The comparative method has taken many forms since Augustus Comte first employed the concept in 1853 in his foundational Cours de philosophie positive. Subsequently a variety of comparative methods have emerged in the social sciences with different goals, units of comparison, and types of data that reflect a variety of theoretical assumptions and interests. Comparison has formed the core of anthropology, sociology and other social sciences, to the extent that Emile Durkheim (1938) viewed all sociological analysis as necessarily comparative. Comparative methods have been employed for both quantitative and qualitative studies of such diverse phenomena as language, political organization, economic relations, religion, myth, kinship, marriage, and the family.
Three strategies are used in comparative methodologies: illustrative comparison, complete or universe comparison, and sampled-based comparisons (Sarana 1975). They are distinguished by the units of comparison (including cultures, societies, regions, or communities) and the particular items or features used to compare the units. Societies as units can be compared by examining items or traits such as institutions or practices. Illustrative comparison is the most common form of comparative analysis and has been employed extensively by theorists from diverse camps. Items are used as examples to explain or exemplify phenomena found in different units. They are chosen for their illustrative value and not systematically selected to be statistically representative. Illustrative comparisons are used in historical reconstructions, and to support interpretations or general assertions. Ethnographic case studies are commonly justified as the source for illustrative comparisons.
The second strategy is complete or universe comparison, in which all elements of the domain within the study, defined geographically (e.g., global or regional) or topically (e.g., analytical concepts or institutions), form the units of comparison. Comprehensive regional ethnographic surveys and analyses of particular topics, such as the national population health indicators of the World Health Organization reports, employ this approach.
Finally, sampled comparison strategically delimits part of the whole, with the goal of selecting data that are statistically representative of the variations within the whole and are intended as the basis for statistical generalizations. While studies of this type abound in sociology and human geography, they are much less common in anthropology. Within anthropology, the most widely known example is the George Murdock's Human Relations Area Files.
General Strategies of Comparative Methods
Comparative methods have been used for three types of goals: the construction of inferential histories, the development of typologies, and the explication of generalized processes (Peel 1987). Theories based on inferential histories dominated the formation of the social sciences until the early twentieth century, while the development of typologies and analyses of processes are now the predominate comparative strategies.
Natural histories of society. During the late nineteenth and early twentieth centuries scholars compared institutions and practices from many societies to construct evolutionary accounts of the origin of civilization, culture, and society. Contemporary primitive societies gave these theorists evidence of earlier social forms. Following the natural sciences' histories of geological formations and biological evolution, widely influential theorists, including Comte, Friedrich Engles (1965 ), Lewis Henry Morgan (1870, 1877), Karl Marx, Herbert Spencer (1898), Max Mueller (1909), James Frazer (1907), and Edward Tylor (1889, 1903), each constructed an historical narrative that traced the emergence of human civilization from ancient, primitive societies into complex and sophisticated civilizations of Europe. They proposed a variety of developmental stages, with characteristic types of social organization, economic activity, and religious practices, that all societies necessarily passed though during their evolution. They shared the belief that the nuclear family was the precursor of more complex forms of social relations such as the clan, tribe, city, and nation-state.
Multi-linear evolutionist and diffusionist theories. A second wave of historical comparativists followed in the early twentieth century. Rather than constructing a single history of human culture or civilization, these scholars attempted to explain the emergence of particular cultures and the historical diffusion of cultural traits. Commonalities and differences among cultures were explained as either independent inventions of social forms, artifacts, and beliefs, or taken to have diffused from a single point of origin. The several different schools of diffusionists preferred to believe that invention was infrequent, so consequently they developed comparative methods to infer relationships among cultural traits and infer their sources. British anthropologists Alfred Haddon (1895) and W. H. R. Rivers (1914) came to the conclusion, based on their research in Melanesia, that social change was the product of migration and culture contact. Taken further, G. Elliott Smith (1928) and W. J. Perry (1923) contended that Egypt was the root of Western European civilization and that culture diffused to ancient Europe as the result of culture contact and migration. A similar approach was developed in Germany and Austria under the tutelage of Fritz Graebner (1903) and Wilhelm Schmidt, who postulated the existence of Kulturekreise, culture centers, presumably in central Asia, from which archetypical cultural items were spread.
The German diffusionists' methodology and conclusions were inspired by the comparative method that linguists including William Jones (1799), Franz Bopp (1967 ), and Jakob Grimm (1967 ) used to identify historically related Romance and Germanic language families. These linguists inferred the previous existence of a common mother-language, Proto-Indo-European, from the systemic variation in sound systems among these languages and Sanskrit.
