4 Cultural Materialism Is Alive and Well and Won’t Go Away Until Something Better Comes Along MARVIN HARRIS At the outset, I wish to disassociate myself from the impression, sometimes carelessly and sometimes deliberately conveyed, that anthro~ pology can be equated with cultaral anchro- pology, or much less, with eshnography.t Cul- tural zathropologists who have carried out ethnographic studies often rely heavily on his- toric, ethnohistoric, and archaeological mater- ials for the fulfilment of their mature profes sional and intellectwal interests (e.g, Sablins, Murra, Wolf, Mintz). Moreover, many influ- ential figures in azchacology (e.g. J. Marcus, Flannery, Hodder, Binford, Sanders} have carried out ethnographic studies on their own in order to enhance their understanding of archaic societies. In addition, ficld studies of nonhuman primates (e.g. by Devore, Van Lawick-Goodall, Imanishi, Teleki) provide an indispensable data base for understanding the emergence of distinctively human cultures. Studies of nonhuman primate sex roles in par- cular are responsible for major advances in our understanding of the evolution of human gender hierarchies. Productive scholarly inver- action involving archaeology, paleontology, and ethnography takes place among many anthropologists who identify their principal interest as medical anthropology or biocultural anthropology. Nor should we forget that applied anthropologists who do surveys, write impact statements, and evaluate development programs outnumber university-based cthnog- raphers. Surely postmodernists must have 62 heard it rumored that the membership of the American Anthropological Association does not consist exclusively of ethnographers. Cultural materialism is a paradigm whose principles are relevant to the conduct of re- search and the development of theory in vir~ tually all of the Gelds and subfields of anthro- polopy. Indeed, it has been guesstimated {Thomas 1989:115) that half of the archacolo- sists in the United States consider themselves to be cultura! materialist to some degree. For cultural materialists, whether they be cultural anthropologists, archaeologists, biological an- thropologists, or Linguists, the central intellec~ tual experience of anthropology is not ethnog~ raphy but the exchange of data and theories among different ficlds and subfields concerned with the global, comparative, diachronic, and synchronic study of humankind: the origin of the hominids, the emergence of language and culture, the evolution of cultural differences and similarities, and the ways in which biocul- tural, mental, behavioral, demographic, envi- ronmental and other nomothetic processes have shaped and continue to shape the human. world. CULTURE Tt is no accident that all of the contributors to this volume who implicitly or explicitly ignore the biological and archaeological facets ofanthropology operate with a definition of culture that is narrowly confined to mental and emic phenomena (see section on episte- mological principles below). Culture for them constitutes a realm of pure idea which is accessible only through interactive discourse between live ethnographers and live “natives” (gee Harris 1980 for the history of the defini- tion of culture as pure idea). This unfortunate bias inevitably isolates the study of humans in the present from humans who have lived in the past and wito have felt no written records. For how can one carry out interactive dis- course with the dead? Crucial to any effort to maintain the nkages among cultural anthro- pology, archaeology, and biological anthro- pology, therefore, is 2 concept of culture that embraces not only the mental and emic (see below) components of human social life but the etic and behavioral components 2s well. ‘The culture in cultural materialism refers to the socially conditioned repertories of activ. ities and thoughts that are associated with particular social groups or populations. This definition of culture stands opposed to the Gxed, “essentialist” notions that inspire those who define culture as a realm of pure and uni- form ideas hovering over the hub-bub of the daily life of specific individuals. For cultural materialists, culture elements aze constructed (more specifically, abstracted) from the bed- rock of the immensely variable thoughis and behavior of specific individuals (Harris 19642). In complete accord with Borofiky’s emphasis upon individual variability in this volume, cultural materialists have long argued that cul- ture is at bottom an unfolding material process (iz. the concept of “behavior stream”) rather than an emanation of a platonic archetype (see the discussion of variability and ambiguity in Brazilian color-tace categories—Harris 1964b, 1970; Kotak 1967). Yet, it would be com- pletely self-defeating to limit the definition of culture and the scope of the social