This volume is a tribute to Professor Hubert Cuyckens on the occasion of his retirement and contains a fine selection of state-of-the-art research papers which discuss the role of context, intent and variation in grammaticalisation. It includes ten chapters, preceded by a stimulating introduction by the editors which revisits the theory of grammaticalisation and discusses the seemingly contradictory views on this phenomenon, understood either in terms of reduction (Lehmann Reference Lehmann1985) or expansion (Himmelmann Reference Himmelmann, Bisang, Himmelmann and Wiemer2004). These perspectives, however, are seen as reflecting different stages of the process, and thus are less in conflict than they might at first seem (see Traugott & Trousdale Reference Traugott and Trousdale2013; Cuyckens Reference Cuyckens2018). Grammaticalisation is considered here as being multi-stage in nature, rather than as a ‘monolithic process that is identical across its substages’ (p. 3), and all the contributions to the volume aim to reflect this broad conceptualisation. Moreover, the editors adopt the metaphor of a river to describe the trajectory of grammaticalisation, which begins with a clear purpose and course, but progressively expands through the addition of more tributaries so that it no longer has the original clear direction. Eventually, the river ‘may even silt up and only remain as a kind of idiomatic relic in the linguistic landscape’ (p. 3).

Although there is no clear thematic division in the book, the first four contributions deal with semantic-pragmatic change, presenting case studies on degree and happenstance expression, and hence shedding light on the early stages of grammaticalisation in particular. The remaining chapters then address morphosyntactic change, mostly changes in complementation patterns, and deal by and large with long-term variation, characteristic of the later stages of grammaticalisation. Chapter 6, by Dirk Geeraerts, serves as an intersection between these two blocks of studies, emphasising as it does the multidimensional nature of semasiological structure through an analysis of the Dutch preposition over. The overall structure of the monograph is therefore coherent and rather balanced, even though some contributions on grammaticalisation and discourse (markers) specifically would have been welcome and would have added to a full understanding of this phenomenon.

In chapter 2, ‘Looking at grammaticalization from the perspective of short-time changes in real time: A comparative corpus-based study of literally’ (pp. 19–46), Karin Aijmer explores the changes in frequency and use of literally in Present-day English (PDE), using data from the British National Corpus (BNC) and the Spoken British National Corpus 2014. She shows that literally has become more frequent in recent decades, not only in its function as an intensifier, but also as a pragmatic marker, a usage which in fact outnumbers intensifying readings in the most up-to-date data. Moreover, it can be used as a response marker to indicate agreement with the interlocutor(s). Literally thus falls within the grammaticalisation cline of other intensifiers which originate in lexical material (see, among others, Breban & Davidse Reference Breban and Davidse2016 on very), gradually expanding its range of collocates. The increasing frequency of literally as both a degree and a pragmatic marker in Aijmer's data is associated with young, female adult speakers, following the Maxim of Extravagance, and is used to mark their message as interesting.

Chapter 3, ‘A quantum of salience: Reconsidering the role of extravagance in grammaticalization’ (pp. 47–77), by Jakob Neels, Stefan Hartmann and Tobias Ungerer, considers the role of extravagance and salience in the grammaticalisation of a number of emerging quantifier/degree-modifier constructions ([ein N(-chen) ‘small unit’ X]) in Present-day German, specifically ein Tick ‘a tick’, ein Quäntchen ‘a quantum’, eine Handvoll ‘a handful’, ein Tacken/Zacken ‘a spike’, ein Hauch ‘a breeze/whiff’, eine Spur ‘a trace’, ein Fünkchen ‘a spark’ and eine Idee ‘an idea’, in addition to the highly grammaticalised form ein bisschen ‘a bit’. Drawing on data from a webcorpus, they apply a collostructional and network analysis to identify the top hundred collexemes of each construction and the degree of functional overlap or difference of these. Their results reveal that the most innovative quantifier/degree-modifier constructions tend to co-occur with extravagant collexemes and are the preferred option in emotionally marked contexts. In addition, many of these novel forms show limited productivity and persistence (Hopper Reference Hopper, Traugott and Heine1991), in that they retain traces of their original lexical meaning. Moreover, all these items are strongly interconnected, with the exception of ein Fünckchen ‘a spark’ and ein Quäntchen ‘a quantum’, which are less strongly linked to the other forms in the network. The authors argue that these constructions emerged as alternatives in a functional domain monopolised by the prototypical ein wenig and ein bisschen and can be regarded as cases of grammaticalisation, although they also ‘exhibit shades of what is traditionally classified as lexicalization’ (p. 70). The development of such constructions shows that their emergence was not only triggered by a functional need, but also by the speaker's need to be extravagant.

