The volume Gender in World Englishes, edited by Tobias Bernaisch, explores genderlectal variation in native, second- and foreign-language varieties of English. As outlined in the editor's Introduction, the eight contributions (excluding the Introduction) approach this theme from two angles: on the one hand, linguistic phenomena that have previously been found to display genderlectal variation are here analysed in the World Englishes context (chapters 2–5); on the other hand, phenomena already shown to vary across World Englishes are now approached from the perspective of gender (chapters 6–9). This rationale is very promising as it takes a position between Sociolinguistics and (corpus-based) World Englishes research and focuses on one particular social variable – thus, the book very successfully fills a gap in the well-trodden field of variation studies in the wider sense, and this in itself is a praiseworthy achievement. If this thematic orientation is one selling point of the volume, another strongly emphasised aspect is its robustly quantitative orientation: individual studies typically (but not exclusively) use different national components of the International Corpus of English (ICE; cf. Kirk & Nelson Reference Kirk and Nelson2018) and rely ‘on the currently most advanced empirical methods’ (p. 17), to quote from the Introduction. Apart from providing brief summaries of individual contributions, this review will therefore focus on these two aspects: (i) how far does the volume add to our understanding of variation along the dimensions of gender and/or World Englishes, and (ii) how well are the analyses supported by the quantitative methods that are employed?
The Introduction by Tobias Bernaisch lays out the rationale of the volume as described above (pp. 1–22). It contains a debate on English-as-a-second-language (ESL) and English-as-a-foreign-language (EFL) varieties and motivates the inclusion of the latter in some of the contributions. The discussion of gender in linguistic research naturally takes centre stage. As the editor explains, contributions in the volume follow a relatively traditional approach in which social gender is effectively measured as biological sex, mainly due to the unavailability of more detailed corpus metadata. Further, and as mentioned above, the editor stresses the importance of using appropriate quantitative techniques in corpus-linguistic research and presents a short typology of approaches. Finally, the Introduction provides summaries of the eight individual contributions to the volume.
In chapter 2 (‘Localisation, globalisation and gender in discourse-pragmatic variation in Ghanaian English’, pp. 23–46), Beke Hansen focuses on the quotative system of Ghanaian English (GhanE) and the rates of occurrence of be like compared to other quotative markers, using the private dialogues sections from a preliminary version of ICE-Ghana. The corresponding parts of ICE-Canada are used as an L1 point of reference. The author finds that language-internal factors (such as grammatical subject, tense or the difference between the quoting of thoughts/attitudes and actual utterances) have similar effects in Canadian English (CanE) and GhanE, while gender itself does not play a major role. In Hansen's interpretation, be like has entered the English language as a ‘ready-made template’ used to report inner monologue/thoughts, combining with first-person subjects and employing the historical present as a highlighting device in past-tense reporting. The study benefits from a very careful approach to the data (the author identified all quotatives, essentially by reading the corpus), a very sensible and informative presentation of previous studies, some statistics concerning the literacy and schooling of men and women in Ghana, and a helpful discussion of concrete L1 influence from Akan languages. Concerning the quantitative analysis, implementing fully parallel models for CanE and GhanE would facilitate the direct comparison of these varieties – as it stands, only the model based on GhanE contains a term for the interaction of age and gender. Further, the specification of random intercepts for speakers/texts and a visualisation of predicted probabilities instead of (or in addition to) tables of coefficients would have improved an otherwise very interesting and focused chapter.
