This dissertation brings a contribution to the philosophical debate on adaptive landscapes, an influent "model" or "metaphor" in evolutionary biology. Some elements of innovation are: the distinction between native and migrant metaphor; a processual and communicational idea on what the Modern Synthesis was, and on what role a metaphor could have played in it; a view (taken by Richard Lewontin) of the disunity and theoretical structure of population genetics; the distinction between “adaptive surfaces” (mainly metaphors) and “combination spaces”, two terms normally conflated in the word “landscape”; an analysis of what bridges (including heuristics) may be cast between equations of gene frequency and the genotype space that, due to its huge dimensionality, cannot be handled by mathematics; a specified vocabulary to be used to clear the adaptive landscapes debate, accompanied by a plea in favor of a pragmatic approach - for example, the plurality of available notions of model forces us to choose one notion and see where it brings, otherwise we get stuck in confused, endless debates; an updated analytical comment of recent landscapes - Dobzhansky, Simpson, Dawkins but also the proliferation of combination spaces used in evolutionary biology to address a great variety of problems; the vision (got by Sergey Gavrilets) of a patchwork of tools finally making Mendelian population suitable model also for speciation; the exact position of holey landscapes in this patchwork, and the idea that scientists’s questions - like “how possibly” questions - matter in accessing this patchwork and in deciding “what explains” and “what describes” what in the world; the direct response to some mistakes Massimo Pigliucci made, I think, in his assessment of the adaptive landscape; an analysis of the Extended Evolutionary Synthesis project at its present stage, and some reflections on the conditions that will allow such a project to give a fair treatment and a good position to tools from the past, like the adaptive landscapes.