Universal Quantification And Distributive Marking In Serbian

This paper discusses experimental evidence bearing on the theoretical analysis of so-called distributive-share markers, such as Serbian po, as opposed to distributive-key markers, such as Serbian svaki or English every. Based on the analysis of distributive po as amarker of event plurality, which forces distributivity, but does not involve universal quantification and therefore atomic and exhaustive distribution, contrary to svaki (every) (Knezevic 2015), we tested the following hypothesis: in Serbian, only sentences with both svaki and po will yield obligatory atomic-distributive and exhaustively distributive readings. A total of 98 children were tested, separated in three age groups, between the ages 4; 3 and 6; 9 (N = 22, MA = 5; 0, SD = 0.8), between the ages 6; 11 and 8; 1 (N = 38, MA = 7; 2, SD = 0.5), and between the ages 8; 7 and 11; 0 (N = 37, MA = 9; 4, SD = 0.7), as well as the control group of Serbian monolingual adults (N = 31, f = 21, m = 10, MA = 27; 3, SD = 8.6). Our results suggest the following. First, that the exhaustivity requirement on universal quantification is acquired before the atomicity requirement. Second, that children acquire the truth conditions of the distributive-sharemarker po prior to those of the universal quantifier svaki (every). On Knezevic's (2015) proposal, this wouldmean that in languages that have both pluractional markers and universal quantifiers, such as Serbian, children acquire pluractionals before universal quantifiers.