Human Universals
Page 12
In addition to Hallowell’s paper, described above, there were a few other exceptions—some only implicit or unintended—to the lessened anthropological interest in universals that could be seen in the late 1950s and early 1960s. One was the universalism (of one kind or another) that is implicit in the search for valid cross-cultural generalizations of almost any sort (see chapter 2). These were the subject of much anthropological research and writing from the fifties into the seventies, and I suggest that the paucity of explicit discussion of the universals that such studies implied was yet a further indication of the ambivalence anthropologists feel toward universals.
Another exception was the development of “componential analysis” (see, e.g., Goodenough 1956). Also called “ethnosemantics,” and included under the rubric of “ethnoscience,” componential analysis is an ethnographic method in which a given lexical domain—say the set of words for plants, or kin, or colors—is isolated so that by inquiry and observation the anthropologist may determine what underlying semantic “components” give the domain its form. For example, in the domain of address terms in English, comprising “Mister,” “Mrs.,” “Miss,” and “Ms.,” there are two semantic components: sex (male and female) and marital status (married, unmarried, and undesignated). “Mister” is defined as “address term male,” “Mrs.” as “address term female married,” etc. In a similar domain among Brunei Malays, by contrast, a component of rank is not only present but ubiquitous: it has four gradations that must be kept in mind to use Brunei address terms (Brown 1976:163–164). The presence or absence of a rank component distinguishes English from Malay address terms, but the common presence of sex and marital status links them. By analogy with the linguistic ideas from which this mode of analysis was derived, the raw facts that are the native’s own terms are emic, while the facts derived beneath the surface, and expressed in cross-culturally valid terms, are etic (see the discussion of emic and etic in chapter 2).
The only thing explicitly universalistic about componential analysis was that it was presumed to be universally applicable; it was a universal model. What is particularly noteworthy in the method is its success in remaining faithful to emic fact—starting as it does with the very words by which another people captures its own modes of organizing thought—while routinely penetrating beneath the surface facts to the underlying semantic elements that structure cognition. In its harmonization of these two goals, componential analysis was a remarkable breakthrough in method. As time went by, this method produced evidence that beneath the bewildering variety of words by which peoples classify the world about them there were some important universal conceptions in the underlying semantic components (Bloch 1977; discussed further below and in the next chapter).
Walter Goldschmidt’s Comparative Functionalism: An Essay in Anthropological Theory (1966) contained an explicit consideration of universals. Goldschmidt presented a universal model for the analysis of society that was apparently inspired by Malinowski’s model (1960 [1944]). Goldschmidt argued that underlying the diversity of human institutions is a universal set of problems or functions that must be solved or discharged in all societies. Consequently, these functions provide a common framework for the analysis of all societies. In the course of his essay Goldschmidt drew attention to the profound influence that anthropological relativism had exerted on the moral philosophy of the modern world, and he criticized extreme forms of cultural relativism and anthropology’s overemphasis on exceptional cases. He also defended reductionist explanations against the superorganicists.
But if an interest in universals had slipped somewhat in anthropology for a while, great strides in their study continued to be made in linguistics. The relationship between linguistics and cultural anthropology has been long, intimate, and productive, particularly in the United States (Hymes 1970). Language is often thought of as the epitome of culture, and forms of analysis employed in one field often apply to the other as well. Linguistic phenomena are cited for some of the clearest statements of relativity, the classification of color being a notable instance. It is a commonplace assumption, for example, that speech sounds have only an arbitrary connection with what they signify.12 Thus there is, according to this assumption, no intrinsic connection between the sounds we make in the word “horse” and the creature it stands for. The German word Pferd is just as arbitrary and just as fitting as a way to signify the same creature. This apparent or actual arbitrariness has often been extended to other aspects of language and elevated to a general principle.
But although linguists had no particular reason to be looking for universals—indeed they had many reasons to expect relativity—they found them. There are two important reasons for this. One is the objectivity of their methods, whose scientific power was demonstrated already in the nineteenth century, particularly in the study of sound shifts (as formulated, for example in Grimm’s laws of sound shifts in Indo-European languages),13 which is crucial in the reconstruction of the relationships between languages. The other was the relative simplicity of the materials they treat: linguists delimit the scope of their research more than cultural anthropologists do, and in many if not most contexts pursue a strictly formal mode of analysis that ignores causation.