Criticisms of the historical comparative methods concern the units of analysis used for comparison including similarity and diversity among the societies studied, the comparability of the data used, and the kinds of generalizations that are possible given the nature of the data. Furthermore the inferential histories paid little heed to the contextual factors that molded the particular institutions that they examined.
The historical comparativists and the diffusionists' comparative methods and research suffered several weaknesses. They were unable to adequately respond to Francis Galton's criticism in the discussion that followed Tylor's address to the Royal Anthropological Institute (Tylor 1899) that, if data were gathered from neighboring groups, it would be impossible to determine if similarities resulted from a common history or arose independently from common functions. Questions were also raised about the ability to establish social rules based on historically contingent phenomena. In addition, the inferences they made were based on data that was often gathered unsystematically. Most significantly these theories seemed increasingly less credible as researchers had greater contact with people in the societies they attempted to explain. Diffusionist theories lost currency after World War II with the rise of theories designed to identify social laws rather than cultural origins. Though the diffusionists' theories were largely discredited as inadequately supported by historical data, the explorer Thor Heyerdahl (1952) kept them alive with his attempts to demonstrate the possibility of ancient transoceanic migrations.
Comparison and social laws. Three different approaches to comparative studies superceded the inferential histories of the evolutionists and diffusionists and established the parameters for anthropological and sociological comparison for the twentieth-century. The German-American anthropologist Franz Boas ( 1940) decried the "conjectural history" of the diffusionists' comparative method, in favor of comprehensive ethnographic descriptions that might reveal the "uniform laws that govern the human mind" (p. 271). Boas directed the efforts of the American Bureau of Ethnology to document the many cultures and languages of the native peoples of North America. His goal was to identify and classify the external (environmental) and internal (psychological) factors that shape the expression of these fundamental features of humans societies.
Durkheim's sociology echoed the analytical distinction between structure and process in Comte's positivist method. His goal was to identify structural forms or morphological units and their subtypes. He created a descriptive-analytic typology with analytical units that were examined synchronically for contextual variations. The goals of his sociological analysis were to identify social crucial facts that are elemental in every society and combined in different numbers and combinations into particular social species. He contended that "societies are only different combinations of the same original society" (Durkheim 1938, p. 86). In his studies of religion and social organization, he drew upon examples from Europe, North American native peoples, and Australia to identify elementary structures and their elaborations. Durkheim's study of social morphology laid the foundation for both British structure-functionalism in anthropology and Continental structuralist sociology and anthropology.
The failures of the conjectural histories of the diffusionists spurred a new and different approach to comparative studies in anthropology based primarily on Durkheim's social morphology and comparative sociology. British anthropologists A. R. Radcliffe-Brown (1951), Fred Eggan (1954), and Edward Evans-Pritchard (1963) severely criticized the historical comparativists and responded by developing more systematic, controlled comparisons that focused on systems of kinship, marriage and family.
Max Weber (1968) took a less positivist approach to social analysis and based his comparative method on the formulation of ideal types. He began with the recognition that the researcher plays an important role in framing research questions, identifying units of analysis, and selecting items for comparison. Rather than assuming an objective separation of the researcher and data, he constructed ideal types, or analytical models that did not confuse the researcher's conceptualization of the phenomena with the phenomena itself. These types enabled him to investigate the phenomena from an acknowledged starting point and interrogate other aspects of the object during analysis. He employed ideal types in his comparative studies of the relationship between economy and religion in Protestant Northern and Catholic southern Europe, the differences between charismatic and bureaucratic forms of leadership in Europe and China, and religious practices in Europe, China, and India.
Clifford Geertz (1963, 1968) used ethnographic cases as real types for comparisons of social organization, economic systems, and educational systems, and paved the way for comparisons in interpretative anthropology and cultural studies.
A third response to the inadequacies of the historical comparative methods was to develop sample-based comparisons with ethnographic databases. George Murdock's Human Relations Area Files and accompanying Ethnographic Atlas were the most extensive attempt to identify cross-cultural correlations and make statistical generalizations (Murdock 1963; Murdock and Yale University Institute of Human Relations 1982). To this end, he cataloged existing ethnographic data from 10 percent of the world's cultures identified by the late 1930s. Murdock's approach floundered due to the difficulties of making correlations, identified by Galton, and its dependence upon existing data, gathered by others who did not use comparable research strategies or common definitions of phenomena.