Lauren Fonteyn and Enrique Manjavacas’ contribution (‘Maximizing the data-drivenness of grammaticalization research: A case study of to death’, pp. 79–107) is concerned with the grammaticalisation of to death from a resultative phrase to an intensifying expression. In this corpus-driven diachronic study, spanning the years 1550 to 1949, the authors conduct a diachronic cluster and sentiment analysis to track the changes in the collocational patterns of to death over time. The application of advanced quantitative methods to automatically filter and annotate examples with minimal intervention allows them to minimise subjective decisions and opens up new possibilities in the quantitative treatment and analysis of data. The results of the statistical analyses indicate that although the verbal collocates of to death considerably increase over the course of the eighteenth century, this prepositional phrase only expands to new semantic domains centuries later, which also coincides with a shift of polarity towards positive collocates.

In chapter 5, ‘From chance to epistemic possibility: On the grammaticalization of happenstance expressions in English’ (pp. 109–34), María José López-Couso and Belén Méndez-Naya trace the evolution of a number of happenstance expressions, specifically of per- and may-adverbs, using data mainly from the Historical Thesaurus of the Oxford English Dictionary (HTOED), the Oxford English Dictionary (OED) and the Middle English Dictionary (MED). The inventory of epistemic happenstance expressions in earlier periods of English was much wider and included not only the two major PDE adverbs perhaps and maybe, but also other formations with a variety of nouns and verbs, yielding forms such as peradventure, by case, may chance and may fortune. The lexical source for these expressions in most cases features a noun meaning ‘chance’ or an adverb originating in a clause with a verb meaning ‘happen (by chance)’. The only exceptions are the formations perhappen and mayhaps, with a less straightforward origin. The development of these per- and may-adverbs constitutes a clear case of grammaticalisation, since it involves processes such as semantic bleaching, decategorialisation, specialisation (Hopper Reference Hopper, Traugott and Heine1991), fusion and (inter)subjectification (Traugott Reference Traugott, Davidse, Vandelanotte and Cuyckens2010). In addition, the evolution of these forms fits two cross-linguistically attested pathways for epistemic adverbs: prepositional phrase > adverb, on the one hand, and matrix clause > adverb, on the other.

Dirk Geeraerts’ chapter, ‘The structured nature of prepositional meaning’ (pp. 135–57), deals with the polysemous structure of the Dutch preposition over ‘over’; however, rather than discussing its uses within the radial network model, he emphasises the multidimensional nature of semasiological structure. The semantic structure of over is said to involve three dimensions: (1) a spatially relational dimension, (2) a ‘motional’ dimension (including cases of real, fictive and zero motion), and (3) the contact between a trajector and a landmark. A number of additional uses may also be derived from a subset of readings related to the three image schemas distinguished by Geeraerts: (1) readings within the spatial domain, (2) metaphorical uses of over, and (3) over as a synonym of ‘about, on’, introducing an object of a cognitive or psychological activity, a reading which is very frequent in Dutch. Furthermore, the semantics of over is ‘multiperspectival’, since language users are able to consider its structure from different angles, which in turn allows for more flexible uses of the word (p. 155). Thus, the preposition over represents a clear case of secondary grammaticalisation and Geeraerts provides evidence of this form travelling further along the grammaticalisation cline, with additional uses as prepositional complement markers.

