The conventional view in the literature on the legacies of violence holds that the original experience of violence has a direct effect on victims and their descendants. We problematize this view, noting that political entrepreneurs seek to alter the legacy of violence by manipulating historical memory. To test whether such manipulation alters the legacies of violence we study a natural experiment in the Spanish Civil War whereby the assignment of violence was orthogonal to the political characteristics of affected settlements. Leveraging the data on actual violence during the civil war, an original dataset on historical memorials, panel data on municipal-level voting, as well as an original survey, we ask whether (i) the pattern of memorialization faithfully reflects the experience of violence, and (ii) whether present-day voting behavior is primarily a legacy of the original experience of violence or of the subsequent policy of memorialization. We find that memorials are not more likely to be constructed in communities that experienced more violence, but rather that political parties that are most likely to benefit from the presence of memorials are ones that are more likely to build them. And, indeed, we show that these parties reap electoral rewards from such memorialization policies. All in all, we demonstrate that contemporary political behavior is often a product of memorialization, which is highly politicized, rather than of the original experience of violence. These findings suggest an important corrective in the field of legacy studies.