This volume stems from a workshop on ‘Current Methods in Historical Semantics’ at the 15th International Conference of English Historical Linguistics, held in Munich in August 2008. Those of us who were present at the workshop will remember a highly stimulating programme of papers punctuated by lively discussion. Revised versions of five of the papers are included here, with the addition of five others covering different aspects of the same theme. Between them, they encompass a wide range of semasiological and onomasiological approaches to the lexicon, as well as exploring the relationship between meaning and other levels of language. The focus is on new tools and techniques, using case studies to illustrate and to evaluate their research potential.

The volume is divided into three sections, each containing three or four papers followed by a commentary that reviews the issues from the perspective of a scholar working in a related field. Section 1 ‘Data and sources’ begins with a paper on the OED by Kathryn Allan, discussing the restructuring of entries for OED3 to prioritise dates of attestations in the ordering of senses. Case studies of the entries for milksop, pregnant and dull (the latter not yet revised for OED3) demonstrate the importance of looking closely at the chronology of semantic change. Christian J. Kay then describes The Historical Thesaurus of the OED (HTOED), the culmination of a 44-year research project at the University of Glasgow to arrange the contents of OED2, supplemented by Old English vocabulary from A Thesaurus of Old English (Roberts & Kay Reference Roberts, Kay and Grundy2000), into hierarchically structured conceptual fields (see Görlach Reference Görlach2011). A case study of terms from the semantic field of Truth and Error is followed by a brief outline of future plans to utilise the thesaurus database for purposes ranging from automatic text-processing to the study of metaphor.

The third paper, by Daphné Kerremans, Susanne Stegmayr and Hans-Jörg Schmid, presents another new tool, known as the NeoCrawler, designed to identify and to retrieve neologisms from the internet. Following a similar format to Kay, the authors begin by explaining the structure of the tool, and then proceed to a case study: in this instance, of the neologism detweet. First attested in May 2008, 144 occurrences of detweet were found up to April 2010, with five separate senses. The case study focuses on the NeoCrawler's ability not only to monitor the diffusion of the term, but to facilitate the analysis of factors leading to the development of different meanings. The commentary for this section is by Philip Durkin, Principal Etymologist to the OED, who discusses the interface between the electronic versions of OED and HTOED and emphasises the key role of dictionary data in the study of lexical change. His comments articulate with the views of Michael Samuels, the initiator of the HTOED project, whose speech at the launch event in Glasgow in October 2009 also stressed the importance of using HTOED and OED alongside each other.

Section 2 ‘Corpus-based methods’ turns to the analysis of large amounts of data in order to track changes across time. Here the focus is on new techniques rather than on new tools. Dirk Geeraerts, Caroline Gevaert and Dirk Speelman use quantitative corpus methods to compare occurrences of anger, ire and wrath in Middle English texts and to analyse the factors contributing to the rise of anger as the dominant term. Building on material from the second author's 2007 PhD thesis, their results affirm the role of social change in influencing lexical change, and also highlight the significance of lectal distribution. Quantitative analysis is combined with qualitative interpretation in the following paper by Martin Hilpert, who discusses the application of collostructional analysis to track changes in the meaning of grammatical constructions over time. Based on the principle that the meaning of a grammatical construction is reflected by its collocates, with shifting collocational patterns representing changes in meaning, his case study shows how changes in the collocation profile of the keep V-ing construction in the Corpus of Contemporary American English (COCA) from 1990 to 2007 reveal an increasing association with continuous rather than iterative activities.

In the final paper of this section, Eyal Sagi, Stefan Kaufmann and Brady Clark use the Middle English and Early Modern English sections of the Helsinki Corpus of English Texts to test the efficacy of Latent Semantic Analysis to track and to identify semantic changes. Again, the focus is on the surrounding context. Analysis of the well-known examples deer, dog, do and hound supports the hypothesis that the contexts of a word's occurrences over time reflect changes in meaning, such that semantic broadening results in occurrences across a wider range of contexts, while semantic narrowing leads to occurrences in fewer contexts. Reviewing all three papers and offering some additional analysis, the commentary by the computational linguist Stefan Th. Gries argues for a more ‘bottom-up’ approach to historical corpus data than in previous research, alongside a wider use of multifactorial methods.

