The point of departure of this study is the increasing use and popularity of so-called socioscientific problems in educational practice. For the study of learning, such problems are interesting from theoretical as well as methodological points of view. Such problems can be articulated in terms of the relationship between learning in the context of science education, and what Bakhtin (1986) referred to as “heteroglossia.” This exotic term captures something that is, we argue, not so cryptic. In fact, it is close to our everyday experiences. What the term alludes to is the observation that there are diverse ways of communicating and knowing about objects and events. In other words, and to continue in Bakhtinian parlance, in the complex society there are many “speech genres” that reflect how different social groups or institutions communicate about what they do. If we take a simple object as, let us say, an orange, it can be discussed, analysed, and thought about in many different ways and in many different genres. The satisfied consumer may speak of its delicious taste and its juicyness, the dietician will speak of it in terms of nutritional value and richness in vitamin C, and the artist may attend to it in terms of its colour, shape and texture in the context of what is to be a still life. At more abstract levels, we can think of the importer of oranges, the transport companies shipping oranges from their sites of production to consumers all over the world, and the economist, in her role as advisor to a multinational company, analysing the supply and demand in the market for oranges, as thinking and communicating about oranges in very diverse manners. In the latter cases, the terms and concepts that are productive are very different from those that characterize the consumer enjoying his morning fruit or the shop owner trying to persuade customers to purchase fresh oranges. Thus, the orange as a physical object is embedded in diverse social practices where very different “speech genres” and conceptual frameworks are relevant. Learning, in the sense of mastering what Vygotsky (1986) refers to as scientific concepts, implies being able to contextualize phenomena in discourses that are often at odds with those that are used in everyday settings. Furthermore, in many situations there will be multiple, sometimes rivalling, scientific discourses that are relevant. In the present chapter, we want to illustrate how heteroglossia, and complexities of speech genres, are related to learning and thinking in the context of understanding scientific argumentation. Our ambitions are a) to argue for the significance of Å MÄKITALO A JAKOBSSON R SÄLJÖ 2 considering the complexity and diversity of speech genres (or discourses) when studying and theorizing about learning, and b) to illustrate the difficulties students encounter when faced with problems where multiple contextualisation must be handled in face-to-face interaction. We will do this by means of some examples from a case study, which illustrate the difficulties students have in dealing with so-called socioscientific problems (see below) that have become so popular in many science classrooms. We will also address the analytical challenges related to studying the demands on students in such complex settings. In our opinion, the methodological and theoretical consequences of this complexity of speech genres for our understanding of human learning are significant.