Agreement with collective nouns is an area that invites studies based on modern-day large-scale corpora. The phenomenon itself is not very frequent, and therefore requires large amounts of corpus data. This is particularly important when studies involve comparisons across varieties, time periods and lexical alternatives. In recent years, a number of studies (e.g. Depraetere Reference Depraetere2003; Lakaw Reference Lakaw, Hoffmann, Sand and Arndt-Lappe2017) have been published on the topic, but in spite of this, Yolanda Fernández-Pena's monograph on agreement with complex collective subjects manages to produce a whole host of new insights that will have a bearing outside the area of agreement studies per se.
The study examines the influence from morphology, syntax, semantics and lexis (p. 50), and focuses on the types of structures exemplified in (1) and (2) (p. 1):
(1) [The second group of books] is those written by botanists.
(2) Under a nearby tree, [a group of children] were having their school lunch.
While previous studies (e.g. Lakaw Reference Lakaw, Hoffmann, Sand and Arndt-Lappe2017) target prototypical collectives such as family, army and committee, this volume focuses exclusively on collective nouns followed by oblique of-dependents. The complex collectives under scrutiny are defined as ‘binominal phrases of the type a bunch of ideas or a flock of sheep, in which the first noun is assumed … to be a collective noun (NCOLL) which (i) denotes a collection of entities of any nature, (ii) takes an of-dependent (of N) that contains an (often plural) oblique noun (N(PL)) and (iii) may imply some nuance of quantity’ (p. 13).
These types of collective noun subjects have largely been excluded in previous studies on agreement, and Fernández-Pena's study is therefore a timely contribution to agreement studies, filling a gap left by previous studies and opening up avenues for further research. The material is provided by the large-scale corpora found at www.english-corpora.org. Included in the study are referential constructions (e.g. a bunch of keys is placed under a chair), partitives (a bunch of the other guys come over) and pseudopartitives (a bunch of cases are raising those issues) (pp. 13–23). Partitives introduce two separate referents in the discourse, and the second NP is therefore generally in the definite form (the other guys). In pseudopartitives, on the other hand, the first NP functions as a quantifier, and the second NP is typically considered to be the head of the whole NP. The oblique nouns in the various constructions include regular and irregular plurals (of books/children), singular nouns (of the population) and the formally unmarked people, which in some contexts displays patterns that are quite different from the other oblique nouns.
The volume is divided into five chapters. A brief introduction to the topic is given in the first, and an extensive overview of the theoretical background and previous studies is provided in the second. Chapters 3 and 4 present the findings, the former being devoted to diachrony and the latter to synchrony. Finally, chapter 5 summarizes the results and provides an outlook for future studies.
In chapter 2, the author presents a range of theoretical approaches and previous findings. Special emphasis is given to Corbett's canonical model (Reference Corbett1979, Reference Corbett2003) – including his Agreement Hierarchy – the generative framework and cognitive approaches. Corbett's model and cognitive grammar are central to chapters 3 and 4, but generative grammar is somewhat lost on the way and hardly mentioned again. The same can be said of another relevant aspect mentioned in the background, that of prescriptive influence on language usage (pp. 28–9), which is hinted at but not really connected to in the analysis. These minor issues aside, the essential ideas of the different schools of thought are presented in an admirably lucid and concise manner in this chapter.
One feature that is related to cognitive processing, syntactic complexity, is devoted much space in both the background and the analysis. Based on the findings from previous studies, which have indicated that greater distance between the collective noun head and the agreeing verb increases the likelihood of plural agreement, the author presents different measurements of syntactic complexity. Complexity and the different features included in the study are given succinct overviews, which aids readers greatly when following the elaborate discussions of the findings in the next two chapters.
Chapter 3 is the first results chapter, devoted to a diachronic study of seven selected collective subjects in the Corpus of Historical American English (COHA): bunch, couple, group, host, majority, minority and number. Two of these, majority and minority, express relative quantification where the quantity denoted by the collective is assessed in relation to the noun in the of-dependent. The rest express absolute quantification, i.e. quantity in relation to a scale of measurement. The findings are largely connected to the various criteria for syntactic fixation posited in previous studies. The evidence presented is striking and partly puzzling.