In chapter 3 (‘Sociolinguistic variation in intensifier usage in Indian and British English: Gender and language in the inner and outer circle’, pp. 47–68), Robert Fuchs investigates the use of intensifiers (so, very, really, too, especially and particularly) in Indian English (IndE) as compared to British English (BrE), using four text categories from the respective national components of ICE. The expectation that women use more intensifiers than men is confirmed for informal registers only. However, women appear to be more sensitive to differences in formality in IndE. Interestingly, there is also a considerably higher rate of occurrence in all-female groups in this variety, which is levelled in mixed-gender groups. Fuchs interprets the different patterns as symptoms of more traditional gender roles in India, with women responding strongly to formal contexts and/or the presence of men. The chapter asks pertinent questions concerning differences in gender construction between India and the West, and it also offers an excellent review of traditional approaches to gender differences in language. The statistical approach is very tidy (for instance: normalizing frequencies at the level of the individual speaker), the visualization simple but maximally informative. It could be asked if it might have been an option to use a count model with a nonlinear link-function (e.g. Poisson or negative binomial; e.g. Cameron & Trivedi Reference Cameron and Trivedi2013) instead of a linear regression model or – perhaps to a similar effect – a linear model predicting logged frequencies. These approaches would reflect relative differences in frequency, which might be more appropriate when predicting values across a wide frequency range. In the case at hand, however, this would in all likelihood not have changed the general conclusions.
In chapter 4 (pp. 69–93), Claudia Lange and Sven Leuckert investigate tag questions in Indian English (IndE). Based on the private-dialogues section of ICE-India, they take a look at the formal realisation of the tag (canonical, invariant or indigenous no/na) and the pragmatic meaning of the tag (e.g. ‘informational’, ‘confirmatory’ or ‘facilitating’, among others). In a way, each of these two parameters constitutes a dependent variable in its own right, while gender and age are the predictor variables. Indian women use more tags than men, but there is little correlation between formal or pragmatic tag types and gender. Interestingly, the overall rate of tags is highest in all-female groups, while it is lowest in all-male groups and relatively low in mixed groups – a pattern that dovetails with the IndE findings in chapter 3 by Fuchs. Lange and Leuckert differentiate nicely between pragmatic types of tags, particularly concerning their relevance in a gender context, and they also add very important nuances to the discourse on gender in the Indian context. The quantitative analyses in this chapter are strongest when using descriptive statistics based on the raw data. Results from an inference tree and a random forest are not reported, and the additional monofactorial/pairwise association measures that the authors apply produce a relatively abstract picture. A more focused operationalisation of the outcome variable in this admittedly very complex phenomenon would probably make it possible to use (multifactorial) regression modelling and generate more unified results.
Chapter 5 by Tobias Bernaisch focuses on English hedges in Hong Kong, the Philippines and Singapore, using British English as a reference variety (‘Hedges and gender in the inner and expanding circle’, pp. 94–120). It is exceptional in using learner data from the International Corpus Network of Asian Learners of English (ICNALE). There is an excellent short introduction to hedges, and the selection of eight clausal and eight non-clausal formal variants for the analysis (of which five infrequent ones were subsequently excluded) seems very reasonable. The author finds that men use more hedges than women in Britain and Singapore, while the opposite pattern is found in Hong Kong and the Philippines. In discussing his results, Bernaisch comes to the conclusion that each variety essentially contains a set of Kachruvian circles en miniature, with native, second-language and foreign-language users. This systematic intra-varietal heterogeneity of speakers may be unsurprising, but Bernaisch addresses it explicitly and shows how it can actually be included in empirical research, given the right data – this certainly provides food for thought regarding future studies. In the quantitative analysis, neither the conditional inference tree nor the multinomial regression takes the hierarchical organisation of the data (i.e. observations clustered within speakers) into account – at least in the regression model this could have been implemented via random effects for speakers/writers. Moreover, showing the proportions of the n = 11 outcome levels predicted by the multinomial regression model would have further strengthened this highly instructive and inspiring chapter.
In chapter 6, Stefan Thomas Gries, Benedikt Heller and Nina Funke inspect the genitive alternation in BrE and Sri Lankan English (SriLE), using parts of the respective components of ICE in a random forest analysis (‘The role of gender in postcolonial syntactic choice-making: Evidence from the genitive alternation in British and Sri Lankan English’, pp. 121–46). Apart from confirming well-established effects of language-internal factors such as possessor animacy or the length difference between possessum and possessor, the authors show that genderlectal variation is greater in SriLE and that female speakers use a higher proportion of s-genitives. Transfer from Sinhala, Sri Lanka's most important native language, is suggested as one possible explanation, since in that language possessors precede possessums – although it is less obvious why this kind of transfer should predominantly affect female speech behaviour. The authors also discuss how the rather strict constraints concerning the correlation of animacy and definiteness with the selection of different possessive constructions in Sinhala may strengthen those constraints within the English genitive. Ideas of this kind could be of great general relevance in language contact settings. The authors base their decision to use a random-forest model on the fact that the outcome variable is skewed (i.e. strongly leaning towards the of-genitive) and that there are many speakers with relatively few data points. It would be a good idea to strengthen this argument by testing it empirically, i.e. by fitting a mixed-effects model before opting for random forests – after all, the random forest partly addresses the potential issue of data sparseness at the speaker level by ignoring the nestedness of the data.
