Emic and etic (Anthropology)

The terms ‘emic’ and ‘etic’ were widely used in the American anthropology of the 1960s and 1970s, and the distinction between ‘emic’ and ‘etic’ levels of analysis was a commonplace in the areas of linguistic anthropology known variously as componential analysis or ethnoscience.

‘Emic’ and ‘etic’ (derived respectively from ‘phonemic’ and ‘phonetic’) designate two contrasting levels of data or methods of analysis. An emic model is one which explains the ideology or behaviour of members of a culture according to indigenous definitions. An etic model is one which is based on criteria from outside a particular culture. Etic models are held to be universal; emic models are culture-specific.

Just as phonetic and phonemic levels imply different methods of analysis, so too do etic and emic levels. So-called cognitive anthropologists, especially in the 1960s, were interested mainly in emic analysis (Tyler 1969). They saw culture as possessing structures similar to those of language. In contrast, anthropologists influenced by cultural materialism, especially in the 1970s, were more interested in etic analysis. They saw culture in terms of minimal units which defined appropriate behaviour, often in direct response to environmental circumstances (see Headland etal. 1990).

The terms ‘emic’ and ‘etic’ were first employed by fKenneth L. Pike in his monumental topic, Language in Relation to a Unified Theory of the Structure of Human Behavior. As this title suggests, their origin and early use reflect not only the analogy between ^phonological (phonemic) and cultural (emic) data, but also Pike’s theoretical stance (which sees linguistics as closely related to behavioural psychology) and his search for a grand theory which could encompass both language and culture. The sub-fields of f’ cognitive anthropology’, ‘the new ethnography’ and ‘ethnoscience’ which emerged in the 1960s all stem ultimately from Pike’s original concerns. These approaches emphasized emic over etic approaches through the meticulous analysis of semantic fields and indigenous classifications, and practitioners sought to apply Pike’s distinction both as a method of ethnographic research and as an aid to the theoretical understanding of the relation between specific and universal aspects of culture.


Etic and emic in cross-cultural comparison

Etic distinctions are explained in terms of various etic frameworks or classificatory grids. Classic examples of etic frameworks include:

Linnaean taxonomy; disease, in medical science; and the genealogical grid. Linnaean taxonomy is intended as a universal, hierarchical system for the classification of plants and animals on the basis of relative differences and similarities, and it entails an implicit theory of evolutionary relatedness. In contrast, the non-Linnaean classification of plants and animals in different cultures (e.g. the classification of bats as ‘birds’ rather than as ‘mammals’) is based on emic criteria, which may be quite different. Medical anthropologists make a similar distinction between ‘disease’ (a pathological condition, as defined by medical science) and ‘illness’ (the culturally specific understanding of disease). Diseases are defined in the same way wherever Western biomedicine is practised, whereas what counts as a particular illness varies in different cultural contexts.

These distinctions imply a value judgement, that those who have a special knowledge of Lin-naean taxonomy or Western medicine understand the true nature of the universe, and that cultures in which ordinary people have access to this specialist knowledge are superior to those in which ordinary people do not have such access. However, not all etic frameworks carry this notion of superiority and inferiority. In the study of relationship terminology the genealogical grid, which arguably is extrinsic to Western culture, is more neutral. This is a particularly good example for examining the relation between emic and etic distinctions, as well as the problems which can arise in reifying the emic/etic distinction.

The genealogical grid precisely denotates each genealogical position. These positions are presumed to be the same for all languages and cultures. The emic distinctions are those which enable languages to define their kinship categories differently, employing common terms for different combinations of genealogically defined kin. ‘Aunt’ and ‘uncle’, as distinct from ‘mother’ and ‘father’, are not universal notions but rather the specific categories of the English language and of the societies in which this language is used; other languages may classify English-language ‘cousins’ as ‘siblings’ or as potential ‘spouses’, and so on.

Analysts might distinguish the etic notion of the genealogical mother, written ‘M’, from the emic notion of the biological or social mother in British or American culture, written ‘mother’. As the italics imply, this ‘mother’ is a culture-specific one, as foreign to the etic notion as a comparable word in any other language. Yet there are two problems here. First, what ‘motherhood’ might mean in any specific culture is a question beyond the confines of such simple linguistic distinctions and requires further emic analysis. Etically, it can only be defined very loosely. Second, the fact is that anthropologists have cultures and cultural preconceptions like anyone else, and they write in one specific language at a time. Such a language, of course, will have its own emic categories, and the etic grid accordingly remains elusive. In kinship the etic grid is relatively easy to specify, but in other aspects of thought (say, in the realm of religious belief), etic distinctions are very much more difficult to define and utilize with any precision.

