The Dynamic Model of Postcolonial English (Schneider Reference Schneider2003, Reference Schneider2007) has become the most widely used model to account for the relationships among varieties of English in former colonies of the British Empire, alongside a few other territories under the control of other colonial powers. Thomas Hoffmann points out that ‘the Dynamic Model is essentially a sociolinguistic model’ (p. 3), and sets out his proposal for a ‘Constructionist Grammar Approach to the Dynamic Model’ which is intended to enhance the linguistic dimension of the model. He does this by introducing readers to both bodies of ideas, Construction Grammar and the Dynamic Model, before proceeding to work out the connections between them. In linking these two approaches, he develops a specific hypothesis, the Dynamic Model Productivity Hypothesis, which he subjects to empirical evaluation by analysing three grammatical constructions across varieties in different stages of development along the predicted path of the Dynamic Model. Hoffmann's proposed Constructionist Grammar Approach to the Dynamic Model is intended to provide firmer linguistic footing to the study of world Englishes, both in the substantive sense of enlarging the scope of investigation to a more thorough-going account of how constructions can potentially develop and change as varieties move along the path of the five phases of the Dynamic Model, and in the epistemological sense of contributing more rigour to the study of the linguistic features of world Englishes. This publication forms part of the new Elements in World Englishes series, which consists of texts intermediate in length between a journal article and a full-length monograph.

The Element is introduced in section 1 (pp. 1–3) with a review of models of world Englishes, identifying the value of a dynamic model over a static one, to account for ongoing change in English as it continues to spread and expand its uses in different societies. This leads Hoffmann to an endorsement of Schneider's Dynamic Model, while characterising it as primarily sociolinguistic. He notes the strong focus on the forces of identity construction, and identifies the need to examine and better understand the linguistic aspects of ongoing changes in world Englishes. His proposed Constructionist Grammar Approach to the Dynamic Model is intended to fill this conceptual gap in the Dynamic Model.

Given the dual focus of the proposal, section 2 (pp. 4–22) offers a review of Usage-based Construction Grammar and of the Dynamic Model. Both reviews are aimed at readers with limited prior knowledge of the specifics, and serve to create common ground for the last part of the section where the two approaches are brought together. The combination of insights from Construction Grammar with the Dynamic Model in section 2.3 takes the shape of a close look at the nature of linguistic innovations in Englishes as they spread to new settings on the back of colonial occupation and settlement. In phase 1, the foundation phase of the Dynamic Model, where there is limited contact between the settlers (STL strand in Schneider's model) and the indigenous people (IDG strand), the linguistic changes are limited to toponymic borrowing. Hoffmann accounts for this through Construction Grammar as instances of ‘a limited number of fully substantive constructions’ (p. 14). He explains that these borrowings are easy to learn for the STL users, since they have clear referential meaning and high token frequency. Alongside such borrowings by STL users, Hoffmann (pp. 14–17) also applies Construction Grammar to the grammatical patterns that emerge in pidgins and later in creoles in plantation colonies, also during phase 1. These options are not explored in the empirical studies in the second half of this Element, but are interesting suggestions that can be taken up by future research, at the interface between Pidgin/Creole/Contact Linguistics and Construction Grammar, and are potentially relevant beyond world Englishes, but nevertheless maintain a connection between the creole and pidgin Englishes and the broader family of world Englishes.

During phase 2, STL speakers borrow more vocabulary, mostly words denoting classes of objects in the natural environment and objects and practices associated with the lives and culture of the indigenous inhabitants. While less directly referential, these terms are nevertheless still straightforward micro-constructions that are easy to acquire and to be incorporated into the ‘taxonomic constructional networks of the STL strand’ (p. 17). Also during phase 2, in plantation colonies, the slaves who started out speaking pidgin forms in phase 1 expand their pidgins into creoles. This is associated with meso-constructions of a higher level of schematicity than the vocabulary items of the STL users, but these constructions are still analytic with close form–function associations, rather than synthetic constructions where the form–function pairings are less transparent. Hoffmann continues to account for features of pidgin/creole Englishes in terms of the concepts derived from Construction Grammar as part of his account of the linguistic properties of Englishes at different developmental stages, taking his cue from the kinds of linguistic features that Schneider (Reference Schneider2007: 29–55) identified as characteristic for different settlement types in the five consecutive phases of the Dynamic Model.

