This article represents an attempt to reconcile discussions of aspects of educational research with recent developments in complexity science. It is argued that current characterizations of and distinctions among research methodologies in education are potentially counterproductive, in large part because they tend to be defined against or in terms of principles and methods that have been rendered problematic within the sciences. To develop this point, the authors draw on several contemporary discourses: poststructuralist methods are used to foreground the Euclidean (plane) geometric roots of much of the vocabulary of educational research; fractal geometry is taken as a source of images and analogies to support alternative conceptions of knowledge, learning and teaching; informed by poststructuralist and fractal geometric notions, the authors turn to complexity science and argue that it is fitted to and offers important elaborations of current discussions of educational research methodologies. In the process, they suggest that it may be time to abandon some of the prominent distinctions used to describe educational research, including 'quantitative versus qualitative' and 'sciences versus humanities'.