Componential analysis (Anthropology)

Componential analysis is a method of formal analysis or of ethnographic description whose origin is usually traced to Goodenough’s article ‘Componential Analysis and the Study of Meaning’ (1956). Its essence is the study of ‘components’, which are the basic building blocks of meaning in a semantic domain. Proponents of componential analysis see such domains as crucial to the understanding, not merely of languages, but also of cultures or significant aspects of culture. For example, kinship terms are not random words for specific relatives, but rather exist in a culture-specific classification system. Such systems almost always distinguish the sex and generation of relatives, but sets of components such as direct and collateral, or parallel and cross, will be specific to given cultures or languages. Consider the classification con-sanguines (blood relatives) in English kinship (see Table 2 below).

Here the components are sex (or gender), generation (or genealogical level), and lineal versus collateral. Sex distinguishes ‘father’, for example, from ‘mother’ (both father and mother being first-ascending-generation direct relatives). Generation distinguishes ‘father’ from ‘grandfather’, ‘brother’, ‘son’ and ‘grandfather’ (all these being male direct relatives). Directness distinguishes ‘father’ from ‘uncle’ (both father and uncle being male first-ascending-generation relatives). If we were considering affines as well as consanguines, the ‘father’ would be distinguished from the ‘father-in-law’ by the component of consanguinity itself.


Proponents of componential analysis, whether in cultural anthropology or in linguistics, often talk about components by the Latin word sig-nificata (singular: significatum). ‘Male’ would be an example. The terms within the given domain are called designata (e.g. ‘father’), and the elements which make up the category are its denotata (in this case genealogical position of the father, which kinship specialists call F). There is a logical distinction here between ‘father’ and F, as the former is a word in English and the latter is a position relevant to any language or kinship system but whose significance literally depends on the system. In many languages F is classified with FB (father’s brother) as belonging to a different category than MB (mother’s brother), and in societies which speak such languages relations between relatives will differ markedly from those in English-speaking societies. A final notion relevant here is that of connotata. What ‘fatherli-ness’ connotes in English is that which defines ‘father’ beyond the purely formal componential distinctions drawn above: fatherly attitudes or fatherly behaviour, whatever these might be.

Table 2

Generation

Direct relatives

Collateral relatives

MALE

FEMALE

MALE FEMALE

+2

grandfather

grandmother

+ 1

father

mother

uncle aunt

0

brother

sister

cousin

-1

son

daughter

nephew niece

-2

grandson

granddaughter

Componential analysis is ideally suited to the study of relationship terminologies both because of the precision of classification in that domain and because of the sociological importance of such classification. It has been used in other domains too, especially in "ethnobotany and "ethnozoology. Indeed, a loose synonym for ‘componential analysis’ is ‘ethnoscience’ (see e.g. Frake 1980). Another is ‘cognitive anthropology’, reflecting the supposed cognitive reality behind the semantic distinctions which compo-nential analysis reveals. Both these terms were common in the 1960s, when componential analysis was at it height in theoretical interest in American anthropology (see also emic and etic).

One semantic domain where things are much more fluid than kinship or ethnoscience is that of colours, and this domain has also attracted much interest especially in cross-cultural comparisons. Here the components are not clearly definable in terms of structural oppositions, but are arranged on sets of continua: the intensity of light (dark to light) and the wavelength of light (red to violet).

Different languages classify colours very differently, and there may be differences even between languages spoken by people who are bilingual, such as Welsh people who speak both Welsh and English (Ardener 1971). Traditional Standard Welsh (as opposed to Modern Colloquial Welsh) has no equivalent to English brown. Some shades of ‘brown’ are called llwyd. Other shades are called du. Loosely, llwyd means ‘grey’ and du means ‘black’, but as they encompass ‘brown’ too their signification is greater than that of the English word. Much the same is true of the Welsh word glas, which loosely translates into English as ‘blue’ but which also includes shades of what English classifies as ‘grey’ and ‘green’. Therefore the Welsh word gwyrdd, loosely ‘green’, is narrower in meaning than its rough English equivalent. Blue-green might be ‘green’ in English, but it is glas (blue) in Standard Welsh. No language classifies everything. With colours, it would indeed be impossible to classify everything, since there is an infinite degree of natural variation. Standard Welsh simply does not need the distinctions which English-speakers take for granted.

Next post:

Previous post: