The Review of Metaphysics

Tony Cheng's book is intended not as an introduction but as a critical reflection on John McDowell's work. Oxford Kantianism is a movement that begins with Strawson's reading of Kant. Strawson endorsed the use of transcendental arguments but without subscribing to transcendental idealism. McDowell and Cheng both follow Strawson's lead, but Cheng attaches more importance to findings of empirical psychology than McDowell does.

In the first two chapters, Cheng defends McDowell's basic insight that the world and the minded-subject are constitutively interdependent. We cannot, with Descartes, take the mind to be real and wonder whether the world exists, but it is also an error to try to make sense of a mindless world. As humans we, like animals, have natural drives, but as we acquire language and thus rationality we gain a kind of freedom that other animals do not have. This new freedom is not supernatural—it is second nature, a fulfillment of a potential of our animal life.

In chapter 3, Cheng examines perception and knowledge, arguing for strong externalism—mental states are both internal and external, but cannot be analyzed into a separate internal and external part. There is therefore no inner realm about which we cannot be mistaken: We must live with the possibility of bad epistemic luck. In chapter 4, he examines how McDowell charts a middle course between Kripke and Davidson. Davidson is right to say—in opposition to Kripke—that language is not essentially a matter of following rules, but Davidson draws the false conclusion that there could be a language without customs. McDowell's insight—frequently overlooked by analytical philosophers—is that the potential for thought can be actualized only when the thinker is initiated into a tradition. In chapter 5, Cheng examines McDowell's position on personhood and agency. Animals can be said to act for reasons, as when a dog flees from danger, but they lack the human capacity to reflect on those reasons. I might decide that although there is danger, I will not flee, and so I can be said to have beliefs about danger. Herbert Dreyfus misunderstands what McDowell says about the capacity to step back and reflect. Dreyfus thinks of reflection as a barrier to action. But McDowell does not mean that action always follows from reflection. I may act, and then reflect on the reason for my action.

In chapter 6, Cheng presents what he describes as the core of the book, a discussion of McDowell's account of subjectivity. McDowell offers a transcendental argument that establishes that experience has no phenomenal content that is not intentional—otherwise rational cognition would be impossible. But McDowell is a disjunctivist, and Tim Crane argues that disjunctivism is incompatible with the idea that all experience has intentional content. As a disjunctivist, McDowell holds that veridical perception involves a relation to an external object and hallucinations do not—these are two different kinds of mental state. Cheng points out that since McDowell is an externalist about mental states, he can treat the [End Page 811] difference between the presence or absence of an external object as an intentional difference, so from McDowell's perspective, disjunctivism is compatible with the idea that there is no phenomenal content that is not intentional.

It is in the final chapter that Cheng diverges somewhat from McDowell. McDowell once argued that thought and perception are propositional. However, he revised that idea after considering the following example. Two people see a red bird, and just one of them is capable of framing the thought "That bird is a cardinal." The less knowledgeable observer lacks the capacity to express this proposition, but both have the same perception. So McDowell now argues that perception is not propositional, but he maintains that it is conceptual. Making use of recent experimental data, Cheng argues that McDowell would do better to argue that perception has...