sciences (as Vayda’s chapter scems to propose) to the bedrock of individual thought and activity. Although we cannot see or touch entities stich MARVIN HARRIS 63 as a mode of production or a transnational corporation or a sociocultural system, to the extent that these are logical and empirical abstractions built up out of the observation of individual-level events, they possess a reality that is not inferior to any other reality. Indeed, it is imperative for human survival and well- being that we learn to rise zbove individual thoughts and actions to the level at which we can begin to examine the aggregate effects of social life and the behavior of such higher- order entities as institutions and whole sociocultural systems. Political economies are as teal as the individuals who fall under their sway, and a lot more powerful. PARADIGMS Paradigms stipulate the principles which govern the conduct of research. Principles fall into two classes: rules for acquieing, testing, and validating knowledge (Le., epistemologi- cal principles) and rules for generating and evaluating theories ((.e., theoretical principles). A widely misunderstood aspect of scientific paradigms is that neither the epistemological ‘or theoretical principles nor the paradigm as 2 whole has the status of a scientific theory. Principles such as creationism, natural selec- tion, or the priority of infrastructure are not falsifiable. This does not mean however that paradigms are “ships that pass in the night.” Paradignns can be compared with cach other and evaluated from two standpoints: (1) their logical structure and internal coherence and (2) thetr respective abilities to produce scien- tific theories in conformity with the criteria discussed below. From this vantage point, the alternatives to cultural materialism presented in this volume offer slight hope of safe passage. Tee a lot of sunken ships in the muddy waters of post-postmodernisin—sbips built out of flawed accounts of the history of anthropolog- ical theory, parochial agendas, inchoate con- ceptions of the nature of human society and human caltures, and a lack of well-formed64 ASSESSING CULTURAL ANTHROPOLOGY epistemological and theoretical principles or useful substantive achievements that might justify a furure—any furure—for anthropol- ogy. EPISTEMOLOGICAL PRINCIPLES: SCIENCE Cukural materialism is based on certain episte- mological principles which are held in com- mon by all disciplines which claim to have scientific knowledge. Scientific knowledge is obtained by public, replicable operations (observations and logical transformations). The aim of scientific research is to formulate explanatory theories which are (1) predictive (or retrodictive), (2) testable (or falsifiable), @) parsimonious, (4) of broad scope, and () integratable or cumulative within a coherent and expanding corpus of theories. The same criteria distinguish scientific theories which are more acceptable from those which are less acceptable. Scientific theories find acceptance in accordance with their relax tive powers of predictability, testability, parsi- mony, scope, and integratabibity as compared with rival theories about the same phenom- ena. Since one can only approach, but never completely reach, perfection in this regard, scientific theories are held as tentative approx- imations, never as “facts.” This view of science derives from the logical positivist and empiricist philosophical tradi- tions, It might be labeled neopositivism since it embodies and surmounts the critiques made by Popper (1965), Lakatos (1970), and Kubn (1970, 1977). Note that it makes no claim to being “value fee.” Rather it proposes to overcome the inevitable biases of all forms of knowledge by methodological rules that insist upon opening to public scrutiny the opera~ tions by which particular facts and theories come to be constructed. The oft-repeated charge by postmodemist science-bashers that there is no community of observers who can or do scrutinize anthropological, especially ethnographic, operations (see discussion of Tyler 1986 below) is belied by the intense criticisms to which crucial facts and theories are regularly subjected in the pages of anthro- pology’s principal journals. Challenges by other observers to the ethnographic accuracy of the work of Boas, Mead, Benedict, Redfield, Evans-Pritchard, Malinowski, Lee, Vayda, and Chagnon just for starters, whether based on fresh fieldwork or written sources, Clearly do fatfill the scientific model for inde~ pendent testing by other observers. This is one Of the few points of disagreement that I have with Salzman’s paper, which answers “no” to whether ethnographic research involves a community of researchers at this time (page 38). It may take awhile, but ethnographers working in the same region if not the same village do help to keep cach other in touch