ABSTRACT Can people really kill their neighbours – massively and out of hatred? Is it anthropologically possible? Massacres and genocides of neighbours have become a subcategory of mass violence studies over the past two decades. This scholarship implicitly refutes entire parts of anthropology, notably the Durkheimian tradition, for which society is organized to produce bonds and bonds are inherently irenic, while liberally using key anthropological concepts – intimacy, social relationships, and the figure of the neighbour. These terms are rarely defined or explored. This scholarship implicitly makes four articulated assumptions: neighbours entertain social relationships, that is the essence of neighbourliness; vicinity therefore equates sociality; therefore, when neighbours kill neighbours, they commit social violations; therefore, bonds can be overwhelmed by beliefs, hatred, or other circumstances. However, a close reading of the evidence provided by this historiography shows that descriptions of preexisting sociality are briefly sketched, typically stretching vicinity and mixing vicinity with sociality. By overlooking the scope and quality of relationships between victims and perpetrators, this scholarship tends to treat simple connections, or what network sociology would call “weak ties” with intimacy and kinship with sociality; it has also overlooked how human groups tend to organize and limit sociality across antagonist identities. Vicinity does not predict sociality, even in times of peace, even in holist societies, and even more so when relations across antagonist groups are mediated by rumours, envy, and cultural insecurity. This article proposes a way forward with the use of socio-metrics to better understand social ties and their proprieties.