Section 3 ‘Theoretical approaches’ applies methodologies from sociolinguistics, pragmatics and cognitive linguistics to the study of semantic change. Justyna A. Robinson employs the apparent-time construct to explore differences in usage by older and younger speakers and their impact on meaning development. Although she also draws on corpus and dictionary evidence, her analysis is primarily based on data from a survey involving 72 South Yorkshire speakers aged between 11 and 94. Through a case study of the polysemous adjective skinny, she investigates the relationship between patterns of use and the socio-demographic characteristics of speakers, demonstrating that age, gender and class are all significant factors.

In the only paper not focusing on English, Maj-Britt Mosegaard Hansen addresses the role of pragmatic inferencing in meaning development. Presenting a case study of the evolution of the negative reinforcers pas, mie and point in medieval French, she argues strongly for pragmatics as ‘a significant driver of semantic change’ (p. 253). Also in this section, Peter Koch discusses the role of contiguity in both lexical and grammatical meaning change, arguing that it is even more fundamental than metonymy. Finally, Gábor Györi and Irén Hegedüs examine the use of image schemas in semantic reconstruction, with a case study of chin and knee. The commentary by the historical linguist Terttu Nevalainen summarises the main points from each of the four papers, and adds further insights from Finnish.

A major theme running through the volume is what might be regarded as the Holy Grail of historical linguistics: the search for regularity in meaning change. As several contributors point out, semantic change is often contrasted with sound change in that the latter is regular whereas the former is not. Sound changes are predictable because the phonetic system is a closed set, so changes in one area lead to corresponding changes in others. By contrast, the semantic system is open-ended, so a change in one area of lexis need not affect the rest of the system. Moreover, the factors affecting semantic change appear to be more random than the factors affecting sound change, including a wider range of variables. Like the phonetic system, however, the semantic system is interrelated, and this raises the tantalising possibility that a better understanding of sense relationships may make it possible to identify predictable patterns of change. As Allan notes in her discussion of dull (p. 35), ‘One line of enquiry might be to examine the meanings of other semantically related lexemes in the period in which dull changes, to see if these might influence the development of new senses of dull.’ A similar point is made by Robinson (p. 225): ‘Since semantic change does not happen in isolation, taking an onomasiological perspective would allow for a consideration of the effect other items in the lexical field may have on individual uses of skinny.’ Hence the extraordinary significance of HTOED, which facilitates just such a line of enquiry. Especially now that the electronic HTOED is linked to the online OED and revised alongside OED3 – an advantage emphasised by Durkin (p. 97) – it offers an unparalleled tool for onomasiological investigation.

The issue of regularity is addressed most explicitly in the final paper by Györi and Hegedüs, who provide an overview of previous research tending towards the view that semantic change is generalizable although not regular (pp. 315–19). As regards the direction of meaning change, a widely accepted generalisation is the development from concrete to abstract. Even this, however, is challenged by Allan's investigation of three lexemes for which the available dating evidence appears to run counter to such an assumption. Whereas her meticulous analysis of milksop and pregnant establishes that in each case, the physical sense is likely to be etymologically earlier, she suggests plausibly that the mental sense of dull precedes the physical ones (pp. 32–6). Moreover, the metonymic development of witness from ‘testimony’ to ‘person giving testimony’ analysed by Koch again prioritises the abstract sense (pp. 283–4). In some cases, concepts of concrete and abstract may be culturally determined (see Hough Reference Hough2012b), but even taking this into account, there may be instances where the directionality rule does not hold. A definitive answer may eventually be provided by the Mapping Metaphor with the Historical Thesaurus project, described at the planning stage by Kay (p. 48), and now in progress at the University of Glasgow with funding from the Arts and Humanities Research Council (2012–14; PI Wendy Anderson).Footnote ¹ By using the thesaurus database to map the development of metaphor throughout the recorded history of English, the project will explore a range of theoretical issues, including the extent to which the concrete-to-abstract directionality prevails.