For a number of, the data show that this was conventionalized as a quantifying construction already in the early 1800s, being almost exclusively associated with plural agreement (a number of scientists are active). The definite counterpart, the number of, has instead retained its lexical or referential meaning and is in Present-Day English used with singular verbs (the number of guests is limited to sixty). Further evidence of the idiomatization of the indefinite a number of, apart from the use of plural verb agreement, is provided by the decreasing frequency of premodifying adjectives: in the first half of the nineteenth century, half the instances involved adjectives (a great number of new banks), and this number drops to a quarter in the second half of the twentieth century. Perhaps unexpectedly, the variability among the adjectives increases over time, so that the proportion of the most frequent adjectives – such as large, great and certain – decreases, while the proportion of ‘other’ adjectives increases. A final, important finding relates to the overall frequency of the construction – it peaked in the twentieth century, and has since been decreasing rapidly in the early twenty-first, as shown by supplementive searches in the Corpus of Contemporary America English (COCA).
The next phrase under scrutiny, a group of, puzzlingly shows an increase in singular verb agreement (a group of soldiers bursts in), which forces the author to conclude that ‘verb number agreement is not a reliable predictor of syntactic fixation in this case’ (p. 66). Moreover, premodified instances (a considerable group of men) are increasing in frequency, which also contradicts any hypothesized idiomatization. These findings for a group of certainly suggest that further studies would be needed in the area.
Majority of is both similar and different to number of and group of. Plural agreement is used in a vast majority of cases, but this applies not only to indefinites (a majority of Senators were ready), but also to definites (the majority of men are imperfectly educated). The semantic analysis presented does not indicate any differences in meaning between the two alternatives either. Here the author reasonably suggests that the partitive nuance of the noun majority evokes quantification with both sets of determiners, which is why plural agreement is used. The most notable finding regarding adjective modification with this construction is that it is rare, and that this adjective-less fixation of the phrase must therefore have taken place before the nineteenth century. Since the majority of remains stable in frequency and a majority of is decreasing, it is concluded that there is no support for ongoing idiomatization of a majority of in the extensive corpus data.
The corpus data for the remaining constructions are more restricted and the findings less striking. The recent specialization of a bunch of with the adjective whole, accounting for more than half the premodifiers with that construction, and the rapid increase of the non-premodified, plural-verb only a couple of should still be mentioned.
The chapter concludes with a discussion of the findings and comparisons between the constructions. The two that express relative quantification, a majority of and a minority of, are found to be the least idiomatized. Among those expressing absolute quantification, a number of is the most idiomatized, and a group of the least, partly because of the increase in singular agreement with the latter. Absolute quantifiers can usually be substituted by prototypical quantifiers such as many and (a) few, while this is not possible for the relative quantifiers. Taken together, the findings regarding these quantifiers are argued to illustrate how speakers are constantly finding new and expressive ways to convey indefinite quantification (p. 99). This result is further evidence of the constant renewal observed across languages in previous studies.
Chapter 4 focuses on synchronic variation. The first part considers synchronic American and British data from the Corpus of Contemporary American English (COCA) and the British National Corpus (BNC), a sub-study comprising 25 variables and more than 5,000 tokens. This is an illustration of the impressive scope and thoroughness of the study. In the following, only a few of the most important variables and results can therefore be described. As observed in previous studies, British English uses more plural agreement than American English, but in the present case the difference is rather small. The verb types with a preference for either the singular or the plural are also similar to those identified in previous studies – verbs associated with animate referents prefer the plural (e.g. see, design and prepare), while the singular is used with verbs referring to the size (increase) or composition of the collective (consist of). The collective noun itself was associated with most of the variance. The nouns majority, minority, couple, host and bunch are linked to plural agreement, while, for instance, herd, flock, set and pack are more frequent with singular. The nouns with the strongest affinity for the plural share an inherent quantifying function, as suggested in some previous studies.
