Representing Anthropological Knowledge: Calculating Kinship
Michael D. Fischer
Analyzing and Understanding Cultural Codes


Kinship Introduction
Learning Kinship with
the Kinship Editor
Use the Kinship Editor
Kinship Editor Results

Kinship Contents

Kinship Contents

Kinship Contents

Kinship Contents

Kinship Contents

Kinship Contents

Kinship Contents

Introduction: Kinship and Kinship Terminologies

The scientific study of kinship began with the publication of Lewis Henry Morgan's Systems of Consanguinity and Affinity of the Human Family, published in 1870. Morgan had amassed a huge amount of data on kinship terminology, and using this he worked out a classification of kinship systems. Morgan assumed that human society had evolved through a series of stages from primitive savagery to civilization, and he saw kinship terminologies as reflecting these stages. Primitive promiscuity, for example, is signalled by a Hawaiian type of kinship nomenclature. Morgan made two major criterial distinctions between kinds of kinship terms: classificatory terms, which subsume a relatively large number of biological kin types, and descriptive terms, which subsume relatively small numbers of types - preferably having unique referents. He imposed this scheme on whole terminological systems. He then fitted the typological scheme to his evolutionary framework, where he said that "primitive systems were classificatory, whereas civilized systems were descriptive. He ignored the problem of how to analyze degrees of extension, or the how to discover the semantic criteria by which people made distinctions between kindred.

Kroeber, in a 1929 paper, exploded the notion of any kind of simple classificatory/descriptive typology. This was quite an important paper. He looked at the principles that were used in separating kinds of kin, and suggested eight: generation, affinity, collaterality, sex of relative, bifurcation, sex of speaker, relative age, and decedence. Lowie added a ninth - polarity. He examined the kinds of differences which can be employed to distinguish kin. For example, in English and American kinship terminology, father and son differ only on the dimension of generation. Father and father-in-law differ in that father-in-law is an affine, that is, a relative by marriage. Father and uncle are distinguished by collaterality. The criterion of collaterality rests on the distinction between siblings and lineal relatives. In English, only cousin ignores the distinction of collaterality. Cousin also ignores the distinction made concerning sex of the relative concerned - the female child of our mother's brother is called by the same term as his male child. Grouping lineal and collateral relatives under the same term is technically called "merging", and in kinship systems in general the relatives most frequently merged are a parent and sibling of the same sex, a sibling and parallel cousin, or a son or daughter and nephew and niece.

Bifurcation means "forking", and recognizes that relatives may be traced through either a male or female connecting relative. Our own kinship system ignores this - an uncle may be traced through either parent, a grandparent is a parent of either or our parents.

Polarity recognizes that a relationship consists of two parties, and thus two terms - aunt, niece. If this criterion is ignored, the two parties call each other by the same name. In English, cousin is an example. "Brother is not, although both brothers use the same term to indicate each other. Brother results from the fact that the same criteria is being used to assign both people to the classification. In English, we almost always acknowledge polarity. In some kinship systems, this is not the case: it is, for example, fairly common for grandfather and grandson to be called by the same term.

The last three criteria are much less commonly used - the criteria of relative age recognizes that within a generation people differ in age. They have different terms for elder brother and younger, elder brother of the father and younger brother of the father, for example. Some groups have different terms depending on the sex of the speaker - male speakers use different terms than female speakers. Decedance assigns a different term to a relative depending on whether the relative is alive or dead.

Kroeber's 1929 paper is important because, first, it destroyed the notion of a simple classifactory / descriptive typology of kinship terminologies. All kinship systems have both classificatory and descriptive terms. Second, because it suggested a way of dealing with the mass of kin terms from different groups. Malinowski, in 1930, said that "We badly need a large collection of native terminologies, collected on a frame of reference that excludes direct translation to our own kinship terms. Until this has been done, discussion of their meaning would appear to be premature." Much of the work that has been in kinship terminologies has been an attempt to describe systems of kin terms on their own terms, without reference to direct translation. Componential analysis is one such example. Most of the work you read on kinship terminology will involve arguments about either how to describe the terms or the relationship of the terminology to behaviour within the society.

Kinship Terminologies

Kinship is one of these more complex systems of culture. All human groups have a kinship terminology, a set of terms used to refer to kin. Many parts of life are impacted by kinship, and in most societies kinship relations influence things like who one can and can not marry, who one must show respect to, who one can joke with, and who one can count on in a crisis.

Kinship terminologies vary in different societies from as few as twelve to more than fifty terms. English kinship terminology is in the middle, and contains the following principal terms:

mother, father, son, daughter, brother, sister
uncle, aunt, nephew, niece
cousin (differently elaborated in different English speaking cultures)
grandfather, grandmother, grandson, granddaughter
granduncle, grandaunt, grandniece, grandnephew (in many dialects)
great-grandmother, great-great-grandmother etc.
great-grandfather, great-great-grandfather etc.

there are also the affinal terms:

wife, husband, brother-in-law, sister-in-law, mother-in-law and father-in-law as well as uncle and aunt.

