This is Stalnaker’s highly anticipated, book-length treatment of the concept of common ground, of what participants in a linguistic discourse agree to take for granted. He develops an account of the structure and dynamics of common ground that, he argues, allows both a simpler semantics and a better explanation for certain linguistic constructions and devices. A methodological strategy threads through the book: idealized formal models are useful, not just for describing complex phenomena, but in separating the data from the problem, and in identifying what a solution needs. I will summarize the book and then comment on this strategy.

In Chapter 1, he contrasts three notions of a linguistic activity’s context. One is the intuitive idea of the objective situation in which the activity takes place. The second is a semantic idea meant to capture how what a participant says can depend systematically on variable features of that situation. The third is the psychological idea of the information that each participant presumes is accepted by the others—the ‘common ground.’ In enormously influential early work, Stalnaker developed a formal model of this common ground, which he further develops and applies in this book.

In Chapter 2, he explains how the model represents, not just what each participant presumes the others accept about the facts they are discussing, but also the fact that they each presume this in their discussion. This iterative structure is essential to common ground. As conversation proceeds, the common ground evolves with different sorts of changes being possible. At the heart of Stalnaker’s project is the hypothesis—dubbed the ‘autonomy of pragmatics’ (1)—that understanding the structure and dynamics of common ground independently of language yields insight into otherwise puzzling linguistic constructions and devices.

In Chapters 3 and 4, he contrasts his notion of common ground with a semantic notion of sentence presupposition. Of special value is his discussion of the need to clearly distinguish descriptive versus explanatory uses of theoretical terms. He argues that the phenomena the semantic notion was introduced to explain are better described and more clearly understood in terms of common ground. Central to this argument is the claim that a notion of common ground is needed anyway for understanding rational linguistic activity, unlike appeals to special semantic or syntactic structures.

Indeed, we need the notion to understand all coordinated rational activity. Drivers on a highway, doctors in surgery, cooks in a kitchen, and even kids on a soccer pitch coordinate their activities by relying on common ground. The common ground always has the same iterated structure, but its dynamics are special in the case of discourse, where participants can question what is taken for granted or suppose something new, and where they can disagree about which questions remain open. Stalnaker models these dynamics using special ‘derived’ contexts, determined as a function of the information in the common ground (the ‘basic’ context), and shows (in Chapters 6 and 7) how various linguistic devices (including assertoric and other forces, conditionals, epistemic modals like ‘might’ and ‘may,’ and so-called ‘subjunctives’) are used to bring about these changes. He argues that this strategy allows for a simpler truth conditional semantics for these constructions.

Stalnaker’s model uses a possible worlds framework to represent informational content. This framework has well-known difficulties representing the information a person has about who, where and when she is. The iterated structure of common ground compounds these difficulties, since it must represent the self-locating information that each participant presumes the others have. Stalnaker develops a modified version of David Lewis’ centered-worlds framework, but, contra Lewis, insists that uncertainty or error about who, where or when one is in a world requires uncertainty or error about what world one is in (115). This is the topic of Chapter 5, which develops some of his recent work. ^{Footnote 1}

In Chapter 8, he discusses recent work on relativism and contextualism, arguing that his account of common ground can clarify what seems true in these accounts. The book ends with an appendix dealing with several technical topics. The final section summarizing relations between basic and derived contexts is especially helpful.

Stalnaker’s strategy employs idealized formal models to clarify complex phenomena. While this strategy may not be to every philosopher’s taste, I found it enormously helpful, both for identifying disagreements and for clarifying solutions. But any strategy has potential risks.

One is that the model may make what is deviant seem perfectly normal. Stalnaker’s discussion of indexical information runs this risk. The cognitive state of someone uncertain or mistaken about who she is ought to seem disorderly, as fundamentally defective. While I think Stalnaker agrees, his modeling of Mark Richard’s notorious phone booth case makes the characters’ beliefs seem too orderly; he even claims that their beliefs are true (125). This makes it hard to recapture the confusion at the heart of their cognitive states.

A second risk is that the model may raise explanatory problems of its own. Stalnaker’s discussion of what he calls propositional ‘detachment’ illustrates this. On his model, common ground determines not just what participants in a discourse say, but also what there is for them to say, what propositions there are to be expressed. This means that propositions are, in the model, context-dependent. ^{Footnote 2} Stalnaker appeals to this in his accounts of epistemic modals (144-146), conditionals (149), disagreement (163); and in his response to relativism (207). Now, Stalnaker is clear that adopting an explanatory framework is not neatly separable from making a substantive claim (180-4). So the idea of context-dependent content may be viewed as an insight made possible by the model. Still, one may find the idea of a proposition’s truth conditions being “fragile” (163) even more puzzling than the phenomena it is meant to clarify.

Stalnaker’s book is an extremely impressive achievement. In this review, I give but a small sense of its remarkable depth and breadth. Written with his characteristic wit, elegance and admirable generosity for opposing views, it makes a forceful and compelling case for the autonomy of pragmatics. Because its topics are subtle and complex, it is not always an easy read. But it is guaranteed to repay the effort and is destined to become the focus of considerable research among philosophers and linguists alike.