The IDG strand comes to play a much more prominent role in phase 3, when they adopt and adapt constructions in English during the course of their encounters with the English language – be that in direct contact with the STL speakers or possibly through formal education. Hoffmann (p. 19) identifies meso-constructions with an intermediate degree of schematicity as the area in the grammar of Englishes in phase 3 where most of the action takes place. His thesis is that the earliest adoption of constructions will stay close to those lexical items that instantiate the prototypical uses of the construction, and thus that constructions will tend to be less schematic in these early phases of adoption, whereas at more mature levels of use the constructions will become more general, more abstract and less tied to prototypical lexical forms to fill the schematic slots.

Developments in phases 4 and 5 of the Dynamic Model are presented very briefly and not so much in terms of further grammatical changes, but rather in terms of the social status of constructional networks. Hoffmann (p. 20) observes that the construction network stabilises among STL and IDG users alike in phase 4, hence appearing to be homogenous as set out by Schneider (Reference Schneider2007: 51). In phase 5, internal diversity within constructions may acquire stronger associations with new identity positions within the postcolonial society, at which point Hoffmann (p. 20) observes that the Dynamic Model shares concerns with cognitive sociolinguistics, which is a matter dealt with in more detail by Hollmann (Reference Hollmann, Hoffmann and Trousdale2013), to which the reader is referred.

Based on the joint exploration of the phases of the Dynamic Model and a Construction Grammar perspective on the linguistic developments along these phases, Hoffmann formulates the Dynamic Model Productivity hypothesis as an empirically verifiable/falsifiable account of the differences between phases: ‘Varieties in later phases of the Dynamic Model show (1) more productivity of the slots of (semi-)schematic construction[s] than varieties in earlier developmental phases due to (2) less reliance on prototypical and frequent fillers’ (p. 21).

Section 3 (pp. 22–30) presents the data and methodology for the empirical evaluation of the Dynamic Model Productivity Hypothesis. A justification is presented for why a corpus analysis is a suitable method for this hypothesis. The corpus selected is the corpus of Global Web-Based English (GloWbE; Davies Reference Davies2013), given that it is ‘by far the largest corpus of postcolonial Englishes available’ (p. 24), the relatively low frequency of the constructions under analysis and the need for the widest possible range of lexical items used in these constructions in order to evaluate the hypotheses. Those national varieties of English represented in the corpus are placed into phases 2 to 5, in accordance with available literature, with the grey areas for placement afforded some discussion to justify the classification decisions. A detailed presentation of the two main statistical analyses – the Large Number of Rare Events (LNRE) and the Configurational Frequency Analysis (CFA) – conclude the section, with explanation of the two complementary aspects of productivity that they measure. The LNRE evaluates how many different lexical items are selected to fill the schematic slots in the constructions under investigation – with a wider diversity of lexemes corresponding to more schematicity for the construction, and thus predicted to match a variety more advanced towards phase 5 of the Dynamic Model. The CFA determines the most characteristic lexemes for the construction, in order to establish the extent to which the use of the construction goes beyond the semantic prototype or stays close to that prototype, the latter option being matched to an earlier developmental phase than the former.

