Concessive Constructions in Varieties of English is the published version of Ole Schützler's postdoctoral thesis (German: Habilitationsschrift), for which he received the University of Bamberg's habilitation award (Habilitationspreis) in 2018. Schützler currently holds the position of Professor for Varieties of English at Leipzig University. The aim of the present monograph is to investigate correlations between functional and formal properties of complex-sentence constructions with a subordinate clause introduced by although, even though or though in nine varieties of English (British, Irish, Canadian, Australian, Jamaican, Nigerian, Indian, Singapore and Hong Kong English), based on both written and spoken data from the International Corpus of English (ICE). American English is notably absent, but only because the American component of the ICE has no spoken section and would thus be incomparable to the other components (p. 95). To my knowledge, this is the first book-length study of concessive constructions in English.

The book consists of twelve chapters. They are not grouped into larger sections, but can roughly be classified as follows: the first chapter is the Introduction, chapters 2–5 form the theoretical section of the monograph, chapters 6–11 focus on the methodology and empirical findings, and the twelfth chapter concludes the study.

Chapter 1, ‘Introduction’ (pp. 1–8), briefly discusses common approaches to concessive constructions in the literature (section 1.1) and it sets out the functional and formal aspects of concessives to be studied in the present monograph (section 1.2). The following four features take centre stage:

1. (i) the clause-linkage level between the subordinate and matrix clause (in terms of Sweetser Reference Sweetser1990);

2. (ii) the position of the subordinate clause relative to the matrix clause (final vs non-final position, the latter including both initial and medial placement of the subordinate clause);

3. (iii) the selection of the concessive marker (although, even though or though); and

4. (iv) whether the subordinate clause is finite or non-finite. Indicative and subjunctive verb forms are considered finite, while present and past participles as well as verbless (i.e. ‘reduced’) clauses are classified as non-finite.

Next, the author states the aim and scope of the investigation before giving an overview of the structure of the book (section 1.3). The chapter ends with a note on open data (section 1.4).

Chapter 2, ‘Concessive clauses: Development, function and form’ (pp. 9–34), elaborates on the aforementioned four features. It discusses diachronic aspects of concessive constructions, e.g. the development of the concessive markers although, even though and though from Old to Modern English (section 2.1). It also defines the clause-linkage levels found in concessives (section 2.2), distinguishing between ‘anticausal’, ‘epistemic’ and ‘dialogic’ concessives, which according to Schützler correspond to Sweetser's (Reference Sweetser1990) ‘content’, ‘epistemic’ and ‘speech-act’ levels, respectively (but see below). Finally, the chapter describes relevant syntactic features like the position of the subordinate clause and finite vs non-finite clause structures (section 2.3).

Chapter 3, ‘Corpus examples’ (pp. 35–53), illustrates the aforementioned clause-linkage levels with examples from the ICE corpus (sections 3.1–3). It also gives examples in which the linkage level is ambiguous (section 3.4) and discusses a handful of tokens that are syntactically interesting in various ways (section 3.5). This chapter thus serves as a qualitative description of the different types of concessive constructions found in the corpus and as an implicit reminder that the author's approach is strictly speaking semasiological, not onomasiological as claimed on p. 212, given that it investigates a formally defined set of constructions rather than all constructions (paratactic as well as hypotactic) serving a pre-defined function.

Chapter 4, ‘Dimensions and mechanisms of variation’ (pp. 55–73), first gives a brief overview of the work's theoretical framework, namely Construction Grammar, and discusses constructionist approaches to linguistic variation (section 4.1). Next, it introduces the main differences between spoken and written language (section 4.2) and discusses variation across varieties of English, including descriptions of several models of World Englishes (section 4.3).

Chapter 5, ‘Previous findings and research questions’ (pp. 75–92), is concerned with the status quaestionis. It gives an overview of how reference grammars of English have treated concessives and summarises the findings of several empirical studies (section 5.1). Based on this overview, the research gaps are identified (section 5.2) and the present study's research questions and hypotheses are developed (section 5.3).

Chapter 6, ‘Methodologies’ (pp. 93–114), discusses the ICE corpus and its components (section 6.1), data retrieval and annotation (section 6.2) and the different sorts of mixed-effects Bayesian regression models used in the study (section 6.3). These statistical models are picked up again in the following empirical chapters, each of which begins with a detailed description of the statistical model used in the respective chapter (sections 7.2, 8.1, 9.1, 10.1 and 11.1).

Chapter 7, ‘Frequencies of conjunctions’ (pp. 115–27), and chapter 8, ‘Frequencies of semantic types’ (pp. 129–35), discuss the overall frequency of the concessive markers, although, even though and though, and the frequency of the different clause-linkage levels, respectively. The following chapters go beyond text frequency analyses and investigate which functional and formal aspects correlate with each other. Chapter 9, ‘Clause position’ (pp. 137–49), looks into which factors determine whether the subordinate clause is positioned in final vs non-final position. Chapter 10, ‘Choice of the conjunction’ (pp. 151–78), investigates what determines the choice of the concessive marker: although, even though or though. Chapter 11, ‘Clause structure’ (pp. 179–99), shows which factors correlate with finite vs non-finite subordinate clauses.

