Morgan’s classification of relationship terminologies included only two types: †‘descriptive’ and †‘classificatory’. ‘Descriptive’ terminologies were defined as those which distinguish direct relatives (direct ancestors and descendants of †ego, plus ego’s siblings) from †collaterals (all other †consanguineal relatives). ‘Classificatory’ terminologies were defined as those which fail to make such a distinction. By these means, Morgan classified all the Amerindian terminologies he came into contact with. He also sent questionnaires to *missionaries and American consuls throughout. the world, in order to compare terminologies with each other and decipher through them what he could of world history.
However, as Morgan’s own work suggests, the division of all the world’s terminologies into only two types proved to be inadequate. It is not merely the presence or absence of one distinction (direct/collateral) which is important. Other distinctions are equally interesting. †Kroeber (1909) found eight distinctions altogether. Of these one emerged as being of special importance: the distinction between †parallel relatives and †cross relatives. On the basis of these two distinctions—Morgan’s direct/collateral and Kroeber’s parallel/ cross applied in the generation of ego’s parents—†Lowie (1928) recognized four ideal types. Consider females of the mother’s generation, for example. One might (1) call mother, mother’s sister, and father’s sister all by a single term (as in the Hawaiian language). Alternatively, (2) one might classify mother and mother’s sister by one term and father’s sister by a different one (as in Iroquois). Another possibility (3) is to classify mother by one term and mother’s sister and father’s sister by another (as in English). Finally, (4) one might classify each genealogical position by a distinct term (as in Gaelic). Lowie called these four ideal types †‘generational’, †‘bifurcate merging’, †‘lineal’, and †‘bifurcate collateral’, respectively. ‘Generational’ terminologies make neither direct/collateral nor parallel/cross distinctions. ‘Bifurcate merging’ terminologies make parallel/ cross distinctions. ‘Lineal’ terminologies make direct/collateral distinctions. ‘Bifurcate collateral’ terminologies make both sets of distinction.
While Lowie emphasized the classification of relatives in the first ascending generation, †Murdock (1949) emphasized the classification of relatives in ego’s own generation. His typology consists of six classes. †‘Hawaiian’ terminologies make no distinction between siblings and cousins. †‘lroquois’ terminologies distinguish cross-cousins from parallel cousins and often classify siblings by the same term as that for parallel cousins. †‘Eskimo’ terminologies do not make parallel/ cross distinctions, but rather, distinguish cousins (collaterals) from siblings (direct relatives). †‘Sudanese’ terminologies, like Lowie’s ‘bifurcate collateral’ ones, make both kinds of distinction. In other words, they lack any general word for ‘cousin’ and call all cousins by strings of possessives.
Murdock’s other two types, *‘Crow’ and ‘Omaha’, like ‘Iroquois’, distinguish parallel from cross-relatives. The difference is that they treat generational differences in a peculiar way. The defining feature of Murdock’s ‘Crow’ (or ‘Choctaw’) type is that it classifies father’s sister and father’s sister’s daughter by the same term. The defining feature for ‘Omaha’ is that it classifies mother’s brother and mother’s brother’s son by the same term. Such terminologies often make further equations across the generation lines. For example, ‘Crow’ terminologies may classify father and father’s sister’s son by the same term, and ‘Omaha’ terminologies, mother and mother’s brother’s daughter. These classifications reflect descent group structure. In societies which possess ‘Crow’ terminologies, †matrilineal groups are usually present, and ego simply assimilates all members of his or her father’s matrilineal group and calls them by two terms (one for males and one for females). ‘Omaha’ terminologies tend to be found in strongly †patrilineal societies, where similar equations are made in reference to ego’s mother’s patrilineal group. Where such terminology structures generate (or reflect) *alliance, they are known as *Crow-Omaha systems.
The formal analysis of relationship terminologies has been used to provide a key to understanding different systems of marriage and descent. Much depends, though, on what connection is thought to hold between the classification of relationships and kinship practice: the confusions and arguments about the terms *prescription and preference, for example, stand or fall on this connection. In general, anthropologists have grown increasingly sceptical of claims that there is any necessary connection between systems of classification and kinship practice. Similarly, grand typologies, of the sort advanced by Murdock, have fallen almost entirely out of fashion.
