Not all languages embed speech the same way—and that changes how we think about language evolution
A new cross-linguistic study upends the assumption that all languages use the same grammar structures to report what others said. Researchers found that complementation—a core feature linguists thought was universal—is actually absent or rare in some languages, which challenges fundamental theories about how human language develops and evolves.
Originaltitel: Is complementation a universal strategy? A cross-linguistic corpus study
This article examines the question of whether complementation structures are cross-linguistically universal by using two different cross-linguistic corpora, each drawing on the same thirteen languages, spanning every continent. One is SCOPIC, the Social Cognition Parallax Interview Corpus, specifically designed to elicit material rich in grammatical categories relevant to social cognition; for each language in our sample this was balanced by a “general corpus” of roughly the same size with no specific targeting of domains. We find that, while complementation is widespread, it is not universal within the languages in our sample: in some it is absent entirely and in others it is extremely rare. Of the structural alternatives used to achieve the same functional goal by far the commonest is quoted speech, suggesting that in the evolution of linguistic structures it is heteroglossia, the embedding of one person’s words in another’s, that is a more basic phenomenon, from which complementation structures then evolve in many but not all languages.