Chapter 12 concludes the study (pp. 201–25), summarising its most important findings (section 12.1), situating the investigation in a broader context (section 12.2) and briefly returning to the relationship between concessives and Construction Grammar (section 12.3), before rounding off the main part of the monograph (section 12.4). The book also contains an appendix on the ICE corpus and its components (Appendix A), an appendix on the statistical models used in chapters 7–11 (Appendix B), a list of references, and indices of names, languages (mostly varieties of English) and subjects.

Concessive Constructions in Varieties of English can be downloaded for free on the website of the open-access publisher Language Science Press. Readers who prefer a print copy can order one via Amazon (print-on-demand). A part of the dataset can be accessed at the Tromsø Repository of Language and Linguistics (TROLLing), allowing other researchers to rerun Schützler's R scripts and replicate all statistical analyses. More information on open data is found in section 1.4 of the introductory chapter (p. 8). This level of replicability and commitment to free and accessible science is highly commendable, and will hopefully serve as a positive example for future studies in the field of linguistics.

Overall, empirical robustness and methodological transparency are key strengths of the present study. It is based on a solid amount of data from a well-balanced corpus, which is unfortunately still relatively uncommon in studies of adverbial clauses, especially concessivity. Each methodological step is explained in detail. Whenever data were simplified in the statistical analysis, this is clearly indicated and motivated. For example, Schützler subsumes subordinate clauses in initial and medial position under ‘non-final’ because a binary variable (‘final’ vs ‘non-final’) is required for his statistical model. This is justified given that initial and medial clauses are similar in terms of processing (p. 27) and because medial position is extremely rare (p. 83). Throughout the empirical part of the monograph, clearly written text passages are frequently complemented by graphs and other visualisations. This will make it easier for many readers to interpret the study's main findings, while appreciating the probabilistic nature of the data and the complexity of the many intercorrelations between the features under investigation.

Another commendable aspect of the present book is its honest and transparent attitude towards both negative and unexpected results. Whereas negative results are often deemed too uninteresting in the literature to be published at all and unexpected results are often brushed aside because they may contradict the researcher's hypothesis, Schützler takes them seriously. This is most notable in chapter 9, ‘Clause position’, where Schützler finds (i) that there is no clear correlation between clause-linkage levels and the position of the subordinate clause, and (ii) that non-final position is more common in L1 varieties than L2 varieties, contradicting his working hypothesis that L2 varieties prefer non-final order given that it is considered cognitively optimal (p. 147). To the extent that he does find negative or unexpected results, Schützler usually makes suggestions as to how future research could improve on his work. For example, he suggests that clause position in concessives is most likely determined by information-structural factors, which are not taken into the account in the present study (p. 149). Such factors are explicitly excluded in order to enable a quantitative approach (p. 149).

A further instance of negative results is found in section 12.2.4, where it is stated that concessive constructions do not show many systematic differences across varieties of English except for a few idiosyncratic features such as the high frequency of though compared to although and even though in Indian English. According to Schützler, adding data from more varieties would only result in even more of such idiosyncratic patterns. Some readers may find this result disappointing, given that varieties of English are so prominent in the monograph's title, but it is not entirely unexpected. It is well known that concessives, especially hypotactic ones, primarily occur in written standard registers (pp. 76, 84 and references cited there), suggesting fewer inter-varietal differences than in informal spoken registers. While the inclusion of several less-studied L2-varieties of English is laudable, one wonders whether some apparent idiosyncrasies found by Schütlzer (e.g. some of the ‘syntactically interesting’ examples in section 3.5) might not be due to language contact, a topic that is not mentioned in the present study at all.

Although the book is generally written in an accessible and transparent style, chapter 6 on methodology and the descriptions of the statistical models at the beginning of chapters 7–11 (see above) are at times rather technical. Schützler explains the basic concept of Bayesian statistics (section 6.3.2), but this does not mean that novices will understand the models. A more elaborate explanation would obviously be beyond the scope of the present study, however. Fortunately, interpreting the study's main findings is possible even with limited experience with statistics.

Schützler adopts a constructionist framework in the present study. While the study clearly approaches concessives from a functionalist and usage-based perspective, Construction Grammar itself plays a relatively minor role throughout the monograph. Construction Grammar is discussed in sections 4.1 and 12.3, but barely mentioned elsewhere, including the main empirical chapters 7–11. Construction grammarians may be disappointed by this and would likely disagree with certain passages, e.g. when Schützler equates allostructions and constructs as if these terms were synonyms (p. 57). By contrast, readers who are unfamiliar or disagree with Construction Grammar will not be distracted by an overabundance of framework-specific terminology.