In chapter 7, ‘Competition in antagonistic verb complementation: A diachronic, corpus-based study of fight, oppose and protest in (Late) Modern British and American English’ (pp. 159–87), Marianne Hundt explores the complementation patterns of the antagonistic verbs fight, oppose and protest in (Late) Modern British and American English (henceforth BrE and AmE), on the basis of evidence from several diachronic corpora. The data examined were annotated for a number of predictor variables, specifically time frame (decade), register, verb form, definiteness of the Noun Phrase (NP) and the distance between the verb/preposition and the complement, measured in number of words. Overall, it is shown that there have been no major changes in the complementation patterns of these verbs over time in either variety: in BrE fight and oppose show a clear preference for bare NP complements, while protest is overwhelmingly found with Prepositional Phrase (PP) complements. Similarly, in AmE bare NP complements are also the preferred complements of fight and oppose, but also of protest, which shows a drastic decline in the proportion of PP arguments over time. A closer inspection of the verb protest in AmE reveals that the variables decade, genre and distance are significant predictors, whereas definiteness and verb form do not have a significant effect. Thus, protest occurs mostly with PP complements before the 1950s (largely with against), but this trend is reversed in the second half of the twentieth century, when bare complementation prevails. In addition, this prevalence of bare NP patterns is most notable in informal text types. Moreover, the greater the distance between the verb and its complement, the more likely it is for protest to select PP complements. Hundt also draws on BrE data from the News On the Web (NOW) corpus to trace ongoing changes from the 2010s onwards. The findings suggest that BrE follows suit, with a clear increase of bare NP complements, yet this variety shows ‘stylistic stratification’ (p. 181), since the change in this case is clearly led by tabloid journalism. As in AmE, the range of available prepositions with protest is drastically reduced, against being the default preposition, which shows ‘paradigmatic attrition’.

In ‘Promoting and inhibiting forces at work: A corpus-based analysis of negative contraction in the recent history of English’ (pp. 189–214), María José López-Couso and Javier Pérez-Guerra explore the role played by different intra- and extralinguistic factors in the selection of contracted and uncontracted negatives in (Late) Modern English. Drawing on data from ARCHER (A Representative Corpus of Historical English Registers), they consider 12 individual variables and apply a regression and Random Forests statistical model to identify which of these has a significant effect on the choice of contracted and uncontracted negatives. Their findings suggest that the major determinants of variation are extralinguistic, namely the variables ‘period’ and ‘register’. Contracted forms increase significantly over time, which supports the growing acceptance of negative cliticisation from Late Modern English (LModE) onwards. Regarding register, negative contracted forms tend to appear in texts with high degrees of speechlikeness. Some intralinguistic factors, however, also show high statistical significance, as is the case with person number and verbal tense, where there is a clear preference for first- and second-person singular subjects and for present-tense verbs, respectively. Other contexts in which contracted forms are likely to occur involve pronominal subjects, pronominal elements occurring after not/n’t, main clauses and sentence-initial positions. The variable ‘operator’, by contrast, contributes far less significantly to the variation between contracted and uncontracted negatives, although contractions are more frequent with auxiliaries than with main verbs.