with basic ethnographic facts. However, certainly agree with Salzman that the future of ethnography Lies in greatly expanding the use of field teams and the number of restudies rather than, as Marcus proposes in his contri bution to this volume, increasing the number of experimental, personalistic, and idiosynera- tic field studies carried out by untrained would-be novelists and ego-tripping narcissists afflicted with congenital logo-diarrhea. ‘The claim is often made that even the natural sciences have had to abandon “objec tivity” and determinism (because of Heisen- berg’s quantum indeterminacy or because of chaos theory). The idea that objectivity is no longer an issue in the physical, chemical, and biological sciences runs afoul of what it is that several million researchers worldwide actually do to earn their living. Let our anthropologi- cal science-bashers get up and tell an audience of the sick that there is no objectively valid treatment for AIDS or leukemia and that there never will be one; or tell 2 group of physicists that ie is impossible objectively to decide whether cold fasion occurs or not; or tell the NIH panels on scientific fraud not to worry because ail scientific data are equally cooked, subjective, and culturally “constructed.” Thereason that cultural materialists favor know!- edge produced in conformity with the episte- mological principles of science is not because science guarantees absolute truth free of sub- jective bias, error, untraths, lies, and frauds, It is because science is the best system yet devised for reducing subjective bias, error, un axuths, lies, and frauds. As for determinism the assertion that phenomena are causally determined by certain events or principles—its death is greatly exaggerated. In the social sci- ences nineteenth-century formulations of absolute laws were long ago qualified by the realization that science does not yield cer~ tainties and laws, only probabilities and gen- eralizations, Incidentally, chaos theory does not Jead t a renunciation of deterministic systems but to the extension of probabilistic determinism to realms of phenomena (such a5 hydrodynamic turbulence} which hitherto seemed to be entirely unpredictable, Following the lead of Clifford Geertz and under the direct influence of postmodern philosophers and literary critics such as Paul De Man, Jacques Derrida, and Michel Foucault, interpretationist anthropologists have adopted an increasingly arrogant and intolerant rhetoric aimed at ridding anthro pology of all vestiges of scientific “totalizing” paradigms, According to Stephen Tyler, for example, sociocultural anthropologists should abandon the inappropriate mode of scientific rhetoric that entails “objects,” “facts,” “descriptions,” “induc tions,” “generalizations,” “verifiation,”” “experi ment,” “inuth,” and tke concepts that, except as empty invocations, have no pancilels either ia the experience of ethnographic fieldwork or in the writing of ethrographies. The urge to conform to the canons of scientific rhetoric has made the easy realism of natural history the dominant mode of ethnographic prose, but it has been an illusory realism, promoting, on the one hand, the absurdity of “describing” nonentities such as “culture” or ‘soiety” as if they were fully observable, though somewhat ungainly, bugs, and, on the other, the MARVIN HARRIS 65 equally ridiculous behaviorist pretense of “describ- ing” repetitive pattems in isolation from the discourse that actors use in constituting and situating their action, and ail in simpleminded surety that the observers’ grounding discourse is itself an objective form sufficient to ihe task of describing acts. (1.986:130) Tyler’s totalizing renunciation of the search for objects, facts, descriptions, inductions, generalizations, verification, experiment, trath, and “like concepts” (!) in human affairs mocks itself so effectively that any attempt at rebuttal would be anticlimactic. 1 do think it may be aseful, however, to point out that the “simpleminded surety” with which positivists and behaviorists are alleged to view human social life flagrantly distorts the entite history of science in general, during which all sureties, simpleminded or not, have been subject to relentless skepticism, and the history of logical positivism in particular, during which the struggle to create objective data languages has constizuted the central focus of a vast and continuing philosophical effort. Anthropology’s dedicated science-bashers are not mollifed by the assurance that cultural materialists seek probabilities rather than certainties, generalizations rather than laws. Shanks and Tilley (1987:38), for example, question the validity of making any kind of genenization. They ask how general 2 state ment must be before it counts 2s 2 generali- zation: “two cases? three? fifty?” They also ask: If the