Anthropologists have learned many interesting things about kinship terminologies. In terms of our topic today, how cultural ideas are organised, kinship terminologies are organised in at least two important ways.

Firstly, they provide a means of classifying relationships with other people, for every person in the society. When different kinds of genealogical relationships are merged into one category, such as (in English terminology) all male siblings being denoted as brother, or all mothers of parents being called grandmother, this reduces the information that might have been needed (many terminologies have different terms for male siblings, often based on relative age, and many have different terms for father's mother and mother's mother) to describe kinship relationships. Thus genealogical relationships are different from kinship relationships. There are a large number of genealogical relationships. For example, in a society of 1000 individuals, there may be nearly one million genealogical relationships, though this number will more typically be a few hundreds of thousands .

In most kinship terminologies this large number of genealogical relationships can be denoted with the 12 to 50 terms that make up human kinship terminologies. What makes this possible is using a limited number of classificatory criteria to define terms, limiting the distance that counts as a denotable kinship relationship, and the fact that the use of these terms is relative to each individual in the society. That is, every person in a society will denote a different person as mother, father etc.

Kinship terminologies are thus systematically limited by classificatory restrictions and relative application.

Preparation of data for genealogical processing

To begin any analysis using a computer, the data upon which the analysis is to be based must be represented in an arbitrary form which an appropriate computer application can use. Ideally, a computer would accept data in the same form in which it was originally collected. This was the speculation of Gilbert (1971:), who predicted that 'soon' genealogical data might be input directly in the form of diagrams using scanners. Unfortunately, this has not come to pass (yet) and it is rarely the case that the original form of the data will resemble that used for data input.

If you intend to use an existing application, such as those by Gilbert (1971) or Ryan (1988), these applications will have specific formats in which the input data must appear. If you use the programs in Chapters 7, or you design and write your own application, each program will have some pre-existing format for the data, although you will have control over this format.

Even if you have entered your data directly into a computer, it is unlikely to be in the correct format, unless you have taken the required format into consideration at the time of entry. In most cases it would not be desirable to record the data directly in such a format. The data collection model usually has quite different demands from the data analysis model, whether the analysis is to be undertaken by computer or manually. For example, if the data collection model conforms to the recommendations of Barnard and Good (1984:30):

(1) F, (2) FP, (3) FG, (4) FGE, (5) FGC, (6) FGCE, (7) FGCC
(8) M (9) MP, (10) MG (11) MGE, (12) MGC, (13) MGCE, (14) MGCC
(15) E, (16) EP, (17) C, (18) CE, (19) CEP, (20) CC, (21) CCE, (22) CCEP
(23) EG, (24) EGE, (25) EGEP, (26) EGC, (27) EGCE
(28) G (29) GE, (30) GEP (31) GC, (32) GCE, (33) GCEP, (34) GCC, (35) GCCE
(36) GEG, (37) GEGE, (38) GEGC, (39) GECGE

G = Sibling, E = Spouse, P = Parent, C = Child, F = Father, M = Mother

Suggested order of procedure in genealogy collection.
Reproduced from Barnard and Good (1984:30):

a sample (a very orderly one at that) of the data as collected might appear as:

House 705 D2BIII, Consultant/Ego: Muzhar Ali Khan, Male, 27 years b 1960, Head, engraver

F Kazeem Ali Khan, pio, Male, 55 years at death, b circa 1930, clerk on depot
FP Qadar Ali Khan, baba, Male, 60 years at death, b. c. 1905, farmer
Saifia Khan, mama, Female, 57 years at death, b c 1910
FG Abdul Ali Khan, taiya, Male, 61 years, b c. 1927
Mustaffa Ali Khan, chacha, Male, 58 years, b.c. 1912

Sample of genealogical data in format of Barnard and Good (1984:30)

For manual or computer analysis the data must be converted from a set of extended genealogies relative to a few people, to some sort of 'normalised' form, from which the data can be cast relative to an arbitrary person. The usual options for conversion are:

a) Hand process the data from the collection model into the required form, and enter the data using a text editor. This most resembles the 'traditional' process of copying, cutting, pasting, indexing and sorting.
b) Hand process the data from the collection model into the required form, and use a computer program which emulates a data entry form. The form which will organise the entry of data and perform some simple error checks on the values entered, improving the 'correctness' of the data entry. There are a number of readily available programs available for most computers for this purpose.
c) Code and enter the data in a form close to the collection model, using a text editor or a database manager, and use an application to translate it to the required form. This, in effect, will require a customised program specified by yourself or a computing partner. This is a particularly easy type of program to design, assuming that the collection model itself is fairly orderly and logically consistent with the required target format.

In the next section we look at another possibility.