Whereas quantitative empirical robustness and methodological transparency are notable strengths of the present volume, a few weaknesses can be identified in the theoretical part. The first issue pertains to the definition of concessivity. Concessivity is not explicitly defined until p. 15, where a definition by Quirk et al. (1985: 1098) is given: ‘[c]oncessive clauses indicate that the situation in the matrix clause is contrary to expectation in the light of what is said in the concessive clause’. Given how well Schützler knows the literature, this definition is rather simplistic: while ‘contrariness to expectation’ is indeed a key feature of concessivity, what is missing is the fact that the propositions in the subordinate and matrix clause of a concessive construction are both factual, a feature often mentioned in the literature (e.g. König Reference König and Hawkins1988: 146). This inaccuracy becomes apparent when Schützler compares concessives proper with concessive conditionals (or ‘conditional-concessives’, an outdated term adopted by Schützler from Quirk et al. Reference Quirk, Greenbaum, Leech and Svartvik1985). In short, concessive conditionals are ‘quantified conditionals’ (Leuschner Reference Leuschner2020) that express a set of hypothetical antecedents in the subordinate clause that all lead to the same consequent in the matrix clause: ‘if {p₁ , p₂ , …}, then q’ (Leuschner Reference Leuschner2020). Both concessives proper and concessive conditionals (typically) have a factual matrix clause, but the key difference is that the subordinate clause in the former contains a single, factual value, whereas it contains a set of hypothetical values in the latter. A concessive conditional like In the mornings I scoured the breakfast pans, whether or not it was my turn (ex. 3a, p. 12), therefore, cannot felicitously be paraphrased with a concessive proper like This morning I scoured the breakfast pans even though it was not my turn, as Schützler suggests (ex. 3c, p. 12). The concessive-conditional clause expresses two hypothetical possibilities (‘my turn’ and ‘not my turn’), whereas the concessive proper expressed only one factual value. Fortunately, this issue remains largely inconsequential, as none of the conjunctions under investigation in the present study has a concessive-conditional reading in present-day English.

The second issue, however, does have considerable implications for the entire study: the definition of the so-called ‘dialogic’ level of clause linkage. This is significant because clause-linkage levels (cf. Sweetser Reference Sweetser1990), to which Schützler refers by the vague and imprecise term ‘semantic types’, play a prominent role throughout the monograph; see sections 2.2, 3.1–4, 9.2.2, 10.2.2, 11.2.2 and all of chapter 8. According to Schützler, dialogic concessives correspond to what Sweetser (Reference Sweetser1990) and Crevels (Reference Crevels, Couper-Kuhlen and Kortman2000) call ‘speech-act concessives’ (pp. 14–15 inter alia). Speech-act concessives are defined as concessive constructions in which

An example would be Although it's none of my business, your behaviour is a disgrace, in which the subordinate clause could be paraphrased as ‘I know I have no business commenting on your behaviour, so I normally wouldn't do so’ (Crevels Reference Crevels, Couper-Kuhlen and Kortman2000: 318).

Schützler instead proposes the new term ‘dialogic concessives’, arguing that ‘a broader designation is needed … since different notions exist as to what precisely constitutes a speech act’ (p. 21). The new category is so broad, however, as to be no longer co-extensive with the original category of speech-act concessives. Dialogic concessives are essentially treated like a sort of residual category, comprising all complex-sentence constructions with although, even though or though that are not clear instances of either anticausal or epistemic concessives, e.g. Although surgery is best, it is not always possible (ex. 19, p. 23, cited from Hilpert Reference Hilpert2013), where there is no logical opposition between what is in principle the best option and what is actually feasible.

Whereas speech-act concessives as defined by Sweetser (Reference Sweetser1990) and Crevels (Reference Crevels, Couper-Kuhlen and Kortman2000) form a niche category, the category of dialogic concessives is ‘broad’ (p. 21) and ‘heterogeneous’ (p. 42), representing the most frequent clause-linkage level in Schützler's data (pp. 131–2). This is presented as a new and surprising finding that contradicts the existing literature (pp. 131–2, also p. 204), but it is actually a rather unsurprising consequence of the present study's much broader definition.

A plausible alternative could be to treat the majority of problematic cases as an entirely different category, viz. adversatives or adversative-like concessives. In fact, Schützler himself notes that in many dialogic concessives ‘[t]he relationship between propositions might be argued to be adversative, rather than concessive in the strict sense of the word’ (p. 16). What his study finds is thus not that speech-act concessives are more frequent than previously thought, but that concessive conjunctions more often code adversativity instead of genuine concessivity than noticed so far. This seems to be especially true of although and though but to a lesser extent of even though, which shows a positive correlation with anticausal clause-linkage (section 10.2).

These pertinent shortcomings notwithstanding, Concessive Constructions in Varieties of English is a welcome addition to the field. The study's findings are based on a large amount of corpus data, a feature missing in most of the seminal papers on concessivity from the 1980s. It uses state-of-the-art multivariate statistics to untangle a complex web of intercorrelated variables, revealing probabilistic patterns in the choice of conjunction, position and structure of the subordinate clause, and clause-linkage levels. Furthermore, it is the first study on English concessives to use data from varieties other than British or American English. Hopefully it inspires future studies on concessives in one or more varieties of English.