This volume presents a collection of nine chapters with the overarching topic of lexis and, by extension, lexicogrammar. All contributions arise from presentations at the 36th ICAME conference, held at Trier University in 2015. In the Introduction to the volume, and against the backdrop of the conference theme, ‘Words, Words, Words – Corpora and Lexis’, the editors draw attention to two issues that ‘figure prominently in the linguistic study of lexis – one is paradigmatic, one is syntagmatic’ (p. 1). They also stress the role of corpus-linguistic work in highlighting the overlap and connections between the two; indeed, the two dimensions, as well as their interwovenness, are represented in this collection.

One of the stated aims of the International Computer Archive of Modern and Medieval English (ICAME) is to ‘to promote all branches of English corpus linguistics, in particular the compilation and analysis of corpora of English and applications of English corpus linguistics’ (ICAME constitution).Footnote ¹ The current volume fulfils this aim in that it provides new insights into all of these areas of corpus linguistics. In addition, as pointed out in the book's blurb, one of the chapters offers ‘a lexicographer's view on the data structures implicit in the Oxford English Dictionary (OED)’ and how these may be visualised in the digital age.

In the first chapter, entitled ‘Modelling lexical structures in the Oxford English Dictionary’, Edmund Weiner, Deputy Chief Editor of the OED, starts out by stating that the business of ICAME and that of the OED are rather different. However, there are overlaps, particularly in that both adopt computer-based methods to gain knowledge about the English language. Following a brief summary of early thoughts on the computerisation of the OED and predictions that were made at the time (mid 1980s), Weiner compares some of the features in the first CD-ROM version and OED online, and to some extent also the paper version. Although the OED in many ways has developed as predicted, there is still one dimension that has not (yet) been fully tackled, namely the ‘relational database that would store similar items together for examination and analysis’ (p. 9). Weiner goes on to present thoughts on how the OED could accommodate such a functionality that would explicitly visualise the systematic structural facts that are implicitly present in the dictionary, for example a visualisation of how entries are connected by etymology (p. 28). The various proposed networks containing structural information about the English lexicon immediately offer conceptual links that would be hard to discover otherwise. These convincing visualisations notwithstanding, Weiner stresses that the diagrams presented need not be the only way of representing the contents of the OED. They should be seen as a supplement to, rather than a replacement of, the current print-like linear presentation.

In chapter 2, ‘Investigating the circumstances of coinage’, Antoinette Renouf seeks to develop a methodology to investigate the circumstances of coinage, thereby gaining better knowledge about the nature and typology of coinage. Coinage is defined as ‘the act of coining or inventing a new word or phrase’, and in addition, it is deemed to be ‘a special case of neologising, distinct in that the act of creation itself is our [the author's] focus’ (p. 41). Coinage, when used as a count noun, is defined as the product of that act. Drawing on data from a newspaper corpus of 1.5 billion words spanning the 30-year period 1984–2014, Renouf employs a three-step method before analysing the coinages. The first step involves automatic extraction of potential coinages and identification of signalling devices in their immediate context. Next, further candidate signalling devices are searched for in the context of the potential word coinages listed in Bauer (Reference Bauer2001). Finally, the third kind of search process concerns past coinages and signalling devices associated with these. Following these three steps, the author develops a provisional typology of coinage in terms of which signalling device is used: no signalling, tentative signalling, performative signalling and past signalling. Renouf successfully demonstrates how the categories in the framework behave differently. Performative signalling (e.g. I hereby dub it), for example, is used where ‘the act of coinage takes place within the text itself’ (p. 67), while tentative signalling (e.g. if that's a word) occurs with coinages ‘which are not yet deemed to properly exist’ (ibid.).

In chapter 3, ‘Synonym selection as a strategy of stress clash avoidance’, Julia Schlüter and Gabriele Knappe set out to put one of Bolinger's (Reference Bolinger and Bolinger1965) observations to the test, where he claims that stress clash avoidance can determine lexical choices. In previous research, rhythm is presented as the main explanatory factor as to why glád dáy seems objectionable whereas glád occásion is acceptable. Taking a diachronic approach, the authors hypothesise that two synonymous adjectives with different rhythmic shapes do not occur equally often in all syntactic functions, notably prenominal position. In four case studies, each with its own challenges, support is found in favour of the hypothesis in that preferred choices of lexically specified adj + noun combinations tend to avoid stress clashes. As the authors state: ‘[t]he rhythmically less appropriate members of the [adjective] pairs (or triplet) are strikingly underrepresented before nouns’ (p. 93). The consistent results are also substantiated in both spoken and written present day English. Schlüter & Knappe acknowledge that there are some important limitations to their study, for example the fact that each of the individual lexemes studied has its own idiosyncratic (semantic) properties and phraseological potential. Indeed, one of the case studies in this thorough investigation offers a more detailed discussion of the adjective glad. Although this has some relevance to the overarching aim of the chapter, it could perhaps have been left for another occasion, as glad seems to deserve an in-depth study of its own. Despite the restrictions and challenges mentioned, the study clearly contributes new insights regarding synonym selection as a strategy of stress clash avoidance.