Chapter 7 by Melanie Röthlisberger investigates the variability of the ditransitive construction (also known as the dative alternation) in Jamaican English (JamE), using data from ICE-Jamaica (‘Social constraints on syntactic variation: The role of gender in Jamaican English ditransitive constructions’, pp. 147–75). She finds a small preference for prepositional datives by female speakers. As an explanation, Röthlisberger discusses different degrees of Jamaican Creole contact to which men and women are exposed. The chapter is truly impressive in its careful approach to the coding and preparation of the data. There is an excellent discussion of the two methodologies that are applied and their respective roles in the analysis: a random forest is used to explore the data and motivate the selection of predictor variables for the subsequent regression-based investigation, whose outcomes are then presented with clearly interpretable effects plots. Röthlisberger sensibly decides to include only those two internal factors shown (by the random forest) to have the strongest effects, namely the recipient/theme weight ratio and recipient pronominality. Actively simplifying the model structure to meet the goals of the analysis, rather than including every possible internal factor, results in a strongly governed and theory-led quantitative analysis. One could perhaps argue that the exclusion of non-significant random variables (in this case speaker and recipient head) is not necessary, as leaving them in the model comes at no cost.
In chapter 8 (‘Linguistic colloquialisation, democratisation and gender in Asian Englishes’, pp. 176–204), Lucía Loreiro-Porto explores the role of gender in the colloquialisation and democratisation of Hong Kong English (HKE) and Indian English (IndE). Colloquialisation features are the contraction of primary and modal verbs, analytic not-negation (vs integrated no-negation) and the future with going to (vs will-future); tokens of democratisation are semi-modals of obligation (instead of must), gender-neutral occupational nouns and gender-neutral epicene pronouns (singular they or he or she instead of generic he). This representation of each process by three features is an excellent idea. The author finds that HKE is characterised by higher degrees of both democratisation and colloquialisation, with stability in apparent time, while there is a female-led democratising/colloquialising development in IndE – a difference for which highly plausible historical explanations are offered. In the quantitative analysis, using a chi-squared test to compare groups within a large set of non-independent observations seems somewhat problematic, and the same could be said for the absence of a random part in the logistic regression models: p-values for effects at the level of the speaker (age and gender) are bound to be overly optimistic – in other words, Type I errors become more likely. In this chapter, this seems particularly important because odds ratios for non-significant effects are not even reported in the respective tables; that is, whether or not a certain piece of information is conveyed to the reader partly depends on p-values.
Chapter 9 (‘Gender, writing and editing in South African Englishes: A case study of the genitive alternation’, pp. 205–32), by Melanie A. Law and Haidee Kotze, focuses on the genitive alternation in Black South African English (BSAfE), Africaans English (AfrE) and White South African English (WSAfE). Instead of the standard corpora typically used in World Englishes research, they use their own corpus, comprised of unedited texts and their edited counterparts. The authors establish a link between the norm-generating linguistic behaviour of women and the move of (post-colonial) varieties towards endonormativity, as in Stage 4 of Schneider's (Reference Schneider2003) Dynamic Model – a highly promising idea that should be explored more in future research. In all three varieties, and in both writing and editing, gender is not a strong predictor of the alternation; previously described internal factors are much more important. The main analysis proceeds from very informative descriptive statistics to a somewhat less informative generalised linear mixed-effects model tree. According to the logic of the tree, rather specific conditions are discussed, as in the following: ‘Where a final sibilant is present in cases with animate possessors where orum_lendiff < 9, definites show a much stronger preference for the s-genitives than indefinites (node 9 and node 10)’ (p. 223). This is not easy to comprehend, and a more generalising presentation in this part of the chapter would provide better support for an otherwise excellent and inspired (in fact: inspiring) analysis.
