This autumn, De Gruyter’s “Comparative Handbooks of Linguistics” was enriched by another voluminous work: “Non-verbal predication in the world’s languages”, in two volumes with over 1200 pages, edited by Pier Marco Bertinetto, Luca Ciucci, and Denis Creissels.
I was particularly interested in this volume because I wrote a paper on a similar topic a while ago that was published earlier this year (“Nonverbal clause constructions”, in Language and Linguistics Compass 2025). The introductory paper to this volume (“Non-verbal predication: An analytical framework”, by Creissels et al.) is similar to my paper in that it provides a list (with discussion and examples) of the main types of non-verbal clause constructions, and the main purpose is to offer a framework of comparative concepts.
(They call it an “analytical framework” in the title, and a “theoretical framework” in the abstract; and in the concluding section, they say that they offer “a consistent set of descriptive concepts”. But if description/analysis is what linguists do when they deal with particular languages (p-linguistics), then their framework is not “descriptive/analytical”, but comparative. And if a theory is an explanatory account from which hypotheses can be derived, their framework is not theoretical, but remains at the methodological level, just like my 2025a paper.)
Now while there is of course a lot of agreement and overlap between Haspelmath (2025a) and Creissels et al. (2026), there are also some differences, and in this post, I would like to focus on different ways of treating “adjectival predication”, or “attributional clauses” (as they are called in my paper). While Creissels et al. treat “adjective” as a kind of construction-strategy, I think it’s better to treat them as a construction-function (see Haspelmath 2025b for this distinction).
“Adjectives” are widely known among typologists to be “less universal than nouns and verbs” (e.g. Panagiotidis 2015: 41; see also Evans & Levinson 2009: 435, among many others), and it has been a perennial question how they should be treated in language description and in typology. How should we deal with predicative property-concept words as in (1)-(3), which are unlike English in that there is no copula in an attributional clause such as ‘this flower is red’.
(1) Russian:
Ètot cvetok krasnyj. (Этот цветок красный.) ‘This flower is red.’
(2) Japanese:
Kono hana wa akai. (この花は赤い) ‘This flower is red.’
(3) Haitian Creole:
Flè sa a wouj. [flower this DEF red] ‘This flower is red.’
Are the words for ‘red’ in these languages “adjectives”? At least Japanese akai and Haitian Creole wouj are verb-like also in that they can combine with a past-tense marker (Japanese aka-kutta [red-PST] ‘was red’, Haitian té-wouj [PST-red] ‘was red’). Are they verbs, or are they adjectives after all?
Creissels et al. (2026: 10) define adjectives as roots that can occur adnominally “without additional devices”:
“Our proposal is that the notion of adjective is relevant for the languages having a class of underived lexemes characterized by the possibility of fulfilling the function of adnominal modifier without the additional devices that are necessary for underived lexemes belonging to other morphosyntactic classes to fulfill the same function. Such lexemes can be designated as primary adjectives.”
This would seem to mean that Russian krasnyj, Japanese akai and Haitian wouj are adjectives, because they can be used adnominally without relativizers (such as participial affixes or relative pronouns): ‘red flower’ is krasnyj cvetok, akai hana, and flè wouj, respectively.
But does the definition work in general? It seems that it fails to exclude nouns which can be used adnominally, and this is not so uncommon. In a sample of 72 languages worldwide that have noun juxtaposition of some kind, Mizuno (2024) found 24 languages that use juxtaposition for adnominal possession (e.g. Nivkh ətək χaj [father pigeon] ‘father’s pigeon’) – in these languages, nouns can “fulfill the function of adnominal modifier without additional devices”. Moreover, some languages do not even need special “devices” for relative clauses; for example, Japanese says tobu dōbutsu [fly animal] ‘animal that flies’, without a relativizer. Does this make tobu ‘fly’ an “adjective” in Japanese?
(In the definition, there is a strange formulation “without the additional devices that are necessary for underived lexemes belonging to other morphosyntactic classes” – but which are the “other classes”? It seems that we can know this only if we already know what an adjective is.)
Creissels et al. refer to the most recent handbook article about adjectives, Beck (2023), who defines “adjective” as follows:
“The term ‘adjective’ will be used to describe a lexical–syntactic class of word that contains primarily expressions of property concepts and whose syntactic distribution can be distinguished from that of other word classes in that it shows a specialization for adnominal modification.” (Beck 2023: 366)
Here the term “specialization” may refer to the absence of a “special adnominal device” (i.e. an absence of a relativizer, as in Russian, Japanese and Haitian Creole), but it could perhaps also refer to special devices that do not occur with adnominal nouns or verbs.
Perhaps the gender-agreement suffixes in Indo-European languages can be regarded as showing such “specialization”? For example, Modern Greek has mikr-ó spiti ‘small house’, but mikr-í póli ‘small town’, because spíti is neuter and póli is feminine.
But what about, say, German, which has very similar gender suffixes (klein-es Haus ‘small house’, klein-e Stadt ‘small town’), but in addition has agreeing participles (e.g. spiel-end-es Kind ‘playing child’, lach-end-e Frau ‘laughing woman’)? Should we say that in German, property-concept words do not show a specialization for adnominal modification, because verbs can be used in this way, too?
We see that it is not easy at all to define the term “adjective” formally in such a way that the criteria can be applied to all languages. For this reason, I have proposed that adjective (as well as verb and noun) are defined notionally for comparative purposes (see this earlier blogpost, and my 2023 paper on word classes). As I said above: “adjectival predication” is best defined in functional terms, as predication of a property concept, not as a formally defined type of strategy (I use the term “attributional clause”, adopted from Dixon 2010).
Admittedly, this goes against the most common view in the literature, which is reflected in statements such as the following:
“When a language has no adjectives or only a small closed class of adjectives, property concepts are encoded either as verbs, as in Chinese and Korean, or nouns, as in Quechua and Hausa (see Stassen 1997 for an overview).” (Nikolaeva & Spencer 2020: 7)
It is an extremely widespread view among linguists that some languages have a special adjective class, while others do not. However, if “adjective” is defined notionally, then all languages have such a class, because all languages can express notions such as ‘big’, ‘small’, ‘old’, ‘good’ by simple roots. Creissels et al. (2026: 10) also take this widespread view:
“Recognizing adjectival predication as a particular type of non-verbal predication makes only sense for languages with a class of adjectives whose predicative behavior is distinct from that of verbs.”
In this sentence, they say that the “predicative behaviour” should be “distinct”, in contrast to the earlier characterization of “adjectives” with respect to their adnominal function. And what exactly is “distinct”? Verb-like adjectives often lack the full range of tense-aspect forms that can occur with action verbs, but the occurrence of tense-aspect forms typically depends on actionality characteristic (for example, progressives do not occur with stative verbs such as ‘know’ – does this make them “adjectives”?).
Thus, according to the view that I presented in Haspelmath (2025a), a clause such as (4a) (where wouj ‘red’ occurs with the past-tense prefix té-, like verbs), is attributional, and counts as non-verbal, in contrast with the verbal clause in (4b).
(4) Haitian Creole:
(a) Flè sa a té-wouj. [flower this DEF PST-red] ‘This flower was red.’
(b) Chen sa a té-ale lakay. [dog this DEF PST-go home] ‘This dog had gone home.’
The term nonverbal clause construction should therefore be interpreted in functional terms: It refers to clause constructions in which there is no action or process root (= no verb, in the comparative-concept sense). From a comparative sense, a copula or a ‘have’ verb in predpossessive constructions is not a “verb”, though for language-particular description, such elements are of course often labeled in this way.
Creissels et al. (2026), by contrast, want to exclude clauses with property-concept roots that are are too similar to action and process roots, though they are not very specific (would they even exclude Japanese, where property-concept words such as akai ‘red’ are fairly verb-like?). They also want to exclude clauses with ‘have’ expressions that are called “verbs” in the language, but again, they are not very specific (how verb-like does the ‘have’ word have to be? is English have sufficiently verb-like, even though it lacks an imperative and a passive construction?). It may be that they are doing this because they are not fully aware of the distinction between construction-functions and construction-strategies (Haspelmath 2025b): Their notion of “nonverbal predication” is a kind of mixture of the two, and their comparative concepts are not defined in a way that can be applied straightforwardly to all languages.
References
Beck, David. 2023. Adjectives. In van Lier, Eva (ed.), The Oxford handbook of word classes, 365–382. Oxford: Oxford University Press.
Bertinetto, Pier Marco & Ciucci, Luca & Creissels, Denis (eds.). 2026. Non-verbal predication in the world’s languages: A typological survey (Volumes 1-2). Berlin: De Gruyter Mouton. (https://www.degruyterbrill.com/document/doi/10.1515/9783110730982-001/html)
Creissels, Denis & Bertinetto, Pier Marco & Ciucci, Luca. 2026. Non-verbal predication: An analytical framework. In Bertinetto, Pier Marco & Ciucci, Luca & Creissels, Denis (eds.), Non-verbal predication in the world’s languages: A typological survey (Volume 1), 1–56. Berlin: De Gruyter Mouton. (https://www.degruyterbrill.com/document/doi/10.1515/9783110730982-001/html)
Dixon, R. M. W. 2010. Basic linguistic theory: Volume 1. Oxford: Oxford University Press.
Evans, Nicholas & Levinson, Stephen C. 2009. The myth of language universals: Language diversity and its importance for cognitive science. Behavioral and Brain Sciences 32(5). 429–448. (doi:10.1017/S0140525X0999094X)
Haspelmath, Martin. 2025a. Nonverbal clause constructions. Language and Linguistics Compass 19(2). e70007. (doi:10.1111/lnc3.70007)
Haspelmath, Martin. 2025b. Construction-functions and construction-strategies. In Däbritz, Chris Lasse & Budzisch, Josefine & Basile, Rodolfo (eds.), Locative and existential predication: On forms, functions and neighboring domains, 9–44. Berlin: Language Science Press. (doi: 10.5281/zenodo.16838056)
Mizuno, Shogo. 2024. Noun juxtaposition for predication, possession, and conjunction: Beyond ambiguity avoidance. Linguistic Typology at the Crossroads 4(2). 220–272. (doi:10.6092/issn.2785-0943/18271)
Nikolaeva, Irina & Spencer, Andrew. 2020. Mixed categories: The morphosyntax of noun modification. Cambridge: Cambridge University Press.
Panagiotidis, Phoevos. 2015. Categorial features: A generative theory of word class categories. Cambridge: Cambridge University Press.
Stassen, Leon. 1997. Intransitive predication. Oxford: Oxford University Press.
OpenEdition suggests that you cite this post as follows:
Martin Haspelmath (October 14, 2025). When is “adjectival predication“ non-verbal? Some comments on Creissels et al. (2026). Diversity Linguistics Comment. Retrieved November 10, 2025 from https://doi.org/10.58079/14y8j