Componential analysis is a formal, qualitative technique for studying meaning. There are two objectives: (1) to specify the conditions under which a native speaker of a language will call something (like a plant, a kinsman, a car) by a particular term and (2) to understand the cognitive process by which native speakers decide which of several possible terms they should apply to a particular thing.

The first objective is descriptive, but the second is a kind of causal analysis and is what the developers of the technique had in mind in the 1950s and 1960s (see Conklin 1955; Frake 1962; Goodenough 1956; Wallace 1962). Charles Frake, for example, described componential analysis as a step toward ‘‘the analysis of terminological systems in a way which reveals the conceptual principles that generate them’’ (1962:74). This created a lot of criticism, but more on that later.

Componential analysis is based on the principle of distinctive features in phonology, the branch of linguistics devoted to the study of the sounds of a language. To understand the principle, think about the difference in the sounds represented by P and B in English. Both are made by twisting your mouth into the same shape. This is a feature of the P and B sounds called ‘‘bilabial’’ or ‘‘two-lipped.’’

Another feature is that they are both ‘‘stops.’’ That is, they are made by stopping the flow of air for an instant as it moves up from your lungs and releasing the flow suddenly. An S sound, by contrast, also requires that you restrict the air flow, but not completely. You kind of let the air slip by in a hiss. The only difference between a P and a B sound is that the P is voiceless while the B is voiced—you vibrate your vocal cords while making a P.

If you add up all the phonological features of the words ‘‘bit’’ and ‘‘pit,’’ the only feature that differentiates them is voicing on the first sound in each word. The ‘‘pitness’’ of a pit and the ‘‘bitness’’ of a bit are clearly not in the voicelessness or voicedness of the sounds P and B, but any native speaker of English will distinguish the two words, and their meanings, and can trace the difference between them to that little feature of voicing if you push them a bit.

There is a unique little bundle of features that define each of the consonantal sounds in English. The only difference between the words ‘‘mad’’ and ‘‘bad’’ is that the bilabial sound M is nasal, and not a stop. These distinctive features carry meaning for native speakers of a language.

This principle can be adapted to the study of other domains of culture. Any two ‘‘things’’ (sounds, kinship terms, names of plants, names of animals, etc.) can be distinguished by exactly one binary feature that either occurs ( +) or doesn’t occur (— + ).

Table 17.5 A Componential Analysis of Four Things with Two Features

Feature 1

Feature 2

Thing 1



Thing 2


Thing 3


Thing 4

Table 17.5 shows that with two features you can distinguish four things: Thing 1 can be ( ++), thing 2 can be ( + —), thing 3 can be ( —+ ), and thing 4 can be ( — — ). Each bundle of features is different and defines each of the four things. With three binary features, you can distinguish eight things; with four, 16; with five, 32; and so on.

When componential analysis was introduced into cultural anthropology, it was applied to the set of English kinship terms (Goodenough 1956) and it continues to be used for understanding kinship systems (Pericliev and Valdez-Perez 1998). A daughter in English, for example, is a consanguineal, female, descending generation person. So is a niece, but a niece is through a sibling or a spouse.

Table 17.6 shows the distinctive feature principle applied to a list of 22 barnyard animals. (This example was first suggested by Hjelmslev 1961:70.) Stallions are adult male horses and foals are baby horses. Notice that there is no column in table 17.6 labeled ‘‘male’’ and no column labeled ‘‘juvenile.’’ A parsimonious set of features for distinguishing among these animals does not require all that information. A stallion is a nonfemale, gendered (that is, not neutered), adult horse. Any horse that is not female and gendered and not an adult must be a colt. A barrow is a neutered, adult hog, and a wether is a neutered, adult sheep.

Actually, if we wanted the most parsimonious set of features we would drop one of the last four columns in table 17.6. Those columns identify the general class of animals— horses, cattle, sheep, and swine—to which each named animal belongs. If the domain of barnyard animals comprised just those four classes, then a not-horse, not-cattle, not- sheep animal must be a swine. I’ve left all four animal classes in table 17.6 because there are other classes of barnyard animals not yet represented (chickens, rabbits, goats, etc.).

Componential analysis can be applied to any domain of a language where you are interested in understanding the semantic features that make up the domain. Table 17.7 shows a componential analysis of seven cars, using three features elicited from Jack (he of the taxonomy shown in figure 17.9).

A Corvette is an expensive car, not very practical, and not foreign; a Mercedes is an expensive, practical, foreign car; and so on. Each of the seven cars is uniquely defined by the three features Jack mentioned (box 17.2).

< Prev   CONTENTS   Source   Next >