This project is funded by the
Idex Emergence en Recherche scheme.

We aim to understand, through the development of a new toolset, what factors influence the organization of descriptive grammars, and the effects of these choices on linguistic typology and theory. The toolset consists of a relational database connecting the full tables of contents of a set of 250 grammatical descriptions from 1850 until the present and metadata on both the grammars and their authors.

Since Boas, the assumption that most contemporary linguists hold is that the structure of a grammar of a language is exclusively determined by its inherent features and characteristics, and that a language should be described “on its own terms”. Yet if it is the nature of a language alone that determines the structure of its grammar, grammatical descriptions would be expected to show similarities according to phylogenetic groupings, rather than according to external factors. There are a number of factors, apart from inherent characteristics of the language being described, that can affect the shape of the resulting grammar: these may be variables related to the description itself (such as date, size, number of chapters, metalanguage, language area) or to the grammarian (such as native language, place of linguistic training, supervisor, past and present institutional affiliations, theoretical leanings).

The way a language is presented through a descriptive grammar can significantly influence linguists' and typologists' perceptions of the distribution of linguistic features across the languages of the world. This raises the important issue of how grammaticographical decisions may impact the contribution to typological theory of data from endangered languages, whose description is often limited to a single grammar, frequently written early on in the career of a fieldworker.

Research into contemporary grammar writing ranges from the descriptive to the prescriptive, but is always based on impressionistic reviews of how grammars are assembled and organized. The two topics that have been discussed in the literature concerning the structural organization of grammars are, on the one hand, whether they feature ascending (from smaller to larger units of speech) or descending order, and on the other, whether the principles of presentation are onomasiological (function to form) or semasiological (form to function).

Apart from a few constant features (phonology tends to be described at the beginning, glossary and texts tend to be provided at the end), modern descriptive grammars are highly heterogeneous in their structures. Recent work (Kelly and Lahaussois, in press) has established a few of the observed organizational schemes of descriptive grammars, based on a corpus of 30 grammars of Himalayan languages: these include parts-of-speech prominent grammars, nominal-verbal prominent grammars, linguistic subfield prominent grammars (with sections for phonology, morphology, syntax, pragmatics), areal topic prominent grammars (with an organization highlighting topics of areal interest), but many more types exist, revealing evidence of various influences (cultural, linguistic, training-based, etc.) on grammarians. It must be noted that, even within a certain organizational scheme, there can be considerable variability: part-of-speech prominent grammars, a widespread model for missionary grammars (Zwartjes 2011), may not include the same word classes from one grammar to another, or may treat them in a different order. Apart from the preliminary work by Kelly and Lahaussois, based on selected Himalayan grammars, there are no established taxonomies of grammars based on the organization of their tables of contents.

Historians of linguistics, however, routinely compare grammars of certain periods and traditions, looking at the treatment of the parts of speech and other basic terminology, as well as at how grammars are structured (Colombat and Lahaussois 2019). This project therefore aims to bring to contemporary grammaticography methods developed for the history of linguistics, and to bridge a gap between historians of linguistics and descriptive linguists.

This proof-of-concept project will be built on a prototype corpus of 250 grammars (of Ethiopian, Himalayan, Gaelic, Australian, Algonquian languages) produced from 1850 to now. The choice of language areas is based on: a) interests and expertise of team members; b) wide geographical coverage and genetic diversity; c) their distinct descriptive traditions, including, for some of these languages, a cohabitation of local (säwasǝw and Semitic for Ethiopia; Paninian and Tibetan for the Himalayas) and imported grammaticographical.

The power of our toolset is that 1) it will provide a new, empirically-based methodology which, once developed and proven successful, can be extended to any number of descriptive grammars from around the world; 2) it will make it possible to establish a taxonomy of organizational schemes for descriptive grammars, something that has not yet been done in linguistics ; 3) it will make it possible to connect these types to a number of variables in order to establish what factors influence the shape of the produced grammatical description ; 4) it will make it possible to establish the genealogies of grammatical descriptions, tracing them to one or multiple source model(s). An additional benefit of the project is that studying the structure of grammatical descriptions is an effective pedagogical tool for teaching students (some of whom are future grammaticographers) about descriptive grammars, their organization, and terminological variation.

Jump to top of page