Collective nouns are a fascinating topic for sociolinguistic study. This is because they respond to quite complex linguistic constraints and are of course variable by their nature, being singular in form but referring to a group of individuals or entities. From a social point of view, this latter attribute makes them a target for the tidy-minded, or those anxious to show that they can count up to two. Even in the populist UK, some people are capable of getting upset over a headline like ‘England suffer heavy defeat’, and the US grammar checker doesn't like it either; so what on earth are the standardising French likely to do?

The book under review is largely based on a speaker survey investigating the linguistic and social factors that influence variable use of verb agreement after some French collective nouns. The nouns selected for analysis were partie, foule, minorité and majorité. Variation after nouns of this type has been the object of codification since the time of Vaugelas (he thought, with surprising tolerance, that une infinitude de personnes should have plural agreement), and variation attracts the fury of prescriptivists to this day. After an introductory chapter (5–14) setting out the scope of the study, the second chapter (15–44) considers the linguistic constraints capable of influencing variation. Since verb forms vary, it goes without saying that French morphosyntax is involved; so is semantics, since lexical items inevitably convey meaning. Relevant semantic features include ‘(in)definite’, ‘animate’ and ‘abstract’. Syntactic factors, on the other hand, mostly concern the nature and position of the post-modifier, as in une foule de + plural noun. There is a semantic dimension to the latter sequence as well, of course: the presence of a post-modifier tends strongly to encourage plural agreement. But this is not the entire story, since reference to the wider textual context can also be relevant.

Chapter three (45–67) describes the methodology used to gather and analyse data. Tokens of individual words are much less frequent than variants in phonology, but the author circumvented this by asking her speaker sample to fill in a cloze or gap-fill test to supplement the more usual interviews. Speakers (from Normandy) were sampled in three age groups (ranging from 15 to 84), and by gender and educational level, the latter on the reasoning that more time spent in education might induce respondents to approximate standard norms more closely. Similarly, it was rightly thought worthwhile to test the usual sociolinguistic assumptions about the influence of age and gender in the still under-researched French context.

The linguistic results discussed in chapter four (69–95) defy adequate summary in a short review, which is unsurprising in view of the complexity of the situation (as outlined above). However, the very broad social picture that arises conforms to expectations: older respondents resist the trend, while women, as ever, are ahead of the game. Educational level showed an effect in the standard direction, at least in the cloze test, a finding discussed in more depth in chapter six.

Change over time is the topic of chapter five (97–124), which considers patterns that emerge from an array of written texts, the oldest of which go back to around 1600. The general trend is towards pluralisation, but singular verb agreement with la foule, for instance, has remained stable at about 95% since 1850. Diachronic findings are of course solid, and some are satisfyingly monochrome: for example, agreement after la plupart has changed very neatly over five centuries from singular to plural. On the other hand, unmodified une partie continues to show more or less evenly divided variation.

Chapter six (125–137), entitled ‘Explorations of education’, follows up the finding that respondents with a higher level of education supplied a greater number of singular verb forms in the cloze test. A historical survey of what school grammars say on the subject shows that the correlation between education and prescription is not straightforward, as these grammars show a good deal of tolerance and rationality. A qualitative study focuses on those respondents who gave categorical or near-categorical replies in the cloze test, whether singular or plural (the test contained sequences that invited answers in either direction). The results illustrate through glimpses into individual identity the seemingly anodyne point that educational level is an abstract category to which people respond for often imponderable reasons; ‘seemingly’, because such insights shake sociolinguistics to its foundations, the discipline having no way of analysing these elements of variation.

The book, based on the author's PhD thesis, is well written and clearly structured, and avoids the postgraduate traps of over-signposting and taking nothing whatsoever for granted. Anyone teaching variation in French will want to talk about the findings and reflections reported in this study. A remarkable amount of ground is covered in a small compass. This is a highly welcome addition to the Legenda list, and one must hope that further linguistics titles will be added to it before very long.