A pioneering but very brief essay on linguistic universals by the Aginskys (1948) was closely linked to the anthropological revival of interest in universals in the United States that began just before World War II. But the most important works appeared in the late 1950s or 1960s. The linguist Noam Chomsky’s influential review of B. F. Skinner’s Verbal Behavior was published in 1959. This review had three notable consequences. First, it was a devastating criticism of behaviorism, along with all that behaviorism stood for in the nature-nurture controversy. Second, it posited “deeper processes” of language acquisition that were innate and therefore presumably universal. This led to a search for grammatical universals (the “deep structure” of language) that remains a preoccupation in linguistics to the present. Third, by referring to work then in progress by Eric H. Lenneberg (see, e.g., 1967), Chomsky drew attention to the biological foundations of language. Chomsky was soon to write of the “speech organ,” in order to draw attention to the profound sense in which language is not learned: in a natural environment of other speakers, an individual acquires language as naturally as pubic hair. Referring to the studies of ethologists, Chomsky likened speech to those instinctive behaviors in lower animals that are acquired by “imprinting” during sharply delimited periods of an organism’s development.14
Joseph Greenberg’s Language Universals, first published in 1966, was detailed in its listing of universals and signaled a broader search for linguistic universals that also remains strong to the present. In this work Greenberg gives particular attention to the phenomenon of “marked” versus “unmarked” categories, a phenomenon found in all languages at the three major levels of linguistic analysis: phonemic, grammatical, and semantic. Marking, a universal process, generally produces implicational universals; occasionally it results in unrestricted universals. Since linguistic universals are not generally familiar, even to most anthropologists, examples will be given.
The phenomenon of marking is easily illustrated. In English he is unmarked, she marked. We use the former as a default term when sex is left ambiguous or unknown (at least we did until the recent introduction of unisex forms like “s/he”). Similarly, author is unmarked, authoress marked; nurse is unmarked, male nurse marked. Marking in these instances refers to the addition of s-, -ess, and male to mark one term of each pair and hence distinguish it from the other. The unmarked term is sometimes said to have “zero expression,” meaning that nothing is added to it. Note that the overt marking seen in these examples is not always present. Marking is normally accompanied by a number of other characteristics (Greenberg 1987; Schwartz 1980); if enough of the other characteristics are present, even though overt marking is not, linguistic elements may still be designated as a marked and unmarked pair.
An example o
f marked versus unmarked phonemes is found in the German word-final -d/-t. Whichever phoneme occurs at the end of a word, it is pronounced as though it were -t. Thus Tod (death) is pronounced “tot.” So, too, with the other final consonants in German that form a contrast set of voiced versus unvoiced. In these cases the voiced consonant is the marked phoneme, the unvoiced the unmarked. The unmarked can take the place of the marked, but not vice versa. In the event of the two phonemes ultimately merging into a single one—which is one of the major evolutionary processes of language drift—they will normally merge into the unmarked form, so that no phoneme retains as one of its necessary and distinctive features a feature that does not contrast with the absence of that feature (for example, no initial consonants must be voiced if there are not unvoiced initial consonants).
An example of the marked versus the unmarked at the level of grammar is provided by the contrast between singular and plural. In all languages where one is marked, as occurs in English generally with the addition of -s, it is always the plural. A further example is that the negative of a sentence is always marked, the positive usually having zero expression. For example, conceivably a sentence of the form “it goes” could be understood to mean “it goes [not],” and one would have to add something, for example, “it does go,” to make it positive—just as, in theory, words could be plural unless one added something to make them singular. In spite of the equally sound logic of these two possibilities, neither occurs in any language. A partial exception that Greenberg cites is that in Vietnamese there is a form to indicate the positive, but it is not compulsory.
Straddling the boundary between grammatical and semantic universals are regularities in the expression of “good” and “bad.” In all languages that have a word for “good,” its opposite may be expressed in two ways: some languages contrast it with a word for “bad,” some with “not good,” (and some with both). No language has the words “bad” and “not bad” with no word “good,” although “bad” and “not bad” are the logical equivalents of “good” and “not good” or “good” and “bad.” Thus “good” is universally unmarked; it is never the marked term of a contrast set.
A similar set of contrasts involves terms such as “long” and “short,” “wide” and “narrow,” “deep” and “shallow,” “many” and “few.” In some languages “shallow” is “not deep,” but no language has only “shallow” and “not shallow”; the same holds for all the other contrasting terms just mentioned. The former, unmarked term in each of these sets is the “neutral” form, the one that can stand for the other. Thus in English we normally ask “how deep,” “how many,” etc., not “how shallow,” “how few,” etc. So, too, in other languages, though nothing logically precludes their speakers from doing just the opposite.
Of particular interest to anthropologists is Greenberg’s suggestion that the common tendency for peoples to call themselves by the word for “people” is not necessarily ethnocentrism—i.e., is not a claim that only they are really people (which is the standard anthropological interpretation of this usage)—but rather is just another kind of marking. Thus the people called the Maidu Indians call themselves majdy, which is their word for “people.” But they call blacks pibutim majdy, whites wolem majdy, etc. Thus majdy is an unmarked term which serves at different levels in the hierarchy of labels for peoples of all sorts.
Of even more interest to anthropologists are the semantic universals in kinship terminology (some of his discussion of this topic is updated in Greenberg 1979). Greenberg points out that all languages use different terms for “father” and “mother.” Each might be merged with other kin (e.g., fathers with uncles, mothers with aunts), but in no language is it obligatory to refer to one’s parents with terms that merge them. All kinship terminologies employ at least two specific semantic components in distinguishing kin: generation and sex.