Several contributors focus on the observation of semantic change in progress, whether within a concentrated period of two years, as with Kerremans et al.'s study of detweet, or across several centuries, as with Sagi et al.'s study of deer, dog, do and hound. Again this may help to identify parallels between the semantic and other systems. A significant finding of Robinson's analysis of skinny is that ‘semantic change seems to be following similar routes of development to the types of change observed at other levels of language, such as phonology or morpho-syntax’ (p. 224). It is only comparatively recently that sociolinguistic variation has been linked to the actuation of the Great Vowel Shift (see Smith Reference Smith1993; Reference Smith1996: 86–111; Reference Smith2007: 127–53); such factors may be no less key to meaning shift. Kay's study of the HTOED category of Truth highlights the interface between cultural and linguistic factors in semantic change (pp. 52–5), while Geeraerts et al. show the importance of ‘interpreting the history of the vocabulary against the background of the history of culture and mentality’ (p. 129). Crucially, these complementary theoretical insights emerge from quite different approaches. Whereas Kay's analysis shows how much can be gained from displaying words in their semantic contexts, Geeraerts et al. (like Hansen) focus more on pragmatic contexts. Text type proves to be a significant factor in their analysis of the Middle and Early Modern English corpus, as indeed it is in Kerremans et al.'s analysis of material from the internet.

It may well be the case, as Györi and Hegedüs observe (p. 319), that sociocultural influences affect the content rather than the mechanisms of meaning change and are hence less easily generalisable than issues of directionality. As they also point out, however, certain conceptualisations occur across languages, appearing to represent general tendencies. One such is the semantic extension of the perceptual pattern ‘bend, curve’ to denote body parts with a curving shape, such as the knee, chin or hump. The question of which meaning component is diachronically earlier is raised within their discussion but not fully resolved (p. 326), and Nevalainen returns to this issue, noting that modern Finnish polvi ‘knee’ also means ‘bend’ in phrases such as joen polvi ‘river bend’ (p. 341). A similar polysemy is evidenced in the place-name record for Old and later stages of English (Parsons Reference Parsons2004: 130, s.v. cnēow), but both here and in other languages, it might be attributable to the (arguably universal) landscape is a body metaphor rather than to the bend/curve schema (see Hough Reference Hough2010: 11–13). The role of name evidence in historical linguistics is gradually achieving greater recognition (e.g. Durkin Reference Durkin2009: 266–83; Hough Reference Hough2012a; Scott Reference Scott2004), and although little considered within the present volume, it may be relevant to some parts of the discussion. For instance, the patterns of formation for tree-names discussed by Koch (pp. 296–8) are further illuminated by place-name material (see Hough Reference Hough2003).

A strong emphasis on empiricity characterises the volume as a whole, and reflects an important trend in twenty-first-century linguistics. While most evident in the increasing use of corpora and, in Robinson's paper, experimental data, the same principle underlies the restructuring of OED entries to prioritise dates of attestation over intuition regarding the order of sense development. A data-driven, evidence-based approach is fundamental both to the completed and ongoing research reported on, and to other projects still in prospect, such as the use of the HTOED database to calculate a relative chronology for the evolution of different linguistic categories (pp. 47–8), and the construction of a tool to identify automatically instances of semantic change within a diachronic corpus (p. 180). The contributors are also generally united in following the cognitive semantics paradigm that has become increasingly influential in recent years, and in exploring how cognitive perspectives can inform theories of meaning change. All papers are supported by extensive bibliographies, and it is noticeable how many of the references date from the 2000s. Although earlier, seminal publications are also brought into the discussion, the overwhelming emphasis is on up-to-date, cutting-edge research.

