Composite Predicates in Late Modern English by Ljubica Leone aims at providing an overview of the diachronic evolution of composite predicates (CPs) in the English language between 1750 and 1850. The diachronic approach to CPs and the focus on the Late Modern English (LModE) period are highlighted in the introduction as the main strengths of this study, together with the use of the self-compiled Late Modern English-Old Bailey Corpus (LModE-OBC) as core data source. This corpus contains records of authentic speech events produced at the Old Bailey, the Central Criminal Court in London, which predate the invention of audio-recording technology.

Leone’s goals are ambitious but well justified. The author draws comparisons and signals points of divergence between her contribution and earlier research, effectively establishing the niche of her research. By targeting CPs, her approach intends to contribute to ongoing debates about the interface between morphosyntax, lexicology and semantics from the perspective of historical linguistics in a period in which CPs have received scarce scholarly attention. In a mostly carefully written piece of work, though also with numerous typographical errors and instances of unexpected editorial oversight, Leone’s study combines corpus analysis and theory to fill a general gap in the literature, thus complementing previous research by the author herself on the development of other multi-word structures like phrasal verbs, prepositional verbs and phrasal prepositional verbs (see Leone Reference Leone2023).

Composite Predicates in Late Modern English consists of 84 pages, including references, an appendix and an index. Rather short in comparison to other publications in the same series, the volume is divided into six chapters. Besides the introductory chapter (chapter 1), the literature review (chapter 2) and the conclusion (chapter 6), the three central chapters 3–5 address CPs from different perspectives based on corpus evidence. In what follows, this review offers an insight into the book’s central arguments, methodology and contribution to the broader field of historical linguistics, while at the same time it raises critical questions regarding certain aspects which could have led to substantial improvement.

Chapter 1 (pp. 1–13) introduces the concept of composite predicate, together with the aims and methodology employed (data source, selection criteria and statistical methods) for the analysis of these structures during the target period. Despite being a central chapter in that it sets the working basis for the empirical analysis that forms the book, it contains important inconsistencies and methodological flaws regarding the object of study. For instance, the author acknowledges (p. 2) that the selected label (‘composite predicate’) constitutes only one of the many terms that can be found in the literature, alongside ‘complex verb’, ‘verbal phrases’ or ‘verbo-nominal combinations’ (see further Wang Reference Wang2019: 20–1), but does not justify her selection. The label ‘verbo-nominal combinations’ is in fact a very extended term, the one used in general reference grammars and other diachronic studies, as it has the advantage of clearly describing the category, setting it off from other multi-word structures (see Claridge Reference Claridge2000: 70). The reader would have greatly benefited from a cohesive explanation and justification of the terminology, especially since the selected term leads to the title of the book.

A second issue which could do with further clarification in the introduction is the scope of the study. Leone remarks that phrasal verbs (e.g. look up), prepositional verbs (e.g. look into) and phrasal-prepositional verbs (e.g. look up to) were the object in her previous research (Leone Reference Leone2023) and, therefore, are not included in the present monograph. However, at no point does the author justify the exclusion of other multi-word verbal constructions, such as verb + prepositional phrase combinations (e.g. bear in mind, take into account), verb + verb combinations (make do [with], let go) and verb + adjective combinations (e.g. fall short, break even), which are usually grouped together in standard reference grammars of English and other diachronic studies on the topic (e.g. Claridge Reference Claridge2000: 39–41). There is a general scarcity of studies on group verbs in the LModE period, and hence their inclusion in this book would have resulted in a deeper, richer study.

Leone defines CPs as ‘phraseological verbs consisting of a verb (V) plus a deverbal noun (N) which behave as a single lexical unit, for example, take part of, pay attention to’ (p. 1), a definition allegedly based on standard grammars of English. However, there are no precise page references here or elsewhere in the book, which makes it difficult to locate CPs in these voluminous contributions, as in other cited works. To the best of my knowledge, the list of multi-word verbs in Quirk et al. (Reference Quirk, Greenbaum, Leech and Svartvik1985: 1150–68) does not include CPs, and Biber et al. (Reference Biber, Johansson, Leech, Conrad and Finegan1999: 428) discuss CPs under the label ‘verb + noun phrase combination’. Leone’s definition seems to be based on the introductory chapter of Brinton & Akimoto’s (Reference Brinton and Akimoto1999) collective volume, yet this work is not cited in this context, and when referenced later it is mistakenly referred to as ‘Akimoto (1999)’ (e.g. p. 58).