In English, lineal kin terms are unmarked in contrast with the marked collateral kin terms; consanguineal kin terms are unmarked in contrast with affinal kin terms, which are marked. This appears to be a universal pattern, in that wherever these two contrasts are found the lineals in the one case and the consanguineals in the other are unmarked. Greenberg notes that a great many further universals can be found in kin terminologies.
The significance of marking lies primarily in what correlates with it and thereby suggests the underlying factors producing it. This will be discussed in the next chapter.
Anthropologists with linguistic interests, or who employed linguistic methods, were particularly sensitive to these developments in linguistics, and from their initial relativistic stances were led to universalistic conclusions. The findings of componential analysis, which were already mentioned, were specially relevant. We have also already examined one of the most famous cases of linguistics-inspired research: Berlin and Kay’s (1969) discovery that basic color terms develop in a universal sequence.
One of the most important general discussions of universals after Kluckhohn’s was Ward Goodenough’s (1970) on the role that universals play in anthropological description and comparison. As previously noted, Goodenough was one of the founders of componential analysis, and his thought is clearly indebted to linguistic models. Goodenough’s position is that there are (at least) two basic elements in anthropological description and comparison: the rights and duties of individuals or persons, and the “problems with which all societies have to deal” (1970:38). His discussion also draws attention to the anthropological usefulness of universally valid definitions.
Goodenough argues, for example, that a cross-culturally valid and universally applicable definition of marriage can be formulated (the matter was hotly debated in the fifties and sixties and simmers still; see, e.g., di Leonardo 1979; Sperber 1986). Such definitions are useful because they allow the greatest possible scope for comparison and generalization without precluding narrower analyses for particular purposes. But his definition of marriage is more than merely useful: it borders on explanation. The definition focuses on the regulation of sexual access to women eligible to bear children, and he argues that this regulation is a response to universals of human nature, including male dominance and male competitiveness for access to females (1970:11, 38).
Goodenough links the search for cross-culturally valid (etic) concepts to universals and to what he sees as the grand aim of anthropology: once we have ascertained all the etic concepts that are required to make sense of “the elementary emic units of any culture,” we may then abstract from those etic concepts an empirically determined list of the “universal attributes of culture and, by inference from them, the universal attributes of men as creators and users of cultures” (1970:129-30). This for Goodenough is the foremost aim of scientific anthropology.
A year after the publication of Goodenough’s book, an extensive exploration of the link between universals and human biology appeared in Lionel Tiger and Robin Fox’s The Imperial Animal (1971), to which an important precursor had been Tiger’s Men in Groups (1969). The latter argued that the need of males to bond with each other is a human universal that is rooted in a human biology shaped by an adaptation to hunting.
Identifying an inadequate understanding of human nature as the “most serious failing of social science” (1971:2), Tiger and Fox drew inspiration from evolutionary theory, studies of animal behavior (ethology), the fossil record, and anthropological studies to construct a human “biogrammar,” consisting of “those elements of human behavior that are the lexicon of social action” (1971:7). Universals were important to their argument both to construct the biogrammar and as evidence of its existence. Tiger and Fox argued that if an experimental Adam and Eve could somehow be raised apart from human culture their descendants within a few generations would have societies and cultures that replicated the universal pattern—because the pattern is in our nature.
Both in their joint work (Tiger and Fox 1971) and in a series of his own publications, Fox (e.g., 1967, 1971, 1980, 1989) also argued that the imp
ortant universals are not at the “substantive” level, where anthropologists usually seek them, but at the level of “process.” Processes may be universal even though their results are highly variable. The universal process of reproduction, for example, may or may not give rise to “families” as we understand this term.
All published in the space of three years, the books of Berlin and Kay (1969), Goodenough (1970), and Tiger and Fox (1971) marked a resurgent interest in universals that persists without break to the present. But while universals most certainly were not ignored in the 1970s and 1980s, neither did they inform the bulk of anthropological writings in those years. Although no one has taken a head count, I suspect that most anthropologists—still under the influence of the suppositions that animate Mead’s essay on Samoa—have been, and still are, wary of the very concept of universals.15
In a lengthy discussion of the issues that universals raise, Ronald P. Rohner (1975) drew attention to this wariness on the part of social scientists in general. For many of them, he says, the issues are “slightly ‘indecent’” (1975:165). Rohner surveys the developments, particularly in ethology and linguistics, that were leading some anthropologists and others to look for universals, rethink the problem of human nature, and question extreme forms of relativism. Rohner spells out a “universalist approach” (1975:1–38), which combines psychological research with the community studies and cross-cultural surveys of anthropology, and he applies this approach to the specific problem of determining uniformities in how children everywhere respond to parental acceptance and rejection.