Philosophers have paid attention to some probabilistic aspects of classical evolutionary genetics, while they have left aside the role of probabilities with respect to the origin and impact of phenotypic variation. In this project, I intend to fill an important gap of this situation by philosophically characterizing the interactions between the evolutionary propensities of the classical framework of evolutionary genetics (fitness and the capacity to drift) and those of new methodological approaches, notably evo-devo, interested in the nature and impact of variation in evolution, known as ‘variational tendencies’ (evolvability, variability, robustness, modularity and plasticity). The aim of the project is to develop a coherent, integrative framework for the relationship between these two kinds of evolutionary properties, as well as the relevant notions of randomness and contingency that they demand. Such a coherent philosophical framework will move beyond the current acknowledgement of a pluralism of approaches to evolution towards an encompassing view of the current diversity in explanatory and predictive probabilistic models within the evolutionary disciplines. More particularly, filling this gap is an important step towards the overcoming of two dichotomies in the philosophy of evolution: the opposition between ‘population’ and ‘typological’ thinking, and the one between necessity and contingency.