Abstract Ochsenbauer, Hickmann & Hendriks
Space in language has been the locus of much controversy in recent research across several disciplines among the cognitive sciences. On the one hand, spatial knowledge constitutes one of the most basic domains of cognition, presumed by most psychologists to be universal and determined by our biological heritage. On the other hand, it has been shown that spatial systems across languages present a considerable degree of variation (reference systems, markers available and underlying semantic distinctions, lexicalisation and grammaticalisation patterns). This paradox has led to new research concerning universal vs. language-specific determinants of first language acquisition, and more generally it has revived an old debate concerning the relation between language and cognition.
In this context, this paper illustrates a set of findings concerning the expression of motion in child language. The results stem from an ongoing cross-linguistic project focusing on the acquisition of spatial language across a wide age range (children of 2 to 11 years, as compared to control groups of adults) and in three languages (English, French, German). These languages differ in several ways, including in their spatial systems. With respect to motion, they belong to two distinct families (Talmy, 2000), that are either satellite-framed (Germanic) or verb-framed (Romance). SF-languages typically express manner and/or cause in verb roots and path in verbal satellites (to run up, down, across, into...: to roll something up, down, across, into…). In contrast, VF-languages express path in the main verb, manner (if at all) in peripheral devices, and cause in special causative constructions (monter, descendre, traverser… en courant; faire monter, faire rouler…).
One of our aims is to examine the implications of such typological differences for first language acquisition. A first set of findings stems from longitudinal analyses of early spontaneous utterances about motion in English vs. French from 18 months to 3 years of age. A second set focuses on descriptions of motion events elicited from 3 years on in experimental situations. Two experiments examine descriptions of caused motion shown in actions (performed by the experimenter) that consisted of displacing entities from one location to another (putting into, onto…) or in animated cartoons showing an agent doing something (pushing, pulling) to cause the displacements of objects in different manners (roll, slide) and along different paths (up, down, across, into). A third experiment elicited descriptions of voluntary motion events with cartoons showing characters’ displacements that varied in terms of manner (running, walking, swimming…) and path (up, down, across).
In all studies the results show 1) striking cross-linguistic differences at all ages, indicating the impact of language-specific properties, as well as 2) similar developmental progressions, indicating the impact of general cognitive determinants. The discussion further explores the implications of cross-linguistic differences for further research testing the more general impact of typological constraints on cognitive functioning.