Comparisons of processes. Comparative studies of social process have returned to some topics previously examined by classical evolutionists and the diffusionists, but with much more constraint and caution. Research on social and economic change, migration, and cultural contact have attempted to return a historical dimension to structural analyses. Edmund Leach's (1954) study of the dynamics of ethnic and political relations in highland Burma paved the way for the more complex formulations in the French sociologist Pierre Bourdieu's (1977) theory of social practice, and in Ulf Hannerz's (1992) analysis of creolization, or the synthesis of new cultural forms, under the pressures of culture contact and globalization.
Comparative Methods in the Study of Kinship, the Family, and Marriage
Kinship and family relations were early subjects of comparison and debate in the social sciences. Studies of kinship and the family have formed the core of British social anthropology and have dominated North American and European anthropology throughout the twentieth century. Family and kinship were central to the nineteenth- and early twentieth-century debates about the origins and evolution of society. Henry Maine (1861 ), James McLennan (1865, 1886), and Johann Bachofen (1967 ) examined forms of family and marriage. Maine compared Greek, Roman, and more contemporary British and continental family law. Bachofen, confusing matrilineality as matriarchy, argued that social authority originally developed from mothers' roles in primitive families that were transformed during cultural development into male authority in patriarchies. McLennan traced social evolution though changes in forms of marriage, from primitive promiscuity though marriage by capture and eventually the monogamous marital relationships of Victorian England.
L. H. Morgan, a U.S. lawyer, is considered the father of kinship studies in anthropology, however. He described the legal or jural dimensions of family and kinship among the Iroquois of the state of New York, and compared their family and clan structures with those of European societies and Australian Aborigines (who have figured significantly in comparative studies of kinship) (Morgan 1870, 1963 ). From his analysis of kinship, Morgan developed a theory of evolution in which the division of labor within the family was the basis for the development of more complex forms of social organization including the nation-state. Another enduring contribution was to distinguish between kin terms used to describe and classify individuals. This opened the door to the use of kin terms as the basis for comparisons of kinship terms as cultural systems of classification.
Morgan's evolutionary schema had a marked impact upon another social theorist, Karl Marx. Though Marx initially replaced Morgan's focus on the family with private property in his social and economic analysis, Marx and Frederick Engels returned to the centrality of the family in their discussion of the origin of private property (Engels 1988 ). Studies of kinship and the family took second place in diffusionist theories to explanations of the transmission of material culture, particularly technology and religious beliefs.
During the later half the twentieth century, comparative studies of kinship dominated anthropology. They were of three types, each closely aligned with the theories of Boas, Durkheim, and Weber, and concerned with social structure rather than history. The first is the controlled case study approach recommended by Radcliffe-Brown and Forde (1950) and Evans-Pritchard (1963). These comparative studies of social forms focused on kinship and marriage and the structural relationships among kin groups. They compared societies' rules concerning the rights and obligations that established group membership, inheritance, and succession. They described them with terms they believed were universal features of kinship and family: descent, generation, gender, collaterality (or siblingship), and marital relations. Their units of study were the nuclear family, the lineage, and the clan. They reduced the variability among their comparative units by concentrating their research on regions of Africa with patrilineally and matrilineally based societies. Social organizations were classified by the rules of group membership, inheritance patterns, laws of succession, and patterns of prohibited and preferred marriage and post marital residence.
British structural-functionalist analyses concentrated attention on kinship to the expense of the family, many contending that lineage and clan relations were the logical and psychological extension of ties among nuclear family members. These anthropological analyses of the structures of family and kinship relations were similar to the functionalist analyses of families and family structures that developed sociology. Comparative sociologists examined the functions and structural attributes of families, household composition, and family dynamics as did anthropological studies of the time. In addition they considered more emotional and psychological issues such as love (Goode 1959). Comparisons by sociologists focused on variations across time and national, ethnic, and class lines, rather than across cultures.
Claude Levi-Strauss developed another method based on the comparison of structural principles. His structuralist treatment of kinship and marriage (referred to as alliance theory) examined the nature of relationships among groups, rather then focusing upon groups' rules of composition. Levi-Straus's seminal Elementary Structures of Kinship 1969) began by examining the significance of incest rules and rules of group exogamy (the practice of marrying outside of one's group) that used marriage as a means of both delineating group boundaries (in terms of those whom one may or may not marry) and establishing alliances. From this starting point, he compared the complex patterns of marriage-based alliances among a number of Australian aboriginal groups and societies in Southeast Asia and India, to compare the various conceptual elaborations of the principles of marriage exchange and alliance.