In short, the collection presents exciting and innovative research, and introduces powerful new tools and techniques to facilitate further advances in the field. The phrase ‘current methods’ in the title is thus amply justified. The phrase ‘historical semantics’, on the other hand, may require more justification. It will be evident from the above that the real focus of the volume is semantic change, and that the papers encompass contemporary as well as historical data. The editors attempt to address this in the introduction, explaining:

Nevertheless, ‘historical’ does not mean ‘diachronic’, so most readers will be surprised to find that three out of ten papers focus on the periods from 1990 to 2007 (Hilpert), from 2008 to 2010 (Kerremans et al.), and from the late twentieth century to the present (Robinson). Although defensible on the principle of ‘looking into the present to explain the past’ (p. 200), their inclusion can only be fully justified if the methodologies are transferable. This does seem to be the case for the collostructional approach pioneered by Hilpert. Issues relating to the analysis of older or more extended data are directly addressed during his paper, and robust protocols established. As regards the other two methodologies, however, the position may be less promising. In the final section of her paper, Robinson claims that ‘the conclusions drawn from the current study indicate the considerable benefits of taking a socio-linguistic approach in order to investigate semantic change in current communities and to suggest its potential for tracing semantic change in past communities’ (p. 226). However, since her analysis draws on data elicited through interviews, it is difficult to see how the same approach could be applied to past communities.

The NeoCrawler, designed specifically for the internet, is of even more limited transferability. As Durkin notes (p. 103), ‘this sort of data is not available to us for other historical periods, nor for items that have shown their initial spread in areas of discourse not well represented on the web’. During the discussion session following the First M. L. Samuels Lecture in October 2012, David Crystal remarked that the replacement in 2009 of Twitter's prompt ‘What are you doing?’ by ‘What's happening?’ had resulted in a huge change of vocabulary, and he joked that a PhD student researching the language of Twitter in 2007 would in 2010 suddenly find themselves studying historical linguistics. In this volume, the joke has become reality. Oddly, the traditional distinction between ‘historical’ and ‘diachronic’ is actually asserted by one of the contributors, as Gries observes that ‘historical semantics need not be diachronic semantics since it can be historical but nevertheless synchronic in nature’ (p. 185). I agree, but the volume as a whole seems to be posited on a different view.

One of the many strengths of the collection is the clarity with which complex ideas and methodologies are explained. The papers are written so as to be widely accessible, with only Sagi et al.'s description of Latent Semantic Analysis being somewhat too technical to be easily followed by a non-specialist. The contributors are not only experts in their fields, but skilled in the art of pedagogy, so the book can confidently be recommended to postgraduate students as well as forming a bedrock for future research. The editors have also done a fine job of summing up the main arguments in the introduction, situating each paper in the context of previous and current scholarship. Presentational matters, however, are less well handled. The papers are richly illustrated with figures and tables, but these are nowhere listed, and three lack captions. Capitalisation is inconsistent, even in the table of contents, and there are various anomalies in the indexes. Several papers would have benefited from more rigorous proof-reading, not only to eliminate typos, grammatical errors and non-existent cross-references, but to ensure consistency between in-text citations and bibliographies. Not all publications cited are included in the bibliographies; some appear with different dates, or without differentiating between two works by the same author in the same year. A few lack full publication details, or are listed out of alphabetical order. It is disconcerting to find the same publications cited differently in different papers; still more so to find them cited differently on the same page. These are trivial but irritating flaws: it is a pity that the high academic standards maintained by the contributors and editors are not matched by equally high standards of presentation.

Finally, it is interesting to see that the volume itself illustrates changes in language use. After trying for years to steer students away from the informal use of quote as a noun in academic essays, I gave up on finding the same usage in essay feedback written by younger colleagues. It was thus with a sense of wry resignation that I noticed it in Geeraerts et al.'s paper (p. 115). Another change that appears currently to be in progress is from the etymologically correct use of data as a plural form to a singular use already widely established in the sciences. The latter appears sporadically throughout this collection, leading me at first to assume that different contributors had different protocols. Since, however, both uses appear in individual papers, and even on individual pages (pp. 149, 205), they may reflect genuine variation. It remains to be seen whether both forms will continue to be used alongside each other in the linguistic literature, or if the etymological plural will either re-establish itself or be overtaken by the singular use.

In sum, this is a fine collection of papers, each strong in its own right, and together forming an impressive and coherent volume. The criteria against which the book is to be evaluated are set out in the series aims on the back cover:

These are ambitious objectives, but Current Methods in Historical Semantics meets the challenge. It is a worthy addition to the series.