This paper argues that the content of extreme and non-extreme messages have significant parallels with one another. This argument is tested via a corpus comparison of 250 extreme, 250 mainstream and 250 counter-extreme messages. The paper has two key findings: (1) at a conceptual level, the extreme and mainstream messages, and the extreme and British Official counter messages demonstrate extensive similarity; and (2) the non-extreme Muslim authored counter messages and mainstream messages demonstrate extensive dissimilarity. The findings suggest support for substantial parallels between extreme and non-extreme material, which can be explained using models from social psychological theory, thus challenging popular understandings of a clear distinction between extreme and non-extreme language.