During the 1960s and 1970s comparative studies declined, in part due to methodological and epistemological debates that questioned the concepts employed in comparative research. Studies of kinship and the family were at the heart of these debates. Questions were raised about the nature of analytical definitions and the use of Western European concepts such as descent, marriage, and kinship as analytical constructs for the description and analysis of systems in other cultures (Needham 1971). Examination of other cultures' theories of conception and paternity even called into question the very nature of kinship and its recognition as a universal phenomena. David Schneider (1968) contended that kinship systems were culturally constructed idioms of social relations. Nevertheless, comparative studies of kinship terminologies continued to use Western concepts such as descent as analytical concepts in comparisons of kinship semantics and the cognitive classifications of kin (Tyler 1969). Consequently, Leach (1966) raised serious doubts about the value of the typologies developed to describe the kinship systems. These questions further undermined the already weak reception for statistical studies such as those of Murdock.Networks and Process. Anthropologists also became increasing concerned about variation within the kinds of social units that they had previously used in comparisons. Case studies that were the staple of the method of controlled comparison of British structure-functionalists and Levi-Straussian structuralism treated families, clans, societies, and cultures as closed systems. Migration by members of formerly isolated societies forced researchers to face growing diversity and the disjunction of features—language, common history, religious beliefs and practice—that had coincided in geographically bound populations. Studies of networks and their structures attempted to overcome the restrictions of geographically defined analytical units (Sanjek 1978). The development of network theory and formal models such as directed graphs provided researchers with new ways to describe and compare families structures and systems of kinship and marriage (Hage and Harary 1996), kin terms, (Schweizer and White 1998), and ties between household and family members and their communities (Wellman and Berkowitz 1997).
Not only were classical comparative studies called into question on epistemological grounds, their adequacy in representing kinship and family systems was attacked for their substantive limitations grounds. Earlier studies had focused on the legal and political aspects of kinship that were dominated by men. Feminist critics argued that they generally ignored women and the domestic sphere, thereby undermining the adequacy of earlier conventional studies. This criticism reinvigorated comparative studies of the family, women's roles, socialization, and gender relations (Yanagisako 1979) that found antecedents in the early comparative work of Boas's student, Margaret Mead (Mead  2001; Mead and Malinowski  2001). The reconsideration of the role of women, the family, and socialization also coincided with Bourdieu's attention to the processes of social reproduction (Bourdieu and Passeron 1977).
Comparative analyses remain an essential aspect of anthropology and other social sciences, just as Durkheim asserted (1938). With the growth of literacy and political activism, the peoples who anthropologists had studied and described have challenged professional social scientists' place as ethnographers. At the same time, anthropologists, sociologists, and cultural geographers' comparative analyses take on greater academic significance and practical value (Sperber 1985).
bachofen, j. j. (1967 ). myth, religion, and motherright: selected writings of j. j. bachofen. princeton, nj: princeton university press.
boas, f. (1940 ). race, language and culture. newyork: macmillan.
bopp, f. (1967 ). "on the conjugational system of the sanskrit language: in comparison with that of greek, latin, persian and the germanic languages." in a reader in nineteenth century historical indo-european linguistics, ed. w. p. lehmann. bloomington: indiana university press.
bourdieu, p. (1977). outline of a theory of practice. cambridge, uk: cambridge university press.
bourdieu, p., and passeron, p. (1977). reproduction ineducation, society, and culture. london: sage.
comte, a. (1853). positive philosophy of auguste comte.london: j. chapman.
durkheim, e. (1938). the rules of sociological method, 8th edition. glencoe: the free press.
eggan, f. (1954). "social anthropology and the method ofcontrolled comparison." american anthropologist 56:643–763.
engles, f. (1988 ). the origin of the family, privateproperty, and the state. 1st edition. new york: pathfinder press.
engels, f., and leacock, e. b. (1972). the origin of thefamily, private property, and the state, in the light of the researches of lewis h. morgan. 1st edition. new york: international publishers.
evans-pritchard, e. e. (1963). the comparative method insocial anthropology. london: university of london, athlone press.
frazer, j. g. (1907). the golden bough: a study in magic and religion. 3rd edition. london and new york: macmillan.
geertz, c. (1963). peddlers and princes: social change and economic modernization in two indonesian towns. chicago: university of chicago press.
geertz, c. (1968). islam observed: religious development in morocco and indonesia. new haven, ct: yale university press.
goode, w. j. (1959). "the sociology of the family: horizons in family theory." in sociology today; problemsand prospects, ed. r. k. merton, l. broom and l. cottrell. new york: basic books.
graebner, f. (1903). "kulturekreise und kultureschichten in ozeania." zeitschrift fur ethnologie 37:28–53.
