Geocentric spatial concepts in Mandarin: comparing empirical evidence with theoretical models

The goal of this paper is to analyse geocentric spatial concepts in Mandarin, i.e. spatial relations defined via reference to salient locations (“anchors”) in the geographic environment. We connect this analysis to the SocioTopographic Model, which introduces a complex taxonomy of spatial relation types. We show that Mandarin expresses cross-linguistically well-documented geocentric relations (e.g. Béi ‘North’, Dōng ‘East’), but also language-specific concepts (e.g. yán-hǎi lit. ‘around-sea’ for ‘seaside’). We then show that in previous works on these relations, their role in Mandarin has received little analysis. To ameliorate this situation, we provide a mixed-methods study. First, we verified the existence of possible words expressing geocentric relations via a dictionary analysis of compound forms potentially expressing these relations. Second, we retrieved sentences including these compound forms to verify their grammatical (i.e. morpho-syntactic) properties and distribution, and their use in conversation. Third, we administered an exploratory online questionnaire to native speakers (N = 77; average age = 35.0 years) to assess whether these compound forms can introduce unique anchors in context. The central findings of our mixed-methods study are as follows. First, compound forms can express a wide array of geocentric relations, though the number of forms potentially expressing spatial relations is considerably larger (116/418). Second, compound forms expressing geocentric relations can occur in any sentential position, and regularly introduce reference to anchors. Third, the responses from the questionnaire provided preliminary support for compound forms referring to unique anchors in context. We discuss how these findings are consistent with the language-general predictions of the SocioTopographic Model. We also highlight the language-specific nature of our findings, and how these findings individuate theoretical gaps in the model. We conclude the paper by sketching possible avenues of future research on geocentric spatial concepts, and possible contributions of our study to the SocioTopographic Model.