properties are explanatorily irrelevant (compare the irrelevance of the shape of the bowl in the example above).
Batterman claims that cases of this sort in physics cast considerable light on the fact that many theories in the higher-level sciences concern multiply realizable properties. Such properties are realized in quite different ways at the level of physics. Yet their behavior, when characterized at the level of description proper to some higher-level theory, is the same. Philosophers have thought this fact very puzzling; Batterman suggests that the similarities in behavior may be explained as a consequence of the fact that the differences in realization at the physical level are irrelevant to the higher-level behavior, in the same way that the differences between diverse systems undergoing phase transitions are irrelevant to their behavior near the critical temperature. The higher-level behavior would then be systematically explicable using lower-level theories (though Batterman denies that this would constitute a reduction). A case study from outside of physics would be very welcome here, but none is provided.
Another example from physics illustrates Batterman's contention that certain features of apparently obsolete theories may be essential for scientific understanding. In order to understand aspects of the geometry of rainbows, Batterman argues, it is necessary to disregard details of the wave behavior of light. Once these details are removed, what is left is more or less the old ray theory of light. A proper understanding of rainbow geometry, then, requires a theory whose basic ontology includes rays.
This is interesting enough in itself, accounting as it does for the persistence of old ontologies, often reconceived as idealizations, in science: as well as the ray theory of light, one might point to ideal gas theory, Newtonian mechanics, Mendelian genetics, and perhaps classical economics as examples (Batterman himself mentions only classical mechanics). But Batterman goes further, claiming that it is appropriate to say that rays are “emergent” in the behavior of rainbows, and that the proper theory of this behavior cannot be reduced to the wave theory, as it contains a ray ontology for which there is no basis in the wave theory.
These are exciting ideas. Unfortunately, their appeal is considerably obscured, I think, by Batterman's overriding concern with a particular method for understanding intertheoretic relations that he refers to as “asymptotic analysis.” This method has been, undeniably, scientifically very fruitful. But Batterman writes and reasons as though the method is both necessary and sufficient for thinking about intertheoretic relations, and that therefore, where the method breaks down or behaves in unexpected ways, the relations themselves break down or have unexpected properties.
For example, Batterman states without any reservations that the way that the phenomena described by classical physics are supposed to be explained by quantum physics is by considering the behavior of the latter theory as the Planck constant tends to zero (102). The limiting behavior turns out not to account for classical phenomena in certain respects, from which fact Batterman concludes that quantum mechanics is limited in its ability to predict and explain (at least in a “straightforward” way) the classical phenomenology. Most readers will take the fact rather as a reason to question the universal applicability of the asymptotic technique. They will, therefore, reject Batterman's conclusions about irreducibility and emergence, and perhaps—which would be a great pity—miss the significance of Batterman's arguments concerning the principled neglect of detail, which are not clearly distinguished in the text from arguments derived from “asymptotics.”
What can Batterman's book teach the non-philosopher of physics interested in reduction and emergence? These readers will, I think, find the book hard going: there are many detailed discussions of the physics, and no worked examples from outside of physics. Batterman does discuss the possible reduction of mental states to physics, but only in the most abstract terms.
The value of this book is rather in the conceptual possibilities that it limns. Batterman is surely right that in some domains (whether psychology, I am not so sure) multiply realizable properties are scientifically fruitful because differences in the realizers are explanatorily irrelevant. He is right, too, I think, in holding that the autonomy of higher-level sciences resides in the utility of higher-level vocabulary for systematically ignoring irrelevant low-level details.
On the negative side, the book contains little in the way of sustained philosophical theorizing. We are not given a much-needed general characterization of “asymptotic analysis.” We are not given an account of when a detail is or is not irrelevant to understanding a phenomenon. Nor, for all the comments on reduction, are we given a general criterion for deciding when one theory can be explained by another, or when a theory's ontology is emergent.
Still, there is plenty here for a philosopher of science in search of fresh ideas to feed on. I very much hope that Batterman's book will succeed in putting these ideas near the top of the philosophical menu.
