This paper clarifies the various definitions of additionality currently in use, and explores the relationship between additionality and key evaluation terms, such as impact and causality. It concludes that additionality should be assessed both ex ante and ex post, and that the presence of additionality will depend on institutional structures and on how different public and private interests are addressed. The paper further argues that the relevance of evaluation methods will depend not only on the applied financial and non-financial instruments but also on the types and dimensions of additionality to be evaluated. Several examples of different approaches to assessing additionality are analysed. The analysis provides a useful foundation for thinking through these issues, and will be of interest to both evaluation and blended finance actors. This paper is the second in a series of three working papers from the OECD/DAC EvalNet Working Group on Evaluating Blended Finance.