In chapter 4, ‘Intensification with very, really and so in selected varieties of English’, Karin Aijmer investigates intensification of adjectives in four varieties of English, representing both inner (UK, USA, New Zealand) and outer (Singapore) circle varieties. Focusing on colloquial language, the study draws on material from the private conversation components of the International Corpus of English (ICE) and the Santa Barbara Corpus of Spoken American English. Aijmer compares the use of the most frequent intensifiers – very, really, so – when immediately followed by an adjective. The study reveals some differences in the frequency of intensifier + adj. All three intensifiers have a higher overall frequency in ICE-SIN, while American English has the lowest overall frequency. In British and Singaporean English, very is the preferred intensifier, while New Zealand English prefers really and American so. Following a qualitative analysis of the type of adjectives used with intensification, Aijmer finds that good and nice typically co-occur with very and really, while funny is more prominent with so in all varieties. The adjectives are most commonly evaluative with all the intensifiers, but some differences are noted between the varieties in terms of semantic prosody. Through a quantitative and qualitative analysis, Aijmer's study clearly demonstrates how intensifiers and combinations of intensifier + adj have developed differently. ‘Although the varieties share the history of colonisation by British settlers and having British English as their major influence’ (p. 135), they progress through time to ‘a new dialect or indigenous variety’ (ibid.).

John Kirk investigates the pragmatics of well as a discourse marker in chapter 5 (‘The pragmatics of well as a discourse marker in broadcast discussions’). This study has two main aims: to test a model of the functions of well on new data (SPICE-Ireland) and to compare the distribution and functions of well with those in British data (ICE-GB). Central to the investigation is Aijmer's (Reference Aijmer2013) model of well’s meaning potential, depending on the situational context, to express coherence, involvement and politeness. Applying this model to a new data set, Kirk thoroughly examines how well the model works on the Irish material, i.e. to what extent does well have the functions outlined by Aijmer. Kirk concludes that Aijmer's model is substantiated by the Irish data, thus lending further evidence to Aijmer's claims about well as a coherence marker (coherence function), an interpersonal marker (involvement function) and a politeness marker (politeness function). Kirk's comparison of the distribution of the three functions in the Irish and British data shows that the SPICE-Ireland and ICE-GB data were homogeneous in the sense that the functions are similar in the two varieties as well as their order of preference: coherence > involvement > politeness. However, different preferences in some of the sub-groupings of each function were noted. A case in point is the British preference for direct agreement in the involvement function, compared to the Irish preference for partial or implied agreement. Kirk finally draws attention to the three notions of homogeneity, register comparability and regionality, all of which were tested and ‘shown to be shared between the British and Irish data sets’ (p. 167).

Maïté Dupont's chapter 6 (‘Between lexis and discourse: A cross-register study of connectors of contrast’) offers new insights into the placement of connectors in English. The generally held view that the natural position for adverbial connectors is initial position is questioned and claimed to be too simplistic and thus needs to be qualified. Based on corpora representing three different registers, Dupont investigates seven high-frequency adverbial connectors of contrast, including however, nevertheless and still. Within an SFL framework, she examines the placement patterns by looking at the role played by lexis (i.e. individual connector types), by register and a combination of the two. The study shows that individual connector types have different preferences regarding placement. Moreover, some of Dupont's findings do not lend support to the description given in general grammars regarding individual connectors; still is a case in point. As regards register differences, it is found that the Thematic 1 position is the most prominent one in two of the registers: academic and parliamentary debate. In the third register – editorials – there is more of a division of labour between the Thematic 1 and Rhematic 1 positions. Finally, lexis and register are further shown to have a combined effect on individual connector placement. While some patterns are stable across registers, others are not. Through the thorough research presented in this chapter, we are given new perspectives on connector placement in English, as the placement patterns seem to be more complex than previously thought.