generalizations made are not laws, they cannot be expected to be applicable irs any particular case so why are these generalizations of use to us? Why must the business of doing science necessarily be equated with the ability, or the will to gen- eralize? This appears to be a procedural rule founded on the basis thet generalizing, rather than ‘considering all the particularity of the individual case, i5 a superior kind of activity. There seems to be no compelling reason why we should accept this. (1987:38)| | | 66 ASSESSING CULTURAL ANTHROPOLOGY Questions and Answers ‘The fallacies that embolden these queries are so transparent that one mast wonder if the in terlocutors really intend to be taken seriously. (This is 3 serious concer on my part since Derrida and his followers are not above cele- brating the playful consequences of decon- structionism.) Yet given the current popularity of antiscientism, their questions, serious or not, cannot be left unanswered. Question: just how often does something have to recur in order for it to sexve 28 the basis for a generalization? Answet: The more times the better. Question: If generalizations cannot be ex- pected to be applicable to any specific case, what good are they? Answer! The better the generalization, the more probable its applicability to the particular case, the more usefll is, (lt is definitely use- ful to know that a particular person who smokes four packs of cigarettes a day is ten times more likely to get lung cancer than one who doesn’t smoke, even though not all heavy smokers get lung cancer.) Question: Why must science be equated with generalizing? Answer: Because science is by definition a generalizing form of knowledge. Question: Is the mandate to generalize noth- ing but a “procedural rule”? Answer: Of course, And anyone is free to ignore the rule but to do so is to cease doing science. ([t is also likely to get you killed the next time you step off the curb against che light, or the next time you strike a match to look inside your gas tank.) Last question: Instead of generalizing, why not consider “all the particularity of the individual case"? Answer: Because there are no limits to par ticularity. Any project shat proposes to deliver ail the particularities of any macrophysical event, human or not human, therefore makes 4 preposterous claim on our time and re- sources. For this reason, in science endless particularity is the exact equivalent of endless ignorance, EPISTEMOLOGICAL PRINCIPLES: EMICS AND ETICS In addition to the general epistemological principles shared with other scientific disci plines, cultural materialism i also based on epistemological principles which ae spcii to the sudy of human sociocultural systems, These involve: (1} the separation of mental events (thoughts) from behavior factions of body parts and their environmental effects) and (2) the separation of emic from etic views of thoughts and behavior. (See below for definitions of “emic” and “etic.") The reason for the epistemological distinction between mental and behavioral events is that the operations (observational procedures) used to obtain knowledge of mental events are categorically distinct from those needed to obtain knowledge of behavioral events. In the former, observers depend directly or indirectly on participants to communicate what is going on inside their heads; in the latter observers are not dependent on actors to identify the actor's body motions and the environmental effects of chose motions. The reason for the further distinction between emic and etic events is chat the separation of mental froma behavioral events does not exhaustively specify the epistemological status of the categories (data language) employed in the identification of mental of behavioral events. Observers have the option of describing both kinds of events in terms of categories that are defined, identified, and validated by the community of participants (emics) or by the community of observers (etics). Four types of knowledgestem from these distinctions: (I) emics of thought; (2) emics of behavior; (3) etics of behavior; (4) etics of thought. To illustrate, consider the practice of indirect infanticide in northeast Brazil: () A sample of economically apd socially deprived mothers condemns and abhors infanticide, (2) These mothers insist chat their own behavior has been devoted to sustaining the life of their infants, (3) Observers note, how~ ever, that some of these mothers actually withhold food and drink from certain infants, especially from infants that are first and last born. (4) On the basis of the observed occur- rence of maternal neglect and high infant mortality, it can be inferred that these dis advantaged women have thoughts that are contrary to or that modify their elicited emics of thought and behavior. (The foregoing has been adapted from Scheper-Hughes 