This article reanalyses the Dravidian kinship terminology and shows th
at the distinctions it makes are distinctions of relatedness in which
differences of gender in the links between people make the difference
between cross and parallel kin. The categories of the terminology do n
ot rest on the marriage system, as Dumont has argued, but rather on pr
ior considerations of marriageability which ultimately rest on conside
rations of relatedness. The understandings that underlie the Dravidian
system are intuitive and practice-based and may be overlaid with any
number of more elaborate theoretical discourses. But these understandi
ngs are the minimum conditions which create the possibility of cross-c
ousin marriage.