To be published in the Proceedings of the International Conference on Cognitive Science, Allahabad University, Dec. 16-18, 2004. A cross-cultural comparison of spatial language and encoding in Bali and Geneva Pierre R. Dasen (pierre.dasen@pse.unige.ch) FPSE, University of Geneva CH-1211 Geneva 4, Switzerland Jürg Wassmann (juerg.wassmann@urz.uni-Heidelberg.de) Institut für Ethnologie, Universität Heidelberg, D-69117 Heidelberg, Germany Abstract The paper presents some comparative aspects of the ongoing study, in particular some results from a replication study in Bali, in which a strong impact of socio-economic and cultural variables (that can be subsumed under the concept of acculturation) was found. Results from a study in Geneva, with monolingual and bilingual children, demonstrate the consistent use of egocentric language and encoding in French and other European languages. Our first study in Bali carried out in 1994 (Wassmann & Dasen, 1998) and subsequent research in India and Nepal (Mishra, Dasen & Niraula, 2003), have been reviewed briefly in the introduction to this symposium (Mishra & Dasen, this volume). Overall, we found a strong tendency towards the use of a geocentric frame of spatial reference (FoR) in all of our samples, taken from societies in which the language allows a choice between all three frames, but in which a geocentric orientation system is generally used (NSEW in India and Nepal, and also Up-down in the latter). Large differences between encoding tasks showed that the choice of a FoR is strongly task dependent, and differences between rural and urban groups showed the influence of ecological settings (Mishra & Dasen, in press). While differences between schooled and unschooled children were also explored (Dasen, Mishra & Niraula, 2004), we did not take socio-economic indicators or other family background variables into account. We also did not find the same results in India and Nepal as in Bali in terms of age trends: while absolute encoding seemed to decrease with age in Bali, it increased in all three samples in India and Nepal. One of the first endeavors in the current project was therefore to replicate the study in Bali itself. Bali 2002, a replication study One of the limitations of our earlier study in Bali was the small sample size. This was due to the fact that we carried out the research in a fairly remote area of Bali, in the small village of Bunutan, and we did not have access to the local school which would have provided us with a larger sample of children. In 2002, we were able to work with much larger samples, in schools of two different locations, as well as testing again a small sample of young children in the Bunutan area. The number of children tested is given in Table 1. The city sample was obtained in a private, experimental school attached to the only teachers training college of Bali, in the city of Singaraja, on the North coast of Bali. The pupils attending this school were mainly from middle-class families. Table 1 : Sample characteristics of studies in Bali, 2002, and 1994. Age groups 4-5 6-8 9-11 12 Total Singaraja (city) 15 24 28 5 72 Sambangan (village) 16 35 37 10 98 Bunutan (remote village) 15 18 33 Total 46 77 65 15 203 Bunutan 1994 9 8 7 14 38 The village of Sambangan was situated in a rural area about 15 km from Singaraja, on a north facing slope of the mountain, with a clear view of the sea in the distance. The parents of the children were mainly farmers, except for the 4 and 5 year olds who attended a pre-school in the same locality but closer to town. Information on the family background of each child was obtained through interviews with teachers, or a questionnaire sent home to the parents. For the background variables of mother and