This wide-ranging book investigates the hypothesis that constitutes its title: it claims that, rather than the sentence or the clause, it is the so-called colon, or ‘intonation unit’ (IU), that is the relevant cognitively and pragmatically elementary unit of Ancient Greek discourse. As such, the colon is considered to be at the basis of a number of word order rules, discourse segmentation and discourse coherence.

After a preface and a general introduction, providing a concise overview of earlier work by Fraenkel and others on properties of the colon and Chafe's work on intonation units, the book is divided into three parts, each with its own introduction and conclusion. The first, ‘Word Order Rules’, takes Wackernagel's Law (i.e. the tendency of a number of enclitics to appear in the second position of the clause) as its point of departure, and investigates the types of words that tend to appear in the first, second or ultimate position of the colon, and their different internal clustering orders respectively. Following Denniston's adage of enabling the reader ‘to bathe in examples’ (cf. p. xiv), S. offers an impressive amount of relevant material that is neatly presented in tables, and which is based on his research corpus consisting of the speeches of Lysias and four Platonic dialogues (Cratylus, Theaetetus, Sophista, Politicus).

The word order rules identified here form the basis for what is to come, as they are used to identify the properties of colon boundaries, which are considered by S. to be at the heart of discourse segmentation, an issue that is addressed in the second part (‘Discourse Segmentation’). Taking both the aforementioned word order rules and a practical colon typology based on what is known about IUs, S. formulates a number of segmentation criteria, which are subsequently put to the test through an analysis of passages from his corpus. It is in sections like these that S. is at his best by showing how a colon-based reading of Ancient Greek helps not only to determine its structural characteristics, but also to develop a more natural sense of reading instead of merely translating.

In the third and last part, ‘Discourse Coherence’, S. argues that the colon, as a pragmatic unit, essentially forms the basic building block of discourse. Discourse itself may then be seen as a pragmatically ordered hierarchical tree, and S. befittingly calls his own organographical presentation of this hierarchy a ‘Pragmatic-tree device’ or ‘P-tree’. After describing a number of pragmatic principles and a provisional typology of coherence relationships (the ‘P-nodes’ within the P-tree), S. extensively discusses one relatively long and six relatively short excerpts from his corpus and tries to uncover their internal coherence structure. The visualising qualities of the P-tree prove to be a useful tool for showing how groups of cola cohere both at a lower and higher level, while a clear identification and discussion of the P-nodes helps in assessing the hierarchical structure of the discourse. Although there is a fair amount of repetition in these sections, S. succeeds in demonstrating the benefits of his approach by offering an insightful rendering of often rather complex passages. The general conclusion summarises the argumentation while also briefly placing this study into a wider theoretical perspective, and offers a number of general suggestions. The work is concluded by a bibliography and extensive indexes.

This is quite a remarkable book, in both its scope, argumentation and presentation. It provides often very original in-depth analyses of a large number of phenomena related to the colon and discusses even more issues on the way. Furthermore, S.'s practical approach convincingly shows that investigating the pragmatics of the Ancient Greek language is worthwhile, not only for linguists but also for those who want to understand the texts in the widest sense of the word. However, it is doubtful whether the book will actually reach such an audience, because its originality is obscured by considerable theoretical and methodological idiosyncrasy and problems in structure and presentation. It is problematic that S. has not tried harder, apart from a few paragraphs, to integrate his book into the very fruitful pragmatically oriented research tradition within Ancient Greek and Latin linguistics of the last decades, and to place his approach on a more extensively formulated theoretical basis, especially since both word order and discourse structure and coherence are amongst the areas that have received considerable attention. S. puts considerable effort into stressing the points where his ideas differ from those of others, while missing the fact that many of his ideas are not ‘new’ in the sense that their essence has (often quite recently) been discussed and applied by others, be it in other wordings and within different research frameworks. In this respect it is even more striking that S. is completely aware of his isolated approach, as a sizeable part of his preface (pp. xi–xv, and again on the last page of the general conclusion) is presented as a kind of apology for the fact that the book is what it is.

A second and related problem concerns the scope of the book. Although the theoretical and practical interrelatedness of the three main areas is wholly convincing, S. should have paid more attention to delimiting its scope. There are many digressions and more than once information is provided for the simple reason that it would otherwise have remained unpublished (acknowledged by S. on p. 445), but when it comes to important theoretical issues, a more thorough discussion is often avoided as it is considered to be outside the scope of the study. It may be true that S. takes the predominantly practical instead of theoretical line of approach of this book and its long and complex genesis (cf. p. xiii) as a reason for this (be it again apologetically), but it is a shame, to say the least, as it seriously undermines the book's cogency.

The third and last problem is the structure of the book. As S. admits (p. xiv) the book as a whole is difficult to read, due to numerous digressions, repetitions, subsections and cross-references. Many issues are touched upon at one point in the argument, but worked out only much later, or not at all. This, again, unfortunately obscures the qualities of the book. The extensive indexes at the end compensate for some of these flaws, but not all of them.

This is an interesting, thought-provoking and often very original book, but it could have been so much better if S. had mitigated his idiosyncrasy and had really addressed the problematic issues raised by himself in the preface, instead of merely acknowledging them.