What, then, is the overall achievement of this volume? As stated initially, it will be very much welcomed by the community due to its particular position at the interface of Sociolinguistics and (corpus-based) World Englishes research. It will be an extremely valuable and at present unrivalled reference point for future studies that aim to describe the development of varieties in greater social detail – studies, for instance, that theorise within the framework of Schneider's (2003) Dynamic Model or the External and Internal Forces Model by Buschfeld & Kautzsch (Reference Buschfeld and Kautzsch2017), but strive to go beyond a black-box approach by looking at the actual social mechanisms of varietal development and change. Several chapters make a positive and often substantial contribution to theory-building in the World Englishes context, either by using data that go beyond the standard corpora from the ICE family (e.g. Bernaisch; Law & Kotze) or by actively engaging with the historical or contact-related reasons for differences between varieties (e.g. Hansen; Gries, Heller & Funke; Röthlisberger; Loreiro-Porto; Law & Kotze). Very helpful discussions of gender in the respective societies/cultures are provided by Hansen, as well as Fuchs and Lange & Leuckert. One might still wish for a more comprehensive social elaboration, but in an empirically oriented volume those aspects obviously need to be kept brief. The chapters by Hansen as well as Gries, Heller & Funke and Law & Kotze actively engage with the potential impact of L1 structures in the respective countries. This raises the overall quality and interest of the volume even further, as does the fact that it is well-known linguistic phenomena that are under scrutiny, and the volume as a whole thus builds admirably upon previous research.
The quantitative focus of the volume is foregrounded quite strongly by the editor, who states that ‘[t]he identification and correct application of an adequate statistical model will make or break a corpus-linguistic study’ (p. 7). However, research questions/hypotheses and modelling techniques do not always appear to be seamlessly integrated in individual contributions. By no means does this ‘break’ the respective studies, but it may at times weaken the link between model output and the conclusions that are drawn, and thus make it more difficult for the reader to follow the empirical narrative. In some cases, the analysis could be further strengthened by building models that take the hierarchical structure of the data into account (Hansen; Bernaisch; Gries, Heller & Funke; Röthlisberger; Loureiro-Porto). In some other chapters, the research designs appear somewhat too complex for the quantitative methods used to address them (Lange & Leuckert; Loureiro-Porto) – in this case the way forward might lie in reconsidering the operationalisation of the outcome variables themselves. Finally, supporting all analyses with visualisations that show, in a nutshell, the outcome values and/or contrasts of interest would help to make results more accessible (i.e. comprehensible) for readers. The statement in the introduction that ‘multifactorial designs are the current gold standard’ (p. 11) is certainly true – perhaps even a variationist universal – but it is also in need of qualification: multifactorial methods do not succeed by default but need to be actively motivated and combined with a transparent communication of results. In this regard, the present volume could serve as evidence that the ‘quantitative turn’ in linguistics, apart from the positive effect of making linguistics more robustly empirical, is not without its complications (cf. Sönning & Werner Reference Sönning and Werner2021). On a related note, the publication of data and analysis routines would be of immense value since the community, in collaboration with the authors, could then respond to the above remarks in an empirical fashion. After all, the editor's assessment that linguistic analyses in corpora are ‘intersubjectively verifiable’ (p. 6) is only realistic in an open data environment.
To return to the main topic and title of the volume, Gender in World Englishes, the studies that are presented strike an excellent balance between a conservative, corpus-based approach that is limited by the nature of the available (binary, sex-based) metadata, and out-of-the-box thinking that points towards a more nuanced treatment of gender in specific (postcolonial) cultures and, at a more global level, in the World Englishes paradigm. Operationalising gender in a more complex way was understandably beyond the scope of this particular volume, but when such attempts are made in the future, they can rely on Bernaisch's volume to provide a well-prepared ground of departure. Gender in World Englishes is sure to remain a milestone publication for some time to come, a very successful and thought-provoking approach to an extremely challenging topic within World Englishes research.