Leone’s working definition is also poorly formulated due to the absence of transparent examples, mostly owing to the inexact description of the nominal element of the compound. While answer and attention (in give answer and pay attention, respectively) can be considered deverbal nouns, part (in take part) cannot be classified in the same way: according to the Oxford English Dictionary, in fact, this noun is partly a borrowing from Latin and French (see OED s.v. part n¹). A look through the extensive list of CPs in the appendix (pp. 70–2) reveals that many of the nouns are not deverbal either (e.g. air, garden, time, to mention a few). Later in the book, Leone also states that nominal elements of CPs are always ‘abstract nouns that are etymologically related to the verbal stem (make a speech/speak) or linguistically related to it (have access/access)’ (pp. 31–2). However, the classification of nominal elements like job, alarm and breakfast as abstract is questionable, and doubt is likewise cast on the alleged (etymological or linguistic) relation to a verbal stem in some of the nouns listed in the appendix, such as memorandum. These inaccuracies ask for a reconsideration of the working definition of CPs from the outset, as they are vital and raise concern about the reliability of the data presented in ensuing chapters.

Leone’s methodological approach is unusually presented in the introductory chapter and again in chapter 2, rather than in a separate chapter of its own. In section 1.5, the reader is informed that the analysis is restricted to combinations formed with the light verbs do, give, have, make and take (p. 10), adopting the selection criteria defined in previous works, such as various contributions in the collective volume edited by Brinton & Akimoto (Reference Brinton and Akimoto1999) and also along the lines of Wang (Reference Wang2019). Later in chapter 2, a chapter devoted to the literature review, Leone explains the narrowed scope of her study in a paragraph which could benefit from more polished writing. She comments that Claridge’s (Reference Claridge2000) classification is broader than hers and any other, ‘as the inclusion of a verb such as take into consideration and pay attention to reveals the intention to extend analysis [sic] also to base verbs that are not immediately followed by the nominal part’ (pp. 20–1). Next, the author attempts to justify that her approach is different in that it only considers ‘combinations formed by the light verbs do, give, have, make, and take complemented by a noun’ (p. 20), which she further describes as ‘verb-nominal combinations in a narrow sense’, that is: ‘light verbs immediately followed by a noun working as the semantic focus of the combination’ (p. 20). Leone also mentions that ‘the inclusion of CPs formed by a base verb immediately followed by a preposition is, in the present work seen [sic] as a subtype of CPs, which should be treated separately given its diverse morphological structure’ (p. 20). While the decision to include a small set of verbs to analyze CPs in the narrow sense may be justified in shorter contributions, it feels rather unsatisfactory in a monograph entirely devoted to the category, crucially because it excludes verbs like lose, pay, put and run which have been examined in other major studies on CPs (see Claridge Reference Claridge2000: 70).

Another important absence in the description of the methodology is the precise input introduced in the concordancer for the extraction of examples. In section 1.5.2, Leone explains that WordSmith Tools were used to retrieve corpus data and visualize the examples in concordance lines with a context span of five words to the right and five to the left. However, no explicit mention is made, for example, of the use of the wildcard (*), which in WordSmith represents any string. That Leone considered intervening material between the verb and the noun in CPs is evident in the following chapters but, for the sake of replicability, the reader would have benefited from a more transparent description of the exact input(s) employed.

Similarly, the choice of text type calls for a more nuanced account. Trial proceedings are indeed an enlightening resource to explore authentic speech from a diachronic perspective, particularly when the phenomenon under study is clearly an aspect typical of the spoken language, as is the case with CPs (Wang Reference Wang2019). However, this feature is not fully exploited in Leone’s study given the absence of empirical comparisons with more formal, written texts from the same period which are readily available in numerous diachronic corpora. This restrictive choice of text turns out to be problematic for the interpretation of the corpus findings. For instance, on accounting for the high rates of CPs in the LModE-OBC, she concludes that ‘CPs exhibit a linguistic behavior that is similar to that of other multi-word verbs also for what concerns the medium and colloquiality’ (p. 26), positing two hypotheses: (i) that CPs are close to other multi-word verbs, such as phrasal verbs, which are typically associated with spoken data; and (ii) that CPs are used in colloquial language (p. 26). However, these hypotheses are not corroborated in the current volume against data other than trial proceedings. Leone does refer to other studies which include a wider range of text types (e.g. Claridge Reference Claridge2000; Wang Reference Wang2019), but these are concerned with Early Modern English, a period in which the frequency of CPs (and multi-word expressions in general) is much lower than in the LModE century examined by Leone.

The main corpus findings are presented in chapters 3 to 5, on the basis of a vast sample of more than 3,000 hits, which is sufficiently robust to enable an extensive analysis of the main characteristics of these constructions during the period under scrutinity and to obtain solid quantitative and qualitative results. Chapter 3 (pp. 25–40) focuses on the frequencies and productivity of CPs and the verbal bases that form them. Among the main findings, Leone notes that, like other multi-word structures (Leone Reference Leone2023), CPs present a higher normalized frequency than in earlier periods of the language (cf. Claridge Reference Claridge2000). Leone attributes this to the speech-like character of the texts in the sample (trial proceedings), as well as the natural mechanisms of language change which have driven the English language towards increased analyticity over time (pp. 25–6). An increasing trend is also attested in the data, especially marked from the 1810s onwards. To test the correlation between time and frequency, Leone applies a logistic regression model, which provides a solid empirical background. Although CPs display variation in this respect, the results of the test are statistically non-significant, which the author attributes to the fact that the LModE period is characterized by stability and for being linguistically close to Present-Day English (PDE) (p. 29). After the scrutiny of the individual verbs and nouns involved in CPs, this chapter concludes with an analysis of their productivity on the basis of their type–token ratio (TTR), which also reveals signs of stability yet with ongoing innovation.

