The paper describes a computer program, called KINSHIP, which performs
componential analysis of kinship terminologies. Some theoretical
insights gained from implementing the program are discussed. Three
styles of componential definitions are discerned, according to the
presence/absence of redundant components in kinterm definitions, and
special attention is given to the problem of multiplicity of
componential models. Two constraints are suggested, reducing the
immense number of solutions to unique analyses. The program, called
KINSHIP, with its capabilities to produce all alternative componential
analyses, including the guaranteed-simplest ones, is shown to be
usable as a practical tool in the study of kinship vocabularies.
Componential analyses of Bulgarian and American English are presented.