1984, 1992.) Returning to participants for additional emic data may result in the elicitation of emic thought and emic behavior that correspond to the etic inferences. Emic and etic versions of social life are often but not necessarily contra dictory. (See Headland, Pike, and Harris 1990 for a full discussion of the history and significance of the emic/etic distinction.) But failure to distinguish between emic and eric and between mental and behavioral data renders much of the sociocultural liserature of cultural anthropology useless by Literally pre- venting researchers from understanding the referential significance of their descriptive discourse (Harris 1968; Marano 1982; Headland, Pike, and Harris 1990). Despite a persistent barrage of uninformed ‘or malicious assertions to the contrary, cultural ‘materialists insist that the proper study of humankind is both emics and etics and bot thought and behavior, This brings me to Robert Murphy's paper and his claim that materialists, myself in particular, define culture as “Ieamed and shared behavior” (italics in the Original). I do not know how to account for Murphy's failure to have assithilated the fact that I have persistendly and plainly stated, over 5 ee MARVIN HARRIS 67 and over again, that culture is both thought and action, both emics and etics, and that anthropology involves the study of both mind and body. Perhaps Murphy could not bring hirnself to admit that his main point—namely that the “antithetical relationship between [the head and the torso], and between Act and Concept, should be the very substance of our discipline” (page 59)—does in fact constitute the very substance of my own work (e.g., race relations in Brazil; colonialism in Mozam- dique; cattle in India; warfare in Amazonia; and reproductive controls ia preindustrial societies). While no cultural materialist has ever advo- cated making the subject matter of cultural anthropology exclusively etic or behavioral, the postmodernists and their idealist predeces- sors have relentlesly advocated essentialist exclusions with regard to what cultural anthropologists ought to study. How conven- ient it would be for these critics if cultural materialists were to reject the study of mind and emics just a5 the interpretationists reject, the study of activity and etics. (“Ah! You see, these vulgarians reject that which is most human about humans.) But words to that effect demonstrate at best an indomitable igtiorance of the anthropological literature. THEORETICAL PRINCIPLES: These rest on the assumption that certain categories of behavioral and mental responses are more directly important to the survival and well-being of human individuals than others and thar it is possible to measure the efficiency with which such responses contribute to the achievement of an individual's survival and well-being. ‘This assumption lies at the basis of the “costing” of akemative patterns of behav- ior which in tum is essential for identifying optimizing behavior and thought (see below) and the development of materialist theories of the causes of sociocultural differences and simnilarities.68 ASSESSING CULTURAL ANTHROPOLOGY The categories of responses whose costs and benefits underwrite cultural selection and cul- tural evolution are empirically derived from the biological and psychological sciences that deat with the genetically given needs, drives, aversions, and behavioral tendencies of Homo sopiens: sex, hunger, thirst, sleep, language acquisition, need for affective nurturance, nutritional and metabolic processes, vulner~ ability to mental and physical disease and to stress by darkness, cold, heat, altitude, moisture, lack of air, and other environmental hazards. ‘This list is obviously not intended to encapsulate the whole of human nature. It remains open-ended and responsive to new discoveries about the human biogram and population-specific genetic differences. This brings me to another endlesly repeated although endlessly refuted characterization of cultural materialism. In Murphy's words: “Cultural materialism. remains rooted in a mechanical, naturalistic mode that fails to reckon with the fact that the mind is fir more than a tabula rasa” (page 58). Without con- ceding that the minimum condition for sdvo~ cating a “mechanical” or “naturalistic” mode is adherence to the doctrine of a mental blank slate, cap one find such a slate incorporated into the theoretical principles of cultural mate- rialism? Impossible, because anyone who insists that humans have definite genetically determined biopsychological drives is obvi- ously saying that our minds are not blank at birth. Now the postmoderns and other critics of cultural materialism may not Eke what cul- tural materialism puts into the brain-mind at birth, but that is a different issue, I do not believe for example, as Murphy docs, that