The diversity of extant carnivores provides valuable opportunities for comparative research to illuminate general patterns of mammalian social evolution. Recent field studies on mongooses (Herpestidae), in particular, have generated detailed behavioural and demographic data allowing tests of assumptions and predictions of theories of social evolution. The first studies of the social systems of their closest relatives, the Malagasy Eupleridae, also have been initiated. The literature on mongooses was last reviewed over 25 years ago. In this review, we summarise the current state of knowledge on the social organisation, mating systems and social structure (especially competition and cooperation) of the two mongoose families. Our second aim is to evaluate the contributions of these studies to a better understanding of mammalian social evolution in general. Based on published reports or anecdotal information, we can classify 16 of the 34 species of Herpestidae as solitary and nine as group-living; there are insufficient data available for the remainder. There is a strong phylogenetic signal of sociality with permanent complex groups being limited to the genera Crossarchus, Helogale, Liberiictis, Mungos, and Suricata. Our review also indicates that studies of solitary and social mongooses have been conducted within different theoretical frameworks: whereas solitary species and transitions to gregariousness have been mainly investigated in relation to ecological determinants, the study of social patterns of highly social mongooses has instead been based on reproductive skew theory. In some group-living species, group size and composition were found to determine reproductive competition and cooperative breeding through group augmentation. Infanticide risk and inbreeding avoidance connect social organisation and social structure with reproductive tactics and life histories, but their specific impact on mongoose sociality is still difficult to evaluate. However, the level of reproductive skew in social mongooses is not only determined by the costs and benefits of suppressing each other's breeding attempts, but also influenced by resource abundance. Thus, dispersal, as a consequence of eviction, is also linked to the costs of co-breeding in the context of food competition. By linking these facts, we show that the socio-ecological model and reproductive skew theory share some determinants of social patterns. We also conclude that due to their long bio-geographical isolation and divergent selection pressures, future studies of the social systems of the Eupleridae will be of great value for the elucidation of general patterns in carnivore social evolution.