This is the final book of a great Latinist: J. N. Adams died shortly after its publication. In contrast to his other recent books with their broad, sweeping themes — Bilingualism (2003), Regional Diversification (2007), Social Variation (2013), Informal Latin (2017) — this one focuses on an apparently small topic, the question of when Romans omitted conjunctions meaning ‘and’, ‘but’ or ‘or’ between words or phrases. (A.'s definition of ‘asyndeton’ does not include omission of sentence connectors.) Part of the answer is obvious: conjunctions are very frequently omitted in lists of three or more items, but much less often between pairs of two elements. This latter type, asyndeton bimembre, is what is tricky to explain and accordingly becomes the focus of the book. Thus the most famous example of Latin asyndeton, ‘Veni vidi vici’, is technically outside the scope of this work — though the discussion nevertheless encompasses this and many other longer examples.

The work focuses primarily on literary texts from the Archaic and Classical periods. Its evidential basis is a huge corpus of asyndetic pairs collected by a systematic hand search of selected laws and prayers, most early Latin poetry, Virgil, Catullus, Horace, Caesar's Bellum Civile, the Annalists, Sallust, and substantial portions of Cicero, Tacitus and Livy; significant amounts of Greek, Umbrian and Vedic data are also included as comparanda. A detailed text-by-text analysis of this evidence takes up about two-thirds of the book and reveals that the use of asyndeton fluctuated considerably, not only between but also within the work of individuals. The (highly complex) patterns revealed include a decline over time in the use of asyndeton; although given the Indo-European origins of -que A. sees no reason to believe that asyndeton is actually older than coordination (as many have hypothesised), it is certainly more common in earlier than in later Latin literature. Genre also played a significant role, especially in the late Republic and early Empire, when asyndeton was used above all in legal texts and secondarily in certain varieties of oratorical, political and historiographical prose.

The fruits of this careful study are anticipated in the first third of the book, which builds on that analysis to offer a more synthetic account of other factors that are or could be related to asyndeton. These have nothing to do with conveying speed or agitation (as often stated in comments on asyndeton); in fact it is likely that asyndetic pairs were pronounced with a pause before the second element, hence more slowly than coordinated pairs. One factor that matters is whether the two elements together form a complete set (e.g. M. Antonius Q. Cassius, tribuni plebis when there are exactly two tribunes) or not (e.g. optimus maximus, to which other laudatory adjectives could be added): A. states that pairs of the latter type are much more likely to use asyndeton, though as the distinction is often subjective he does not do much with it. The other factors include grammatical ones, such as pairs with a negative prefix (e.g. inamabilis illepidus), pairs of related words (e.g. sumere consumere or pellerent pellerentur), pairs of imperatives (e.g. i arcesse) and masculine/feminine pairs (e.g. pueros virgines). These turn out to take asyndeton to varying degrees, ranging from the imperatives (regularly asyndetic) to the masculine/feminines (usually coordinated). Two semantic factors are also investigated: pairs of opposites (e.g. vita mors) and family members (e.g. liberos coniuges); the former are more likely to use asyndeton than the latter, but semantic factors were probably never wholly responsible for asyndeton.

More relevant, in A.'s view, are structural factors, particularly the context. Asyndeton bimembre is most common in ‘accumulations’; that is, when a larger list contains a pair of words, that pair is more likely to have internal asyndeton than when it occurs outside such a list. It follows that, in editing and commenting on Latin passages that may contain asyndeton, scholars should not simply consider how that pair of words is joined elsewhere, but also look at the context of the particular passage.

The section on structures also discusses the order in which paired words appear: in many (especially early) authors, the longer one usually appears last, but the tendency to place the semantically stronger word last is also important. There is also a lengthy discussion of ‘end-of-list coordination’, that is, the English tendency to put a conjunction between only the last two elements of a list (‘A, B and C’); it turns out that this phenomenon also occurs in Latin, though much less often than complete asyndeton (‘A, B, C’) or repeated coordination (‘A and B and C’). But A.'s findings here are not as different from the communis opinio as he suggests, and the choice of ‘Friends, Romans and countrymen’ (sic, p. 192) as an example of English usage was unfortunate.

The book is clearly written with remarkably little jargon and is easy to understand despite the high density of information per page. It is also well organised and well equipped with aids to finding specific passages, including a highly detailed fifteen-page table of contents and three indices. Quotations are provided with translations when the reader particularly needs to grasp their content. Typographical errors are infrequent. A.'s last gift to scholarship is worthy of his memory.