In the synchronic material, plural agreement was significantly more frequent with irregular oblique nouns ((a group of) women, teeth, phenomena) than with regular plurals ((a group of) bees, girls, computers). This is a somewhat puzzling finding since previous studies have found the opposite (Haskell & MacDonald Reference Haskell and MacDonald2003). This result is explained by making reference to irregular nouns being less prototypical and more salient than regulars, and therefore more likely to affect verb agreement (p. 135). However, this seems to raise the question as to why similar results were not obtained in previous studies.
The findings regarding the influence of syntactic distance, as counted in number of words between subject and the verb, and syntactic complexity, as measured in the number of postmodifiers, have a bearing on our understanding of language processing. NPs with up to three postmodifiers occur in the material (the number of people [per square kilometre] [in the United States] [in 1980]). Different oblique nouns in the of-phrases yielded different results regarding the distance in words. Unexpectedly, singular oblique nouns (e.g. object and person) form the category that increases its likelihood of plural agreement the most with higher complexity. Regular plural oblique nouns are instead less likely to pattern with increasing distance in words, a finding not conforming to Corbett's (Reference Corbett1979) Agreement Hierarchy or previous results for prototypical collectives (Levin Reference Levin2001). Moreover, increased complexity as reflected in a large number of postmodifiers leads to lower proportions of plural agreement with the unmarked plural people as the oblique noun. No effect was nevertheless noted for regular plural obliques. These notable findings are explained by Rohdenburg's (Reference Rohdenburg1996) Complexity Principle (p. 138), which holds that more explicit alternatives tend to be preferred in more complex environments. Syntactically more complex contexts involving several postmodifiers would then favour the use of the more explicit singular verb number. The unchanged likelihood of plural agreement with regular plural oblique nouns would be explained by the marked plurality of the nouns making the plurality more explicit in contrast to the unmarked people. Thus, the findings corroborate some previous findings suggesting that syntactic distance (i.e. syntactic boundaries) exerts stronger influence on agreement than distance in words (Levin Reference Levin2001: 102).
The influence of animacy on verb agreement is perhaps more straightforward. As could be surmised from Comrie's (Reference Comrie1989) Animacy Hierarchy, complex collective subjects with human animate subject referents are the most likely to take plural agreement, and those with inanimate referents significantly less so (e.g. the following set of commands creates a directory). Of the more than 1,500 types of oblique nouns in the material, only a fifth displayed variation in agreement, a finding that illustrates the strong preferences associated with different kinds of constructions in the material. Of the nouns preferring the plural, most are animate, while those associated with the singular are slightly more varied.
The chapter concludes with a sub-study on regional variation, comparing the American and British data from COCA and the BNC, and those from World Englishes from the Global Web-based English Corpus (GloWbE). The comparisons are here restricted to regular and irregular plural oblique nouns and the noun people. The differences in agreement patterns between COCA and the BNC are quite small considering the greater preference for plural verb agreement in British English attested in several previous studies. The comparisons in GloWbE are restricted to American, British, Australian, Canadian, Irish and New Zealand English. Other varieties (such as Bangladeshi English) are excluded due to the language potentially being affected by substrate languages (p. 171), but it could be argued that this is no a priori reason to exclude these. In recent years, the expanded research focus on Outer Circle Englishes has produced findings that are also highly relevant as comparisons with the Inner Circle varieties. The results from GloWbE indicate a preference for plural verb agreement in British and Irish English and a preference for the singular in American English. As found in some previous studies, Australian and New Zealand English occupy a middle position (in this case together with Canadian English).
In conclusion, Fernández-Pena's monograph on verb agreement with complex collective nouns is a welcome contribution to English linguistics. It presents new findings from a niche that has previously been overlooked. Some of the findings are hard to explain, but they widen the horizons of our knowledge. Some findings are more in line with what previous studies have found, such as the strong influence of the collective noun head on the choice of agreement. The volume condenses a large number of findings connected to an impressive number of variables into a neat format. The amounts of corpus data used allow for well-founded conclusions, and the different stages of the statistical treatment of the data are kept brief, yet clear. The extensive use of examples to illustrate the different kinds of variation makes the discussions easy to follow. The findings will certainly inspire future studies in related fields.
