Spatial communication and computational efficiency across languages and cultures

There are few more compelling questions in the cognitive sciences than whether people who speak different languages communicate about the world in the same way. Speakers of different languages employ the same perceptual apparatus, so one might expect that the world’s 7,200 or so living languages may have evolved to share common properties. Yet the idea that communication systems are universal has been challenged with studies documenting extensive cross-linguistic variation in domains closely yoked to perception. Spatial communication is a critical test case; as Evans and Levinson (2009) note in the Myth of language universals, “spatial cognition is fundamental to any animal, and therefore if Fodor is right anywhere [that languages directly encode the categories we think in], it should be here” (12, p.436). However, arguably the most important spatial terms across languages, spatial demonstratives (e.g. ‘this’, ‘that’, ‘here’, ‘there’ in English), have not sufficiently featured in this debate.
Spatial demonstratives occur in all languages, are among the oldest word forms in language evolution, and are also produced among the first words in language acquisition by children across languages. Yet it has been suggested that these words maybe used in fundamentally different ways across languages (e.g. Levinson et al., 2018), with parameters including the position of an addressee, elevation, object visibility and the eye gaze of addressee all hypothesised to affect demonstrative choice in different languages. They therefore represent an essential word category with which to understand fundamental relationships between language, cognition and environment.
In two complementary Work Packages led by a UK-USA-German collaborative team, we will:
1) Use cutting edge experimental methods to examine the impact of a wide range of parameters on demonstrative choice across diverse languages to establish commonalities and differences in spatial communication systems both between and within languages (UK-led WP1).
2) Examine the relationship between language use (between and within languages) and the nature of the environment in which speakers live (UK-led WP1).
3) Develop a systematic information-theoretic approach to spatial words, combining the cross-linguistic experimental data with computational modelling (USA-led WP2).
The programme of work represents a timely and critical step change in understanding whether speakers of different languages fundamentally carve up the world in different ways, made possible as a result of recent innovations in methods and networks of collaborations across diverse languages of the required scale. To achieve these goals, we have assembled the optimal international team, with complementary expertise in cutting edge experimental cross-linguistic methods and analyses (Coventry, Gudde, UK), and computational modelling (Mahowald, Futrell, USA), further augmented with expertise in linguistic typologies and linguistic field work (Diessel, Forker, Germany).
Outputs will be high impact scientific journal articles in outlets where the PIs have a track record of publication (Nature Human Behaviour; PNAS) and conference presentations (including a conference on spatial communication and diversity to be hosted by the research team), targeting diverse audiences in cognitive psychology, linguistics, psycholinguistics, perception, neuroscience and philosophy. Moreover, the combination of experimental cross-linguistic methods and modelling using the Information Bottleneck approach will significantly advance the state-of-the-art in computational modelling and cross-linguistic research methods, offering a new gold standard in the cognitive sciences. The PIs will also engage a wide range of non-academic stakeholders, using the topic to enrich understanding of diversity between and within languages across all age groups and to encourage learning new languages.