CE27 - Culture, créations, patrimoine

Spatial Asymmetries across Languages: a Typological Approach – SALTA

Spatial asymmetries across languages

The aim of this project is to investigate the conceptual domain of motion in space, with a focus on structural and functional asymmetries in the expression of path of motion, especially asymmetries linked to dynamic deixis, boundary-crossing and verticality, with a comparative study of 30 typologically and genealogically varied languages.

Typological study of spatial asymmetries: beyond the state of the art

The expression of spatial relations has been widely investigated over the last forty years. This constitutes a vast and diverse body of literature, the focus of which has shifted from localism to spatial deixis, frames of reference, “topological” relations and, more recently, motion event descriptions. In the last decades, influential scholars in the domain have addressed this issue from different angles, from the empirical approach adopted by Slobin to the functional stance taken by Vandeloise, or the theoretical approach adopted by Talmy, each approach having its own merits. One domain in particular has received a great deal of attention lately, namely spatial asymmetries in language and cognition, with a specific focus on the cognitive and linguistic bias toward the Goal of motion, as opposed to the Source. The most striking spatial asymmetries are Source-Goal Asymmetry and Dynamic Deixis. Asymmetries in the subdomain of verticality, an important dimension in human cognition, seem to be intricately linked to these two phenomena.<br />Asymmetries in linguistic descriptions of space were first suggested to be a language universal (Ikegami 1987), with a consistent bias across speakers and languages to express goals rather than sources (Bourdin 1997). More specifically, it has been observed that there is a tendency for goals to be expressed (i) more frequently (Kopecka 2012; Lakusta et al. 2006; Miyajima 1986), (ii) more precisely (with finer semantic distinctions, Fillmore 1992) and/or (iii) more lightly (with morphologically simpler forms, Stolz et al. 2014). <br />The notion of goal bias (Ungerer & Schmid 1996, Dirven & Vespoor 1998, Aurnague 2015) is of special importance for the project. Earlier studies have shown that this is at least a clear tendency, with some counter-examples: asymmetric descriptions of space are crosslinguistically common, not universal (Stefanowitsch & Rohde 2004). The goal of the SALTA project is precisely to investigate these asymmetries, and measure their scope, by looking at language structure and language use, working our way toward a typology of spatial asymmetries, understood in terms of locus, semantic distinctions and degree of grammaticalization of the markers.

In order to meet its objectives, the SALTA project is structured in three workpackages, with specific but interrelated goals.
WP1 “Language structure” is a typological survey of spatial asymmetries in languages, with varying depth and width. It includes European languages, of course, but there is also a clear emphasis on non-Indo-European languages (outside of Europe).
WP2 “Asymmetries in Language Use” proposes an empirical assessment of the effect of these asymmetries on language use, on the basis of tests run on a subset of languages. Indeed, this approach is possible only with data, and is realistic only with magnitudes of data which only big corpora can provide. This automatically restricts the pool of languages we can choose from.
WP3 “Methodology and data” aims to build a visual elicitation tool ensuring the replicability of experiments run both within the SALTA project and beyond. This also includes an adapted and systematic procedure for spontaneous data collection. Both will be the foundation of an open-access cross-linguistic database, which will also link back to the results of WPs 1 & 2.

In the context of the SALTA project, we intend to design a series of methodological tools for data elicitation in different linguistic and cultural field sites. This will include both traditional methods such as questionnaires and more innovative visual methods that can be used both for investigating language-specific characteristics and for cross-linguistic comparison.
Results of these investigations will be presented at major national and international conferences, and form the basis of papers which we plan to submit to international journals; the descriptive work pertaining to individual languages will be presented in a collective volume.
Beyond these publications, one of the main goals of the project is to contribute new data and share both raw data and results with the academic community, both students and faculty. In order to share and make the best of our research efforts and results, our plan is to build a linguistically enriched corpus that includes glossed, annotated and formatted data collected in various languages in the context of this project.

Beyond the SALTA project, the results of our investigation will thus lay the foundations of a renewed theoretical approach of spatial asymmetries based on language structure and language use. This will in turn make it possible to prepare for research projects on other intriguing specificities of the way languages enable us to encode space and motion.

Once published, all scientific productions of the project are uploaded to the HAL website, in open access (https://halshs.archives-ouvertes.fr/).

SALTA investigates the conceptual domain of motion in space, with a focus on structural and functional asymmetries in the expression of path of motion. When describing a motion event, we can choose to encode various portions, from Source via Median to Goal, as in she walked from school (Source) through the forest (Median) to her friend’s house (Goal). In most cases, these three portions of path do not receive equal treatment, with overexpressed Goals, underexpressed Sources, and less explicitly expressed Medians. In this project, we focus on such asymmetries, including those linked to dynamic deixis (come vs go) and verticality (up vs down), through comparative studies of 30 typologically and genealogically varied languages. SALTA will (i) provide a comprehensive corpus of spatial asymmetries taking into account not only language structure but also language use, (ii) assess the role of general tendencies and language-specific patterns in spatial asymmetries, and (iii) build a typology of spatial asymmetries, thus gaining insights into the factors that determine and constrain spatial asymmetries in language. In order to achieve this goal, we will build a crosslinguistic database of spoken and written data, making the most of existing data and complementing it with data gathered with new elicitation tools that will be developed within SALTA. This will yield a balanced compilation of elicited and spontaneous data, made available to the research community, thereby contributing to open-access and reproducible science related to linguistic and cultural diversity.
The languages included in SALTA are genealogically and typologically varied, but they do not constitute a representative sample of language diversity. Rather, we aim for an extensive and fine-grained cross-linguistic investigation carried out by experts of these languages – an important prerequisite, if we wish to build the basis for a typology that will allow for further typologically informed language descriptions, and that can be further informed by data from a larger sample of languages. SALTA’s unique contribution is the fact that it investigates languages that have not benefited until now from intra- and inter-linguistic inquiries into spatial asymmetries, examining spatial asymmetries in both language structure and language use, without the usual focus on European languages.
By adopting a systematic methodological approach and by assessing factors that trigger spatial asymmetries, SALTA will provide key insights both for the field of language sciences, language typology in particular, and for the field of cognitive science. Furthermore, SALTA addresses the question of unity and variation in human language. Because language plays a fundamental role in human culture and cognition, languages have long been a main concern in debates on human diversity and human cognitive capacities. While there is a universal genetic basis for language acquisition, we know that languages vary tremendously across cultures in time and space. This raises important questions concerning the degree of variation, the existence of cognitive and/or functional constraints on variation and the possibility that some (culture-specific) conceptual domains are more likely to show variation than others. Although much effort has been made in the past twenty years to describe under-described and endangered languages, many languages have not yet been thoroughly described. While some languages have benefited from grammatical descriptions, there are many conceptual domains, such as space or more specifically the domain of spatial asymmetries, in which much unknown remains. By investigating languages belonging to different genetic groups and languages spoken in different geographical areas and within different cultures, SALTA will contribute to this research field, uncovering constrains on crosslinguistic variation in this conceptual domain.

Project coordination

Benjamin Fagard (Langues, textes, traitements informatiques, cognition)

The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.


CRLAO Centre de recherches linguistiques sur l'asie orientale
LATTICE Langues, textes, traitements informatiques, cognition

Help of the ANR 432,286 euros
Beginning and duration of the scientific project: March 2021 - 48 Months

Useful links

Explorez notre base de projets financés



ANR makes available its datasets on funded projects, click here to find more.

Sign up for the latest news:
Subscribe to our newsletter