The emic model is not the native’s model

A commonplace assumption about emic models is that they are ‘discovered’ rather than ‘invented’ by the analyst. However, emic models, like phonemic ones, are ultimately exogenous constructions, formalized by the analyst on the basis of distinctive features present in indigenous usage. They are not in themselves ‘the native model’, though anthropologists often loosely identify them in this way.

This may be illustrated by Conklin’s (1969) example of the structure of the pronouns in Hanunoo, a language spoken in the Philippines. Conklin argued that the conventional linguistic (etic) distinctions — first, second and third person; singular, dual and plural; and exclusive and inclusive — only describe Hanunoo pronouns in an inelegant and uneconomical way. These distinctions account for all Hanunoo pronouns, but they produce no less than four potential categories which the Hanunoo language does not distinguish. It is better, he suggested, to examine the distinctive contrasts made by the language itself. In doing this, he came up with three sets of emic distinctions for Hanunoo pronouns — minimal membership, non-minimal membership; inclusion of speaker, exclusion of speaker; and inclusion of hearer, exclusion of hearer. The application of these distinctions generates all and only the eight pronouns found in the language,and the resulting analysis is therefore more elegant and economical than the one employing the etic categories traditionally used by linguists. Yet the emic criteria he identified are distinctions which are not named or even consciously employed by the Hanunoo themselves. They are only implicit in indigenous usage.

As this example shows, an emic model is not necessarily a model held consciously by indigenous thinkers. Here it is clearly an analyst’s model, but one which is built up from principles derived from, rather than forced upon, the data. This is equally true of behavioural, semantic or phonological data. Just as no native speaker, simply as a native speaker, can coherently describe the phonological system of his or her language, similarly no indigenous thinker can usually present a complete emic analysis of his actions or of a culturally significant semantic field of his language. Analysis, even emic analysis, is the job of the observer.

Critiques of emic and etic

Although the emic and etic levels of culture are intended to correspond analogously to phonemic and phonetic levels in language, there are nevertheless crucial differences between culture and language which make the correspondence problematic. Most obviously, culture is much more variable than language, and cultural behaviour is much more difficult to assign to a single structure than speech is.

Marvin Harris (1976) has objected to the notion that culture is made of sets of rules or ‘grammar’, in effect denying the possibility of emic models at all. He argued, especially against Goodenough (1956), that the methods of linguistics are a poor example for anthropologists to follow, since there is no anthropological equivalent to a native-speaker or one possessing absolute ‘cultural competence’ in any sense analogous to linguistic competence. Goodenough’s view was that the native ‘authorities’ should be sought and that their ideas should be used in the construction of emic models. In Harris’s view, several problematic questions remain. Is there any such thing as a cultural authority? If so, how can such a person be identified? What about the ideas of those who are not considered authorities, but merely average members, of their own culture?

Others have questioned the existential status of etic models. What guarantee is there that the observer’s supposedly objective, etic model is not in fact his or her own emic one? Since the 1980s, under the influence of postmodernism and reflexivity, critics have challenged the notion of objectivity upon which etic grids depend. These approaches imply instead that an interplay between what might be considered the emic models of the observer and the observed are as close as we can get to an etic level of analysis.

The future of emic and etic

As Levi-Strauss (1985: 115-20) has pointed out, the emic level is the level of perception. People do not understand sounds as sounds, but through the phonological structure of their language. Likewise, people understand actions or words only through the culture they possess. Thus, in Levi-Strauss’s view, the materialist objection to the emic as merely culture-specific and not based on objective principles does not hold. The poststructuralist objection to the etic is more difficult to counter on a philosophical level. However, the simple answer to this apparent dilemma is to seek objectivity, while realizing that it is elusive. Clearly, etic models can exist as heuristic devices, but they are as problematic as emic ones to define precisely.

The concepts ‘emic’ and ‘etic’, although less often discussed today than in the past, are implicit in more recent anthropological approaches, even postmodernist and reflexive ones, where they exist as exemplars of the contradictions in anthropology itself. They are also taking on new significance in regional analysis and regional comparison. A defining feature of the classic emic approach is that ideology or behaviour is studied from ‘within’ the cultural system. This implies that only one cultural system can be studied at a time, and in the past the cultural system was often taken as equivalent to one culture or society. Yet, for those who define cultural systems more broadly, i.e. who draw their boundaries around a wider geographical area, renewed interest in a more elaborate version of the emic/etic distinction shows promise.

Next post:

Previous post: