Last year (in 2020), just before the pandemic caused all kinds of things to close down, I was lucky enough to travel to Vancouver to participate in the West Coast Conference on Formal Linguistics (WCCFL). The paper I presented there has now been published in the proceedings of WCCFL 38 so I’ll write a little about it.
It’s the first publication coming out of my project on case and agreement in ditransitives and it sets out the background and the idea on ten brief pages. The paper deals with an observation about combinations of case and agreement alignment in languages in which the verb agrees with the subject and one of its objects (but not both). The observation is the following:
- Among languages with one instance of object agreement, no language with secundative or neutral case-marking alignment allows object agreement exclusively with the theme argument in a ditransitive construction (indirective agreement alignment).
What this means is that in languages which have a so-called double object construction such as the English sentence in (1), if they also have object agreement with one object, agreement with the recipient (Peter) must be possible.
(1) Mary gave Peter the flowers.
Most of these languages do not have an alternative, that is, they only allow agreement with the recipient and not the theme (the flowers). The observation also covers some other types of ditransitive constructions: some languages mark the recipient with regular object case (accusative or absolutive) and the theme with an oblique case, a little bit like in the English sentence in (2).
(2) Mary provided him with the flowers.
Again, in languages with object agreement which have these constructions, the verb must be able to agree with the recipient.
In my WCCFL paper I test this observation on a relatively large sample of 99 languages from all over the world and I show that, indeed, the observation mentioned above seems to hold. In languages with ditransitive constructions of the types (1) and (2), no language allows object agreement exclusively with the theme. If the theme can control object agreement in such languages, so can the recipient.
Which of the two objects controls agreement in such cases is an interesting question and is subject to variation. In the small group of the Chukotko-Kamchatkan languages, for example, one finds Chukchi, Alutor and Koryak, in which the person of theme and recipient determine which of the two controls agreement, as well as Itelmen, for which Jonathan Bobaljik and Susi Wurmbrand have argued that the more topical object controls agreement. Another possibility is that different ditransitive predicates, such as give or send, show distinct behaviour.
I show how these alternations can be accounted for using Amy Rose Deal’s idea of agreement in terms of interaction and satisfaction features. More generally, I show how the observation about possible alignment types follows from assumptions about
- the structure of ditransitives (recipients are more prominent than themes),
- recipients being the first target of the agreeing verb, and
- the interaction of morphological case and agreement, namely that in languages with object agreement, accusative and absolutive objects can always control object agreement, but dative and other oblique cases can only do so in some languages.
Taken together, these assumptions predict what seems to be happening in the 99 languages I looked at for the paper. For reasons of space, the 10-page WCCFL paper doesn’t have all the relevant information but I’m supplying that in a separate repository on OSF.
Since submitting and finishing the paper, my sample of languages has grown to around 130 languages from 99 genera and the generalisation still holds — but I’ll write about that in a future post!
Bárány, András. 2021. A typological gap in ditransitive constructions. In Rachel Soo, Una Y. Chow & Sander Nederveen (eds.), Proceedings of the 38th West Coast Conference on Formal Linguistics, 43–53. Somerville, MA: Cascadilla Proceedings Project. Download the paper here.
Bárány, András. 2020. Supplementary materials for A typological gap in ditransitive constructions (Proceedings of WCCFL 37/38). https://osf.io/fr4b5/