The following contribution, by Günter Rohdenburg and Julia Schlüter, is entitled ‘Making interpretation redundant: The grammaticalization of that-clauses after verbs of deception’ (pp. 215–52). In line with Hundt's study, the authors here focus on complementation patterns diachronically in both BrE and AmE. In particular they consider the verbs of deceiving delude, deceive and fool, followed either by a directly linked content clause functioning as argument of the main clause verb (e.g. He deceived the public ∅ that he would step down), or by a PP which contains an abstract noun or a verb followed by a that-clause (e.g. He deceived the public into the belief/into thinking that he would step down). These abstract nouns or verbs in fact indicate cognitive processes or states of the experiencers, and hence the label ‘interpretator’ is used to refer to these interpretive phrases. A number of variables, both extralinguistic (‘variety’ and ‘genre’) and intralinguistic (‘interpretator’, ‘complementiser’, ‘transitivity’ and ‘negation’) are considered and tested for significance with logistic regression models, this from analyses which move from long-term (covering the entire data range), through mid-term (from 1810 onwards) to short-term (1990–2019). The results indicate that verbal interpretator types have replaced nominal ones over time in both varieties, although informal BrE leads this change. In addition, the use of the indefinite article and premodifiers with these nouns has reduced drastically, as also has the variety of nouns and their associated prepositions. Similarly, the range of verbal interpretators allowed in these constructions has narrowed. Moreover, BrE seems to lead the change in the loss of interpretators, with the first uninterpreted uses of deceive and delude dating to the eighteenth century, and the nineteenth century in the case of fool. In AmE the loss of the interpretator is also first attested in the nineteenth century. Interpretator omission is facilitated by factors such as reflexive uses and negation of the superordinate verb. The data also point to a growing tendency to drop the complementiser that, this initiating in BrE informal fiction texts. Contexts with reflexive uses also seem to favour omission of the that-complementiser. Moreover, constituent length plays a role, since greater subject complexity correlates with increasing use of the directly linked that-clause variant. Crucially, these changes in complementation patterns, with the gradual replacement of the nominal by the verbal interpretator and the ensuing loss of the interpretator, should be considered more widely as part of the trend from a (more) nominal to a (more) clausal domain.

The chapter by Paul Rickman and Juhani Rudanko, ‘Straddling a syntactic divide: Tracking sentential complements of promise and threaten in CLMET3.0’ (pp. 253–75), is also concerned with complementation patterns. Specifically, they look at changes in subject control and subject-to-subject raising constructions with promise and threaten from the early eighteenth to the twentieth century. These two constructions are related to various senses of the verbs: the lexical sense is associated with the control structure, while the grammatical one is linked to the raising structure. The authors report a significant increase in the number of raising constructions with both verbs over time, this being particularly marked in the case of threaten. The development of these raising structures yields clear examples of grammaticalisation, since promise and threaten here are semantically bleached and no longer express volition, but rather convey epistemic meaning (‘portend, presage’). The more advanced grammaticalisation status of threaten may be triggered by frequency and by pragmatic factors, in that it allows for the expression of negative stance and the establishing of personifications.

The final chapter in the volume is ‘The development of prose style in English horse manuals: Interpreting stylometric findings’ (pp. 277–303), by Bettelou Los and Thijs Lubbers, which also considers syntactic change. However, in this study the authors adopt a data-driven stylometric approach, using a balanced corpus of texts from horse manuals, a genre which has presumably remained largely unchanged over time; the texts all deal with the topic of the feeding of horses and range from Early Modern English (EModE) to PDE. In order to identify changes in syntactic patterns, the authors conduct a trigram analysis of Part-of-Speech (POS) tags. The analysis of these trigrams reveals some differences from earlier to later texts: the earlier ones show a tendency to place given information in end-focus position and to embrace a continuative style, with long sentences and a high frequency of conjunctions; by contrast, passives and increasingly complex NPs are more characteristic of later texts. A further difference is that the earlier texts exhibit a higher incidence of the subjunctive and more pre-subject phrasal adjuncts. Finally, the reported changes are understood in this study not as cases of syntactic change, but rather as changes in the conventions of style or genre.

To conclude, this volume is not only the perfect means of recognising and celebrating Hubert Cuyckens’ distinguished and prolific career, but also a very illuminating book for researchers on (historical) linguistics and grammaticalisation. As such, it will definitely provide further food for thought in the field and inspire additional high-quality, state-of-the-art publications on the phenomenon of grammaticalisation.