grimm, j. (1967 ). "germanic grammar." in areader in nineteenth century historical indo-european linguistics, ed. w. p. lehmann. bloomington: indiana university press.
haddon, a. c. (1895). evolution in art: as illustrated by the life-histories of designs. london: w. scott.
hage, p., and harary, f. (1996). island networks: communication, kinship, and classification structures in oceania. cambridge, uk: cambridge university press.
hannerz, u. (1992). cultural complexity: studies in thesocial organization of meaning. new york: columbia university press.
heyerdahl, t. (1952). american indians in the pacific:the theory behind the kon-tiki expedition. london: allen and unwin.
holy, l. (1987). comparative anthropology. oxford, uk, and new york: blackwell.
jones. w. (1799). works of sir william jones. london: g.g. and j. robinson.
leach, e. r. (1954). political systems of highland burma:a study of kachin social structure. london: london school of economics and political science.
leach, e. r. (1966). rethinking anthropology. london:athlone press.
lehmann, w. p. (1967). a reader in nineteenth centuryhistorical indo-european linguistics. bloomington: indiana university press.
levi-strauss, c. (1969). elementary structures of kinship, rev. edition. boston: beacon press.
mclennan, j. f. (1865). primitive marriage. edinburgh:adam and charles black.
mclennan, j. f. (1886). studies in ancient history: comprising a reprint of primitive marriage. london and new york: macmillan.
maine, h. s. (1911 ) ancient law, its connection with the early history of society and its relation to modern ideas. london: murray.
mead, m. ( 2001). sex and temperament in threeprimitive societies. new york: harpercollins.
mead, m., and b. malinowski. ( 2001). growing up in new guinea: a comparative study of primitive education. new york: harpercollins.
morgan, l. h. (1870). systems of consanguinity and affinity of the human family. washington, dc: smithsonian institution.
morgan, l. h. (1963). ancient society; or, researches in thelines of human progress from savagery through barbarism to civilization. cleveland, oh: world pub. co.
müller, f. m. (1909 ). comparative mythology. london: g. routledge and sons.
murdock, g. p. (1963). outline of world cultures. 3rd rev. edition. new haven, ct: human relations area files.
murdock, g. p., and yale university institute of humanrelations. (1982). outline of cultural materials. 5th rev. edition. new haven, ct: human relations area files.
needham, r. (1971). introduction to rethinking kinship and marriage. london: tavistock.
peel, j. d. y. (1987). "history, culture and the comparative method." in comparative anthropology, edited by l. holy. oxford, uk and new york: blackwell.
perry, w. j. (1923). the children of the sun. london:metheun.
radcliffe-brown, a. r. (1951). the comparative method insocial anthropology. journal of the royal anthropological institute 81:15–22.
radcliffe-brown, a. r., and c. d. forde. (1950). africansystems of kinship and marriage. london and new york: published for the international african institute by the oxford university press.
rivers, w. h. r. (1914). the history of melanesian society.cambridge, uk: university press.
sanjek, r. (1978). "what is network analysis, and what isit good for?" annual review of anthropology 1:588–597.
sarana, g. (1975). methodology of anthropological comparisons. tucson: university of arizona press.
schneider, d. m. (1968). american kinship: a cultural account. englewood cliffs, nj: prentice-hall.
schweizer, t., and d. r. white. (1998). kinship, networks, and exchange. new york: cambridge university press.
smelser, n. j. (1976). comparative methods in the socialsciences. englewood cliffs, nj: prentice-hall.
smith, g. e. (1928). in the beginning: the origin of civilization. new york: morrow.
spencer, h. (2002 ). the principles of sociology. newbrunswick, nj: transaction publishers.
sperber, d. (1985). on anthropological knowledge: threeessays. cambridge, uk: cambridge university press.
tyler, s., ed. (1969). cognitive anthropology. new york:holt, rinehart and winston.
tylor, s. e. b. (1889). "on a method of investigating thedevelopment of institutions: applied to laws of marriage and descent." journal of the royal anthropolocial institute 18:245–269.
tylor, s. e. b. (1903). primitive culture: researches into the development of mythology, philosophy, religion, language, art, and custom. 4th revised edition. london: j. murray.
weber, m. (1968). economy and society: an outline of interpretive sociology. new york: bedminster press.
wellman, b., and berkowitz, s. d.. (1997). social structures : a network approach. greenwich, ct: jai press.
yanagisako, s. j. (1979). "family and household: theanalysis of domestic groups." annual review of anthropology 8:161–205.
william h. mckellin