Spontaneous spoken English: An integrated approach to the emergent grammar of speech by Alexander Haselow (review)

In his monograph, Alexander Haselow outlines a model for the grammatical description of spontaneous spoken English as it emerges in the flow of speech. He thereby focuses on linguistic expressions that are traditionally left aside in grammatical work, namely so-called unintegrated or extraclausal expressions like discourse markers, parentheticals, general extenders, or final particles (Ch. 4), and fragmented syntactic phenomena, such as minimal structures or ellipses, segment-chaining structures, and far-reaching projections (Ch. 5). According to the author, these linguistic structures are typical of spontaneous speech and ‘escape the syntactic formats predicted in sentence-based models of grammar’ (80). This thus calls for an alternative. Indeed, the main motivation for writing the book is given right from the beginning, namely: ‘[t]he search for an alternative to established structuralist and generativist approaches, both of which see language structures as relatively fixed prior to verbalization and as conforming to abstract, allegedly “descriptive” rules rather than as being schematic and inherently open’ (2). I come back to this point later.

The main hypothesis of the work is that of a dualistic organization of grammar. H defends the idea that grammar is a two-component system, where one component deals ‘with the internal organization of emergent syntactic units in terms of linearization and hierarchization, the other one with the organization of language based on cognitive, discourse-structural and interactive principles of language use’ (30). This dual organization is referred to as the grammatical dualism assumption, meant to account for linguistic structures of both microgrammar and macrogrammar. The former accounts for linguistic expressions that are hierarchically integrated (relations of dependency and embedding), while the latter concerns the linearization of linguistic segments that are relatively autonomous from a morphosyntactic and semantic point of view. This reference to micro- and macrogrammar relates H’s descriptive work to the notions of microsyntax and macrosyntax as they are used in the description of (spoken) French syntax (in particular, Berrendonner 1990, 2002 and Blanche-Benveniste 2003, cited in the volume, but see also Avanzi 2007 and Corminboeuf & Benzitoun 2014, among others). The author acknowledges that his plea for a dualistic approach to grammar is also present in work by Kaltenböck and colleagues distinguishing sentence grammar from so-called thetical grammar (Kaltenböck, Heine, & Kuteva 2011) or in the functional grammar approach by Dik (1997), pursued today in Hengeveld and Mackenzie’s functional discourse grammar (2008).

H’s endeavor differs from these previous works by providing neurolinguistic evidence to support his dualistic approach. He does this convincingly in Ch. 6 by referring to a select number of neurolinguistic studies that demonstrate that the forms, structures, and functions which have been identified by H as belonging to microgrammar or macrogrammar are indeed processed differentially. More specifically, it seems that the left brain hemisphere takes care of structures and forms involving morphosyntactic and semantic relations, as well as embedding, belonging to micro-grammar, while the right hemisphere is in charge of discourse organization, speaker-addressee relationships, and formulaic expressions outside hierarchical structures (286). In other words, not only is grammar dualistic, but so is human cognition. Thus, H wants to propose a theory of grammar that is descriptively adequate for spoken (spontaneous) data, which would be at the same time a cognitively plausible model of language. Only then, he argues, will we be able to explain the corpus data as we find them in spontaneous interactive discourse. In his view, this requires revision of the—in his eyes—traditional view of a product-based grammar, and replacement by a process-based one. This is needed in order to do full justice to the structural properties of spoken language as they are produced under cognitive constraints, including mainly ‘limitations in working memory capacity and the quasi-simultaneousness of speech planning and production’ (3). In line with Hopper’s (1998) emergent grammar, H sees grammatical structure-building ‘as an incremental process, based...