Chapter 4 (pp. 41–54) delves deeper into the morphosyntactic and semantic features of CPs which situate them between stability and change in the LModE period. Leone observes the occurrence of two major patterns which reflect a continuation between earlier periods and PDE: verb + noun (e.g. make a mistake) and verb + noun + preposition (e.g. take advantage of). Whereas PDE CPs are prototypically characterized by their occurrence with an article, the LModE data show a marked tendency towards an article-less pattern, which suggests that CPs are not yet a stable category during this period (pp. 44–5). Conversely, the analysis of internal modification within the phraseological units reveals signs of stability within the constructions, although Leone acknowledges that the recurrent tendency of two of the verbal bases (do and give) to occur with adjectives may have biased the results obtained (p. 47). Even so, the possibility of modification is important in that it affects the degree of internal cohesion and thus can be revealing regarding the extent of lexicalization within the compound. The data also show that the nominal elements of CPs in LModE tend to appear in the singular form. This, together with the very low frequency of passive CPs, helps Leone situate them closer to earlier stages rather than to PDE trends. This chapter concludes with a very short note on the semantics of CPs, whereby the author puts forward the existence of a cline of compositionality which also characterizes other multi-word verbs.

In chapter 5 (pp. 55–66) Leone examines processes of change that have affected CPs. The chapter opens with an analysis of the morphosyntactic and semantic features presented in chapter 4, this time through the lens of two major processes of language change: grammaticalization and lexicalization. Corpus evidence leads the author to conclude that ‘grammaticalization and lexicalization affected some instances of CPs, which gradually increased their internal cohesion after the deletion of the determiner’ (p. 56), moving from the pattern V + D + N to a determiner-less pattern. To support this finding, but without an explicit justification, Leone narrows down her analysis to four specific CPs (have use, have information, take notice, take place), from which she concludes that CPs can be contextualized in the fourth stage of the four-stage process of lexicalization/idiomatization described by Brinton & Akimoto (Reference Brinton and Akimoto1999: 16). This is laid out in their work as follows:

According to Leone, if the results from the LModE-OBC are contextualized within this four-stage process, CPs can be situated at the fourth stage, which is characterized by the reanalysis of the constituent structure, ‘moving from [V] [D] [N] to [V] [N], and the consequent lexicalization of the constituents into a single lexeme’ (p. 58). It must be noted, however, that the reanalysis of the constituent structure corresponds to the third (not the fourth) stage in Brinton & Akimoto’s (Reference Brinton and Akimoto1999) description. Leone also argues that the specific constituent analysis can be taken as evidence that (i) CPs are grammaticalized structures concerning their increased bondedness and the loss of syntactic variability, and (ii) CPs are lexicalized in the sense that they display univerbation of the components. Nonetheless, though revealing, these claims require more caution since they are based on four randomly selected compounds which manifestly bias Leone’s conclusions.

Another important finding is the identification of new coinages ‘first established in the period 1750–1850’ (p. 61). The LModE-OBC provides a number of combinations (e.g. make discussion, take warning, give reference, etc.) which are not identified in previous studies and do not occur in other corpora examined (ARCHER), thus proving that ‘they are new coinages in LModE’ (p. 61). As above, this statement reminds us of what Matti Rissanen (Reference Rissanen1989: 17) termed ‘God’s true fallacy’, that is, ‘the erroneous impression that [an authoritative corpus] gives an accurate reflection of the entire reality of the language it is intended to represent’. Leone’s corpus portrays a very small portion of the linguistic reality of the LModE period, which is sufficient to make a survey of CPs, but very limited to make such general statements. In fact, a quick search in the OED quotation database provides examples like (1)–(3) below, which antedate those in Leone’s corpus. It thus follows that Leone’s conclusion that the creation of new forms is favored by analogy in pairs like take warning/have warning and take trouble/have trouble (p. 61) needs to be revisited.

Notwithstanding its limitations, Leone’s monograph stands as an important contribution to our understanding of the diachronic development of CPs in the LModE period. By using data which are less readily available in other corpora, Composite Predicates in Late Modern English presents empirical evidence of ongoing language change in LModE spoken data, which in turn enriches our knowledge of language change in this period and offers new directions for future research.