The existence of palm-dominated forests covering the island since the last glaciation and the recent deforestation by humans are paradigmatic in Easter Island's paleoecological reconstructions. The timing and mode of the deforestation are controversial, but there is general agreement that it actually occurred, and it is often given as an example of a human-induced environmental catastrophe with philosophical implications for the future of the whole planet. To evaluate whether this is the only well-supported hypothesis or if there might be other scenarios compatible with the paleoecological data, this paper reviews all the available evidence on past vegetation changes on Easter Island. The discussion is centered on three main points: 1) the alleged nature and extension of the former forests, 2) the taxonomic identity of the dominant palms, and 3) the nature of the recent ecological changes leading to a treeless island. The potential causes of the assumed deforestation are beyond the scope of this study. Concerning the first point, palynological and anthracological results obtained so far are not only compatible with a forested island, but also with other scenarios, for example a mosaic vegetation pattern with forests restricted to sites with a high freshwater table (gallery forests), which are mostly around the permanent lakes and along the coasts. With regard to palm identity, some extant species have been proposed as potential candidates, but the palms that dominated these forests seem to have become extinct and their identity remains unknown. The existence of a sedimentary hiatus around the dates of forest decline complicates the picture and reinforce the possibility of climatic changes. It is concluded that the hypothesis of a previously forested island has yet to be demonstrated. Therefore, the recent ecological disaster, human-induced or not, is still speculative. Several types of future studies are proposed for a better understanding of Easter Island's ecological history, including: modern analog studies from similar situations, pollen dispersal modeling, high-resolution multi-proxy studies along the cores obtained so far, more coring campaigns in the search for older sediments, and DNA and isotopic analyses of plant remains for taxonomic identification purposes.