we are wired to think dislectally, although { am prepared to change my opinion if someone can provide me with some empirical evidence from the cognitive sciences or neurophysiol~ ogy. To continue: Various currencies can be used to measure the costs and benefits of behavior that have optimizing consequences such a8 morbidity and mortality rates, differential sex- ual access, monetary costs and benefits, encr- getic inputs and outputs, and nutritional inputs and ourputs. (The omission of currencies directly linked to differential reproductive success should be noted, for it encapsulates the basic difference between cultural materialist and sociobiological paradigms. See Harris 199% for a sustained critique of human socio- biology.) Infrastructure, Structure, and Superstructure ‘The components of social life which most directly mediate and facilitate the satisfaction of biogram needs, drives, aversions, and behavioral tendencies constitute the causal center of sociocultural systems. The burden of this mediation is bone by the conjunction of demographic, technological, economic, and ecological processes—the modes of produc- tion and reproduction—found in every soci cultural system. More precisely, it is the etic bekzavioral aspect of the demo-techno-econo- environmental conjuaction that is salient, and hence it would be more precise (but too cum- bersome) to define the causal center as the ete behavioral infastructue (or the etic behavioral modes of production and reproduction). Infra- stracture constitutes the interface between nature in the form of unalterable physical, chemical, biological, and psychological con straints on the one hand, and culture which is ‘Homo sapiens’s primary means of optimizing health and well-being, on the other. It is the unalterability of the laws of physics, chemistry, biology, and psychology therefore that gives infrastructure its inal strategic priority in the formulation of cultural materialist theories. Culeural optimizations and adaptations must in the first and last instance conform to the re- straints and opportunities of the environment and of human nature. In addition to infrastructure, every human sociocultural system consists of two other major subsystems: structure and superstructure, each with its mental/behavioral and emic/cticaspects. Structure denotes the domestic and political subsystems, while superstructure Genotes the realm of values, aesthetics, rules, belieés, symbols, rituals, religions, philosophies, and other forms of knowledge including science itself. ‘The basic theoretical principles of cultural materialism can now be stated: (1) optimiza tons of the cost/benefits of satisfying biogeam needs probabilistically (ic. with more than chance significance) determine (or select for) changes in the etic behavioral infrastructure; (2) changes in the etic behavioral infra structure probabilistically select for changes in the rest of the sociocultural system. The com-~ bination of 1 and 2 is the principle of the primacy of infrastructure? ‘As a guide to theory-making, the primacy of infrastructure enjoins anthropological re- searchers concerned with the explanation of sociocultural differences and similarities to concentrate on and to give priority to the formulation of hypotheses and theories in which components of the etic behavioral infrastructure are treated as independent vari- ables while components of structure and superstructure are treated as dependent vari- ables. The practical consequence of such a commitment of research effort is that the search for causal infrastructural variables will be conducted with decisively greater pessist- ence and in greater detail than is likely under the auspices of alternative paradigsns. ‘The his- tory of anthropological theory demonstrates that those who lack a paradigmatic commit- ment inevitably “quit eatly” when confonted with difficult, puzzling phenomena, The primacy of infrastructure has frequently been criticized as a teleological principle that denies the importance of chance, random processes in evolution. This is clearly an incorrect portrayal, given the formulation of causality in corms of selection processes, B. F. Skinner (1984) has appropriately designated the genre of evolutionary process in question i gtltion by consequences. In biological evo- tion, in operant behavior conditioning, and MARVIN HARRIS 69 in cultural evolution, selection operates on variations whose origins may be indetermin- ate, Cultural materialism is thus no more teleological than Darwinian evolutionism. In this limited sense, cultural evolution is analogous to biological evolution (and to the development of individual learning reper- tories). As in biological evolution, there are innovation and selection for or against. Inno- vations occur at massive rates in socially con- ditioned human response zepertories (culture). Some are selected for (retained and propagated across generations), others are selected against (extinguished). Selection for or against is probabilistically determined by the infrastruc taral consequences (costs and benefits) of the innovative behavior. Despite this analogy, there are specific differeaces in the mechanisms of selection by consequences which characterize biological and cultural evolution, Selected biological innovations are stored in the form of informa tion in the organism's genes; selected cultural innovations are stored in the form of response repertories ia the organism's neueal pathways. It is this difference that dooms the attempts to reduce the great mass of human sociocultural phenomena to the level of biology. Another point of recurrent concern expressed by colleagues is whether culeural materialism insists that every sociocultural difference and similarity can be explained by infrastractural determinisra (Magnarella 1982). ‘This concen is misplaced. Clearly it is very likely that many sociocultural traits are the consequence of arbitrary, idiographic events. But the main task of cultural materialism is to concentrate on building a coxpus of testable theories that is broader, more coherent, and more interpenetrating than the theories of al- temative research strategies while attending to seemingly refractory cases as they are brought to fight? Another aspect of the principle of the primacy of infrastructure that is surrounded by misinformation is the feedback between infra~ structure and structure or superstructure. It70 ASSESSING CULTURAL ANTHROPOLOGY would be convenient for materialist-bashers if the principle of the primacy of infrastructure meant that cultural materialists regard the mental, emic, and symbolic-ideational aspects of socioculeurat systems as mere mechanical reflexes or epiphenomena of infrastructure. (Harris thinks ideas, symbols, values, art, and religion are unimportant aspects of human life. ‘Ugh!”) Again | quote from Murphy's paper: “As for the materialiss, they fil to recognize that cultural forms have lives of their own and are vot mere epiphenomena of underlying “infrastructures” (page 57), The attempt by Murphy and others to portray culeural materialism as a paradigm in which “the ideas by which men {sic} live have no importance for their action” (Bloch 1985b:134) is totally at variance with the prominence of the phrase “sociocultural system” in the specification of caltural materialist principles. Why does one bother to talk about the systemic role of structure and superstructure if infrastructure alone has importance for action? Do cultural materialists propose that people go about producing and reproducing at random and without an idea in their heads? Could sociocultural life a5 we know it exist if there was nothing but infrastructure? Certainly not. No more than one can imagine people living without an infrastructure, ie., living on ideas alone. Franz, Boas was absolutely right whea he noted that “no people has ever been observed that has no social structure.” But who has ever been so certifiably insane as to suggest the contrary? To sy that humans must think to live, however, is to say nothing about the roles of behavior and thought in the processes responsible for sociocultural evolu- tion. The issue is not whether thought is important for action, but whether thoughts and actions are equally important in the explanation of the evolution of sociocultural systems. Cultural materialism—indeed any genuinely materialist paradigm in the social sciences—says no. The system is asymmetrical Infrastructural variables are more decermina- tive of the evolution of the system. But this does not mean that the infrastructure can do without its superstructure. Moreover, to say that structure and super structure are causally dependent on infra structure is not to say that in the processes of continuity and change, selection pressure is exerted only from infrastructure to super- structure. Without structural and superstruc— tural instrumentalities the inftastructural subsystem would have evolved in a radically different direction ftom those which we now observe. Structure and superstructure are not mere passive, epiphenomenal products; rather they actively contribute both to the continuity and change of infrastructures, But they do so within the limitations and possibilities inher- ent in a given set of demo-techno-econo- environmental conditions. They almost always initiate and select for change in conformity with but almost never in opposition to those conditions. To illustrate, consider the changes in U.S. family life since World War If with reference to the disappearance of the male breadwinner role, the demise of the multiparous stay-at- home housewife, and the rise of feminist ideologies emphasizing the value of sexual, economic, and