My copy of Philosophy of Physics is either on my office desk or on the desk in my study at home - I can't remember which. As it happens, the book is in the office, which I can discover either by seeing it there or by seeing it is not at home. Nothing odd here but, according to a common interpretation of quantum physics, extremely odd things happen in an analogous situation involving atomic particles. In this case, the book (which here stands in for the value of the quantity called "spin") is not really in either place, or anywhere else, before there is an observation. It can, however, be brought into existence by being observed in the office. Stranger still, the book can be brought into existence in the office by the observation that it is not at home.

The claim that measurement in one place should simultaneously bring about existence in another almost defies comprehension, but this is what our understanding of nonlocality in quantum physics suggests. Results such as this are not just very surprising but also conceptually disorienting. They call into question our grasp of both the world and our thoughts about it, concepts such as space, time, causation and even existence itself.

Philosophers have always been students of this sort of conceptual disorientation. But many have refused to look to science for inspiration, believing that their investigation of the basic structure of the world must proceed a priori and should include no appeal to the results of empirical inquiry. Other philosophers, however, have rejected this attempt to isolate metaphysics from physics, and Lawrence Sklar shows why they are right. His splendid book introduces the implications of relativity theory, statistical mechanics and quantum physics for such metaphysical issues as the nature of space, the distinction between past and future, and the relationship between observation and existence. He also discusses some epistemological questions posed by the relationship between physical theories and empirical evidence, focusing on the problem of how the sweeping theoretical structures of physics could be rationally believed on the basis of what is necessarily a highly limited range of observation and experiment.

One of the book's many attractions is its balanced view of the relationship between physics and philosophy. If the idea that physics is irrelevant to philosophy is indefensible, so is the scientistic thought that physics will somehow replace metaphysics. As Sklar repeatedly illustrates, physical theories often presuppose substantive metaphysical positions. This shows that, however relevant physics may be to the metaphysical debate, we cannot simply read the right metaphysical answer off the physics, since the answer we get out may be only the answer that scientists put in. The book also shows that the task of interpreting modern physics is best seen as a collaborative project, calling on the skills of both the scientist and the philosopher.

ÁñÁ«ÊÓƵ

Writing a serious introduction to the philosophy of physics is no simple matter, since the reader has to be brought up to steam in two difficult subjects. Sklar's book is certainly serious; it is also genuinely introductory. He assumes no knowledge of either subject, and provides a sustained but informal exposition of modern physics that is a remarkable combination of scope, clarity, and accuracy. His philosophical presentation is similarly impressive, reflecting the important contributions to the subject he has made in other publications, without taking anything for granted. In short, this book is an expository tour de force. The discussion is too detailed for it to serve as an ideal popularisation; neither is it a typical textbook, since it will not take innocent readers to a level where they can pursue the subject themselves. To reach that level will require substantially more reading in philosophy and especially in physics. Sklar's book is, however, an ideal place to start. It should also be read by the many philosophers and physicists who labour under the misapprehension that the other subject is not really relevant to their work.

Peter Lipton is a lecturer in philosophy of science, University of Cambridge.

ÁñÁ«ÊÓƵ