Societies of strangers do not speak less complex languages

Islands are engines of language diversity

Islands have played a prominent role in evolutionary and ecological theory, centring the theoretical framework for understanding biodiversity in terms of isolation and area and providing 'laboratories' of evolutionary change and adaptive radiation. However, a similar role for islands in understanding global language diversity has not been established, even though one-sixth of the world's languages are spoken on islands which account for <1% of the inhabited land area. The striking diversity of island languages remains largely unexplained. We construct a global database which reveals that 10% of the world's languages are endemic to islands (landmasses <11,000 km2) and we test several key theories of language evolution and diversity. We show that language diversity on islands increases with area but does not show a steady decrease with isolation, nor are island languages at elevated risk of loss. However, number of endemic languages per island increases with both area and isolation. We demonstrate that islands shape language evolution, with fewer phonemes (distinct sounds) in island endemic languages with increasing isolation. Our results suggest that islands generate language diversity by accelerating both language change and diversification.

The evolutionary dynamics of how languages signal who does what to whom

Languages vary in how they signal "who does what to whom". Three main strategies to indicate the participant roles of "who" and "whom" are case, verbal indexing, and rigid word order. Languages that disambiguate these roles with case tend to have either verb-final or flexible word order. Most previous studies that found these patterns used limited language samples and overlooked the causal mechanisms that could jointly explain the association between all three features. Here we analyze grammatical data from a Grambank sample of 1705 languages with phylogenetic causal graph methods. Our results corroborate the claims that verb-final word order generally gives rise to case and, strikingly, establish that case tends to lead to the development of flexible word order. The combination of novel statistical methods and the Grambank database provides a model for the rigorous testing of causal claims about the factors that shape patterns of linguistic diversity.

Human languages with greater information density have higher communication speed but lower conversation breadth

Human languages vary widely in how they encode information within circumscribed semantic domains (for example, time, space, colour, human body parts and activities), but little is known about the global structure of semantic information and nothing about its relation to human communication. We first show that across a sample of ~1,000 languages, there is broad variation in how densely languages encode information into words. Second, we show that this language information density is associated with a denser configuration of semantic information. Finally, we trace the relationship between language information density and patterns of communication, showing that informationally denser languages tend towards faster communication but conceptually narrower conversations or expositions within which topics are discussed at greater depth. These results highlight an important source of variation across the human communicative channel, revealing that the structure of language shapes the nature and texture of human engagement, with consequences for human behaviour across levels of society.

Grammars Across Time Analyzed (GATA): a dataset of 52 languages

Grammars Across Time Analyzed (GATA) is a resource capturing two snapshots of the grammatical structure of a diverse range of languages separated in time, aimed at furthering research on historical linguistics, language evolution, and cultural change. GATA comprises grammatical information on 52 diverse languages across all continents, featuring morphological, syntactic, and phonological information based on published grammars of the same language at two different time points. Here we introduce the coding scheme and design features of GATA, and we describe some salient patterns related to language change and the coverage of grammatical descriptions over time.

Languages with more speakers tend to be harder to (machine-)learn

Computational language models (LMs), most notably exemplified by the widespread success of OpenAI's ChatGPT chatbot, show impressive performance on a wide range of linguistic tasks, thus providing cognitive science and linguistics with a computational working model to empirically study different aspects of human language. Here, we use LMs to test the hypothesis that languages with more speakers tend to be easier to learn. In two experiments, we train several LMs-ranging from very simple n-gram models to state-of-the-art deep neural networks-on written cross-linguistic corpus data covering 1293 different languages and statistically estimate learning difficulty. Using a variety of quantitative methods and machine learning techniques to account for phylogenetic relatedness and geographical proximity of languages, we show that there is robust evidence for a relationship between learning difficulty and speaker population size. However, contrary to expectations derived from previous research, our results suggest that languages with more speakers tend to be harder to learn.