intellectual independence for women, As I have proposed elsewhere (Harris 19812), these structuzal and superstructural transformations are the determined outcome of a shift from goods-producing industrialism to service-and-information-producing indus- trialism, mediated by the call-up of a reserve army of housewives into low-paying service- and-information nonunion jobs. The infra- structural transformations themselves were re- lated to the use of electronic technologies and to declining productivity in the unionized smokestack industries which had created and sustained the muale-breadwinner-stay-at- home-housewife families. The rise of a femn- inist ideology which glamorized the wage labor market and the intellectual, sexual, and emotional independence of women was the determined outcome of the same infta- structural force, However, itis clear that boththe structural and superstructural changes have exerted and continue to exert an amplifying, positive-feedback effect on the infrastracrural transformations. As the consequences of the call-up of the female labor force manifest themselves in higher divorce rates, lower first marriage rates, and historically low fertility rates, service-and-information industrialism is in turn amplified into an ever-more dominant mode of production and reproduction, Simi- larly, as feminist ideologies continue to raise consciousness against the vestiges of male breadwinner sexism, men and women find themselves locked into the labor force as competitors, wages for both are driven down, unions are driven out, and the profitability of the service-and-information industries rises, encouraging more diversion of capital from. goods-producing enterprises into service-and~ information production. Although deviation-amplification can occur a8. consequence of changes initiated in any sector of the sociocultural system, the inft structural components remain causally domi~ pant. Relations among the components remain causally asymmetrica! because changes initiated at the structural or superstructural level are most likely to be selecced for if they facilitate or help to optimize crucial aspects of the infrastructure, In the example under con- sidexation, it is clear that neither feminism nor the demise of the breadwinner family selected the technological innovations which initiated the transformation of the smokestack infta~ structure. FEMINISM AND ANTISCIENCE ‘This is an appropriate point for me to make some observations about the relation between, cultural materialism and ferninism, Feminist anthropology is a distinct intellectual tradition Say seeks to establish a balance between -atidrocentic. (male-oriented) and gynocentric (feroale-oriented) perspectives, theories, and ita bases: In-view of the virtual hegemony of MARVIN HARRIS 71 androcentrism in anthropology during the first sixty or seventy years of this century, it is not surprising that feminism in practice often seems to give way to an attempt to substitute gynocentrism for androcentrism. It would require much more than the available space to begin to review the strengths and weaknesses of the specific modifications of anthropologi- cal data bases and theories that have been introduced under feminist auspices or in response to feminist critiques. But there is a general point that must be made if one is to understand the surge in postmodernist anti- scientism and relativism. The corpus of positivist-scientist theories about gender issues was impoverished, when not grossly counterfactual, prior to the rise of feminism. Feminists therefore tended to iden fy science with androcentrism. They saw sci~ entific theories as male concoctions designed to confuse and disempower women. There was a parallel reaction to Marxism, with its claim that it was a science of society while focusing on class exploitation to the neglect of gender exploitation. In this context, post modemist attacks on the distinction between the observer and the observed, and the advocacy of the idea chat truth is relative and a political construction, seemed to constitute an appropriate paradigm for carrying out the feminist project in anthropology. (Some feminists argue that these perspectives had an independent and antecedent development within feminism itself, see, for example, Mascia-Leeds, Sharpe, and Cohen 1989.) From a cultural materialist perspective, however, the ferninist response to past inade- quacies of positivist formulations of gender is both intellectually and politically counter- productive. In the long run anthropological ferninismn has little to gain by throwing its weight on the antiscience side, Scientific anthropology is largely if not yet completely ‘open to feminist researchers and to their con- tribution to the improvement of anthropo- logical theories in any domain of their choice. The most productive intellectual response to