A central theme in discussions about education policy internationally concerns literacy, as is evident from the PISA study. In this paper we specifically discuss how scientific literacy is to be construed. The current argument regarding how scientific literacy best can be achieved has been summarized by Roberts (2007) as Visions 1 and 2. Vision 1 argues that knowledge primarily concerns acquiring concepts and propositions that later can be applied generally in various situations. However, it is well established that concepts which are learnt in a specific activity, as for instance in school, do not automatically transfer to other situations (Lobato 2006). Thus Vision 2 argues that learning science should primarily be situated in authentic contexts. But by just using activities that mimic another activity (like science investigations) does not automatically result in general propositions about the principles of that activity, which can be used universally (Moss, 2001). The analysis builds on new, but well established methods with pragmatists orientations developed in French speaking countries and in Sweden, viz. Practical Epistemology Analysis (Wickman & Östman, 2002) and Joint Action Theory (Sensevy et al., 2005). The theoretical frameworks draw among others on the work of Ludwig Wittgenstein (1953) and John Dewey (1929/1958). Both analytical frameworks have strong didactical purposes and are aimed at understanding and improving the meaning making and learning in classrooms. Before technical definitions of scientific concepts can make sense, students first need to be part of activities where further precisions of concepts are needed. Our examples deal with measurements and hence also deals with the interface between science and mathematics. The concepts of weight, volume and area are learnt as part of certain activities where they need to be measured and where distinctions can be made.

In an attempt to contributing to how concepts are generalized, we here present qualitative research results that suggest an understanding of conceptual development that is supplementary to Vision 2 (Östman & Almqvist, pending; Wickman & Ligozat, pending). Together with theoretical underpinnings we present empirical data that consists of video recordings from science educational settings in Switzerland (grade 4) and Sweden (grade 7). In the analyses we compare how the didactic situations in interactions between students and teachers and the sequencing of the didactical situations supports generalizations and subsequent generalization of concepts.

The concepts are not primarily delivered as representations of things as given in the world, but instead as part of activities that create a need for children to make conceptual use more precise. Further scientific precision as supported by the teacher is needed to draw on communication to further socially agreed purposes. We show that if this is not the case, it is difficult for the learner to value which conceptual use is necessary and sufficient. We demonstrate how not just conceptual representations need to be generalized across activities, but also the norms of deciding what counts in different situation as well as the whole point of the activities. The meaning of a scientific concept is not simply what it represents in nature, but also how well it can be used in communication to accomplish certain social purposes that is valued by society and that can be valued by students.