- Compositional semantics: An introduction to the syntax/semantics interface by Pauline Jacobson
Pauline Jacobson’s Compositional semantics is a comprehensive introduction to formal semantics with a focus on those phenomena with the greatest potential to elucidate the interplay between syntax and semantics in natural language meaning. Let me emphasize the adjective comprehensive. This book contains nineteen substantive chapters covering everything from Semantics I staples like generalized quantifiers and scope ambiguities to negative polarity items, focus semantics, and a host of binding phenomena like weak crossover and donkey pronouns. Though broad in this way, the book is laudably modular and so could be used in a variety of courses focused at a graduate or upper-division undergraduate level. My overall recommendation is to strongly consider this text for your next semantics class, especially if you plan to focus on compositional issues.
All introductory textbooks in a field as large and diverse as formal semantics must make hard choices about empirical coverage, technical depth, and philosophical outlook. There is no one right way to teach formal semantics, and so my goal is to detail the choices made and to consider their necessary tradeoffs. This is a well-constructed textbook. Assumptions do not leak in, either from outside the text or from chapters that are meant to be isolated from surrounding chapters. The exercises vary appropriately in difficulty and are solvable from the assumptions in the text. They require the student to both comprehend the material and be able to generalize from it. The one exception to the book’s usability might be the typesetting of the formal material, which defaults to typewriter standards in many places (e.g. using the letter v for ⋁, <…> for 〈…〉, etc.). This impedes readability, especially where there are (semi)clashing notations. For example, brackets are used for denotations, as in [[…]], but also for delimiting, as in […⋀ [… ⋁…]], as well as for marking expressions as strings, as in […].
Empirically, Compositional semantics is top-notch. J has a vision, namely to teach formal semantics through an investigation of the syntax-semantics interface, and she does not back down from it. The result is an introductory textbook with more discussion of, for example, weak crossover than you might expect. I contend this is a good thing. The focus on the interface means that students are not merely given toy systems, but will leave a course based on this book with real understanding of ongoing debates about the semantics of, inter alia, pronouns, ellipsis, quantifier scope, and relative clauses. Students will understand not only the analytical possibilities, but also the data that can be used to distinguish them. The tradeoff is that there is less here on standard Semantics I topics such as pragmatics and intentionality, but this tradeoff is reasonable given the depth achieved in other areas.
J’s Compositional semantics also strikes a nice balance in terms of technical depth. The text is organized around constructing increasingly more complex fragments, which are summarized in a set of appendices. The fragment method is a tried and true means of introducing semantics and works well. Each time we extend the syntax we get a concomitant extension of the semantics, which gives the resulting system a well-built feel. In constructing these fragments, J makes the interesting choice to postpone the introduction of the lambda calculus until the middle of the book. Up to that point the objects that natural language expressions denote in the model are described in plain English as much as possible, though with the help of a type system. This is a nice approach, but when we finally come around to seeing higher-order logic (HOL), the book seems almost afraid to admit that HOL is a language in its own right, which can be interpreted in a model. This fear seems to flow from its commitment to directly interpreting English in previous chapters. The result is a presentation of HOL that is a bit idiosyncratic and may not help students in their future studies. However, this quirk is made up for...