Brinton reports on a corpus investigation of a selection of comment clauses that are based on common verbs of perception and communication—for example, see, mean, say—and that have diverse syntactic forms—for example, (I) say, (as) you say, look, I gather, if you will, what else. She draws on corpora of present-day English from various parts of the world—Britain, the US, Canada, Australia, New Zealand—and on a range of historical corpora. She notes the methodological issues and difficulties involved in recovering pragmatic meaning, from corpora generally but more so from corpora of English in the earlier periods.

The first three chapters have a theoretical focus. Ch. 1 discusses pragmatic markers in a broad way, then reviews the literature on comment clauses as parentheticals and pragmatic markers. Ch. 2 reviews the semantic and syntactic development of pragmatic markers, both pragmatic markers with adverbial sources and clausal pragmatic markers. Ch. 3 discusses approaches to accounting for the development of pragmatic markers: grammaticalization, pragmaticalization, lexicalization, and idiomaticization. Brinton explains that she made a decision to choose grammaticalization as the theoretical model for her study because comment clauses “undergo decategorialization, lose referential meaning, and acquire functional (pragmatic) roles” (72). The following seven chapters present case studies of a selection of comment clauses: (I) say, I daresay, (as) you say, that is (to say) (Ch. 4); I mean (Ch. 5); you see, as/so you see, see (Ch. 6); if you will, as it were (Ch. 7); look (Ch. 8); what's more, what else (Ch. 9); I gather, I find (Ch. 10). Each case study begins with an analysis of major uses of the comment clause in present-day English, which includes range of meanings and frequencies of occurrence. A presentation of the historical development of the comment clause follows, supported by evidence from relevant corpora. Finally, there is a theoretical discussion of development with respect to grammaticalization and the data for that comment clause.

The concluding chapter in the book (Ch. 11) pulls together the results of the case studies. First, diachronically, the sources of comment clauses are more varied than previously assumed. Second, the matrix clause hypothesis—that parenthetical comment clauses originate in matrix clauses which have become syntactically indeterminate—is called into question. The diachronic syntactic development of comment clauses is much more complex and less clear than the matrix clause hypothesis leads us to believe. The main issue is that the relevant matrix clause constructions are rather rare in the earlier periods of the English language. Third, the case studies provide better evidence for the syntactic origins of other types of comment clauses: imperatives and adverbial/relative clauses. Brinton concludes by suggesting how Construction Grammar may provide a way forward for further study of comment clauses and thus may lead to a viable alternative theoretical model to grammaticalization to account for their historical development.