In recent years, in the areas of science education, cognitive psychology, developmental psychology, and cognitive science, a number of researchers have taken the position that in the course of their everyday interactions with the natural world, children construct theories that are in many ways similar to those constructed by scientists (Carey, 1985; Driver & Easley, 1978; McCloskey & Kargon, 1988). However, other researchers have argued that children’s theories are very different from scientific theories (diSessa, 1988; Inhelder & Piaget, 1958; Scholium & Osborne, 1985; Solomon, 1983a, 1983b). In this chapter we argue that children’s theories embody the essential characteristics of scientific theories. Our argument takes the following form. First, we give some recent evidence about children’s theories in the domain of observational astronomy that shows that children can develop very impressive theories of the natural world. Next we re-evaluate the characterization of children’s theories by those investigators who have argued that these theories are very different from scientific theories. We suggest that a knowledge-based approach to child development allows a reinterpretation of earlier findings, and we review a range of recent data that support this reinterpretation. We note that in these discussions many researchers have contrasted theories constructed by the individual child with positivist accounts of cultural achievements, such as Newtonian physics. We suggest that this is not an appropriate comparison and point out the importance of distinguishing between the child as theory constructor and science as a historical institution. Then we shift the argument to the characterization of scientific theories and suggest that the idealized description of scientific theories used in these debates is not supported by recent work in the philosophy and