In chapter 7, ‘Towards a model of co-collocation analysis: Theory, methodology, and preliminary results’, Moisés Almela and Pascual Cantos challenge the commonly held view in Corpus Linguistics that collocation is a bipartite structure containing a node and a collocate (cf. Sinclair Reference Sinclair1991). Almela & Cantos rather advocate collocation as a multipartite structure where they introduce the concept of ‘mediated’ collocation, or co-collocation. They argue that ‘co-collocations are not amenable to the traditional definition of collocation as a bipartite structure, because they are constituted by the interaction of more than two lexical domains’ (p. 215). The authors outline a six-step methodology for extracting co-collocates from huge corpora such as enTenTen2013, a 20-billion-word English corpus of texts from the Web. The authors illustrate the method in a study of the collocates and co-collocates of the lemma consequence, and argue that a node–collocate pair (e.g. mitigate consequences, long-term consequences and unintended consequences) can be influenced by co-occurrence patterns outside the pair itself. For example, when the node consequences occurs simultaneously with two of its collocates, for instance mitigate long-term consequences or mitigate unintended consequences, each of the collocations may or may not be established independently. While the presence of mitigate seems to strengthen ‘the capacity of consequence to predict unintended’ it does not do so for long-term, and ‘only the former relation constitutes a co-collocational pattern’ (p. 223). Thus, the authors argue that a new technique is needed that accounts not only for collocations consisting of a node and collocate alone, but also for collocations consisting of node, collocate and co-collocate. The interaction between collocates and co-collocates is said to provide a more comprehensive and precise picture when describing word meaning through collocational data.

In chapter 8, ‘The lexicogrammar of be interested: Description and pedagogy’, Costas Gabrielatos presents pedagogy-driven corpus-based research, focusing on the intersection between language description, Learner Corpus Research and the evaluation of pedagogical materials. The purpose is to develop ‘a body of corpus-based lexicogrammatical information for learners’ (p. 240). Motivated by the lack of comprehensive information in existing materials, Gabrielatos focuses on the sequence beinterested. The investigation draws on corpora of L1 (BNC) and L2 (ICLE and LINDSEI) speakers of English. Some interesting observations emerge from the study, not least in the comparison of the use of beinterested in these corpora with how the sequence is treated in pedagogical grammars and dictionaries. For example, the study reveals that the pedagogical materials consulted do not include all the lexicogrammatical patterns observed in the BNC. This observation lends support to the treatment of lexis and grammar as being intertwined, as evidenced by the strong attraction between the adjective interested and copular be. The L1 vs L2 comparison uncovers more differences than similarities. Results suggest that there is indeed a correlation between learner use and the treatment of beinterested in the pedagogical materials, grammars in particular. It is suggested that future materials provide more ‘comprehensive and nuanced information on the complementation and related lexicogrammatical patterns of beinterested (and, more generally, “copula + interested”)’ (p. 266). Frequency and mode information would also be a useful addition.

In the final chapter, ‘Tracking L2 writers’ phraseological development using collgrams: Evidence from a longitudinal EFL corpus’, Yves Bestgen and Sylviane Granger use the Longitudinal Database of Learner English (LONGDALE) to investigate how the use of collgrams develops over time. Collgrams are defined as ‘n-grams to which an association score has been assigned’ (p. 280). The longitudinal data are written by the same students in their first (T1) and third (T3) year at university. With various statistical measures, the authors set out to determine the impact and importance of the time factor in the students’ use of collgrams, focusing on bigrams. Bestgen & Granger identify clear phraseological developments over the three-year span for all bigram categories studied: NN, JN, AJ as well as the total set of bigrams in the material. For example, T3 shows a more sophisticated use of collgrams, ‘as evidenced by a lower frequency of non-collocational bigrams [e.g. past people] and a constant increase in lower-frequency, but strongly associated bigrams [e.g. paramount importance]’ (p. 294). The authors also offer a valuable discussion of a longitudinal vs a pseudolongitudinal corpus design and the effect this might have on the results of a developmental investigation. A comparison is made between the current longitudinal study and Granger & Bestgen's (Reference Granger and Bestgen2014) previous pseudolongitudinal study on the use of bigrams by intermediate vs advanced learners. Very similar trends emerge; notably, ‘the differences between intermediate and advanced learners are completely in line with those obtained when the start and stages are compared over a three-year span’ (p. 290). Finally, some suggested pedagogical applications drawing on results from the study are presented, for example remedial exercises on the least appropriate collgrams and rewriting exercises on overused collgrams.

The volume comes across as a successful collection of good-quality, state-of-the art research in the field of corpus-based lexical studies. It covers a wide range of topics and fields dealing with English lexis or lexicogrammar from different perspectives, including lexicography, phraseology, pedagogy, varieties of English, learner language, register, spoken and written modes. The contributors have done an excellent job in presenting their research in both interesting and coherent ways. The editors are also commended for the fine volume this has become.