In this paper I discuss similarities and differences between a potential new model of language development - lexical selection, and its biological equivalent - natural selection. Based on Dawkins' (1976) concept of the meme I discuss two units of language and explore their potential to be seen as linguistic replicators. The central discussion revolves around two key parts - the units that could potentially play the role of replicators in a lexical selection system and a visual representation of the model proposed. draw on work by Hoey (2005), Wray (2008) and Sinclair (1996, 1998) for the theoretical basis; Croft (2000) is
highlighted as a similar framework. Finally brief examples are taken from the free online corpora provided by the corpus analysis tool Sketch Engine (Kilgarriff, Rychly, Smrz and Tugwell 2004) to ground the discussion
in real world communicative situations. The examples highlight the point that different situational contexts will allow for different units to flourish based on the local social and linguistic environment. The paper also shows how a close look at the specific context and strings available to a language user at any given moment has potential to illuminate different aspects of language when compared with a more abstract approach.