The Grammaticon is an evolving database of grammatical comparative concepts with proposed definitions and links to typological features from typological feature collections such as WALS (World atlas of language structures, Haspelmath et al. 2005), Grambank (Skirgård et al. 2023), and APiCS (Atlas of Pidgin and Creole language structures, Michaelis et al. 2013).
It was inspired by the Concepticon (List et al. 2016), a database of lexical comparative concepts, and it is hoped that it will facilitate comparison of typological datasets in the future. Such comparison is often still limited by variable usage of grammatical terminology and lack of clarity of definitions. The Grammaticon provides not only links between grammatical terms and typological features, but (for many of the features) also discussion of how the grammatical terms relate to each other.
When the definition of a grammatical term contains other technical terms, these are preceded by a symbol (°) in the definition and listed in the box on the right-hand side of the detail page for the term (“defining concepts”). This box also lists the “derived terms”, as well as related typological features.
Definitions which are preliminary are enclosed in double square brackets ([[…]]).
Not at all – it primarily aims to provide clear and sharp definitions of a wide range of terms that are already widely used as comparative concepts by linguists. The set of grammatical comparative-concept terms that one might need is completely open-ended, just like the set of lexical concepts that one might need for lexical comparison is open-ended. Linguists create new concepts all the time, any maybe it would be helpful if they created new terms as well (in order to avoid proliferation of polysemy).
Yes, there is a recently published database of comparative concepts by William Croft. The Grammaticon systematically links to these definitions, as well as to Wikipedia and SIL’s Glossary of Linguistic terms. The Grammaticon offers one perspective, without claiming that it is the only one.
It is true that the reality is often continuous, but scientists must impose sharp boundaries in order to capture this continuity quantitatively. Without sharp concepts as units of measurement, precise quantification is not possible, and all comparison remains vague. Some scientific concepts may not be amenable to sharp definitions, but technical terms of grammar should ultimately be subject to quantitative evaluation and should therefore be defined with clear boundaries.
Linguists typically assume implicitly that their technical terms (including even “word”) have a clear meaning, even though they often fail to provide definitions. But it is only when one provides an explicit definition that one can see problems with approaches that leave definitions implicit. It appears that one of the reasons why linguists are often reluctant to provide definitions is that they think of the terms as referring to natural-kinds concepts (Haspelmath 2018), i. e. entities that exist in nature to be discovered or fully elucidated by linguists. It may be that “words” and “affixes” exist in nature, but as long as we do not know whether they do, it is best to use these terms as comparative concepts with clear definitions.
Each language has its own unique categories (Haspelmath 2020), so descriptive work must normally work with language-particular categories, defined in terms of language-particular notions. By contrast, the Grammaticon’s terms are comparative concepts, which are defined in the same way for all languages (Haspelmath 2018). Thus, they are not directly suitable for language description/analysis, and many of the definitions come from typological work. However, language description can often be inspired by the general concepts defined by typologists, and language-particular descriptions are more transparent if they use similar terms.
Most of them are retro-definitions, i. e. definitions of well-known terms that have been provided retroactively (Haspelmath 2021). Most well-known terms have spread among linguists via salient exemplars, rather than via definitions, and the Grammaticon aims to provide clear (retro-)definitions for them. However, the Grammaticon also contains a number of neologisms which seem helpful, such as “numerative”, “hyparctic clause”, and “duonominal clause”.