The empirical evaluation in section 4 (pp. 30–46, ‘Sample studies’) is focused on three different constructions: the Way construction (‘the boy works his way into the lives of a middle class family’, pp. 30–5), the V the hell out of NP construction (‘Quentin acts the hell out of this scene’, pp. 35–44), and the As ADJ as a N construction (‘The town was as quiet as a cemetery’, pp. 44–6). In each case, the LNRE analysis confirms the Dynamic Model Productivity Hypothesis: varieties in phase 5 have the largest number of different lexemes in the schematic slots of the construction, those in phase 2 the lowest, and the other two in between. Unfortunately, the number of varieties in phases 2 and 4 represented in GloWbE is relatively few, but phases 3 and 5 are well represented, and these two phases are consistently separated by means of the LNRE test. Even if the differences are not statistically significant in all cases, varieties in phases 2 and 4 pattern in ways consistent with Hoffmann's hypothesis. For the Way construction, the closer examination of the specific lexemes by means of the CFA also supports the hypothesis, but no statistical support is obtained for the other two constructions.

Hoffmann (pp. 39–44) undertakes a closer look at the constructional semantics of the V the hell out of NP construction, and finds that the rich data in GloWbE enable a more refined semantic characterisation of the construction cutting across varieties. These are the physical transfer of force from the agent to the patient (e.g. ‘he beat the hell out of me’) and the way a stimulus mentally affects an experiencer (e.g. ‘enough to surprise the shit out of them’) senses.

The conclusion (section 5, pp. 46–7) reiterates that the Dynamic Model Productivity Hypothesis receives support from all three constructions analysed, and then refines the insight by considering the quantitative and qualitative aspects. Quantitative similarity is expected for varieties at the same phase of development, while they are bound to differ more and more qualitatively in terms of which lexemes are selected for the schematic slots of constructions.

Hoffmann's contribution is a convincing one, which shows the value of infusing Schneider's Dynamic Model with insights from Construction Grammar, to elaborate the linguistic, cognitive side of the model that is otherwise stronger on the social side. The manner in which the grammatical phenomena in the different phases are interpreted is particularly insightful, to develop the thesis that different types of contact are likely to have different consequences for the grammars in the minds of speakers. The account is particularly convincing in dealing with phases 1–3 of the Dynamic Model, but is presented in a less detailed fashion for phases 4 and 5. These two phases are those where social factors play a much stronger role, and it would thus enhance this account and the desired connection between the cognitive and social if the implications of the identity rewriting in phases 4 and 5 could also be incorporated into the grammatical representations, even if that means reaching beyond world Englishes to venture in more detail into cognitive sociolinguistics.

The biggest concern with the current study is the extent to which the phases of the Dynamic Model are the underlying force that motivates the observed differences in productivity and schematicity. To the extent that users continue with a limited selection of lexemes for a construction and with a narrower, more prototypical semantic range for constructions, the question has to be asked whether that is due to less extensive acquisition of the language, at least by some of the contributors to websites that were included in GloWbE. Thus, are we dealing with a psycholinguistic learning phenomenon or a sociolinguistic developmental phenomenon? These are extremely hard to disentangle, but it would be worthwhile for future research in support of the endeavour in this publication to take on this difficult question. The explanatory power of the Dynamic Model, including the degree of insight which the Constructionist Approach to the Dynamic Model can add to understanding differences in grammatical patterns, will benefit from such an attempt. There is an undertone of a line of development towards mature native-speaker-like behaviour implicit in the analyses offered by Hoffmann, where innovations or deviations in other varieties are side-lined from the scope of ‘scientific practice’ with testable hypotheses, which in turn would run counter to the spirit of the functional equality of the Outer and Inner Circle proposed by Kachru (1985 [Reference Kachru and Webster2015]). Of course, such equality should not be a matter of belief or attitude, but an empirical matter: whether the development along the phases necessarily converges on a native-like pattern as end-point (assuming implicitly that this is ultimately a psycholinguistic learning problem, extrapolated to societal level), or whether a different line of development is possible with outcomes that are not just points on a journey towards the same destination, but a trajectory to a different destination. Disentangling the psycholinguistic and sociolinguistic dimensions and incorporating matters of acceptability alongside entrenchment are crucial to that endeavour and represent the next challenge in the development of the insights that Hoffmann presents in this publication.