256 THE DEMISE AND REVIVAL OF KINSHIP Yanagisako, Sylvia J., and Janet f. Collier, 1987 Towards a Unified Analysis of Gender Kinship. In Gender and Kinship: Essays towards a Unified Analysis. Jane f. Collar j Sylvia J. Yanagisako, eds. pp. 14-50. Stanford: Stanford University Press. Yanagisako, Sylvia J., and Carol Delaney, eds. 1995 Naturalizing Power: Essays in Fe Cultural Analysis. New York: Routiedge. ,.J: 14 What Is Kinship All About? David M. Schneider Tve are here to celebrate the centenary of Morgan's Systems, and the topic I have chosen for this occasion is "What Is Kinship Ail About." ] ;et us look at the one who began it all and ask what he thought kinship was all about.1 t need not remind you that Morgan was concerned to discover the history and origin ' of the Indians of North America. He believed that he could reconstruct their history and i? locate their origins by their manner of classifying kinsmen. He argued that it was not the words or the language that could be taken as reliable indices but rather the mode of classification regardless of the words or language used. a:;': Morgan's reasoning was that if the system of relationship of consanguinity should be found to be- the same among all the Indians of America and should also be shown to be the same as ■those from India, then it would follow that the Indians of America brought that system with riiem from Asia. Why? Because it is "a system v.hich is transmitted with the blood" {1871:4). ;fty "blood" Morgan meant precisely what we \hiean: genetics and biology. He says elsewhere : in the book: "In the systems of relationship of the great families of mankind, some of the oldest memorials of human thought and experience are deposited and preserved. They .: have been handed down as transmitted systems trough the channels of the blood, from the earliest ages of man's existence upon the ; earth; but revealing certain definite and pro- gressive changes with the growth of man's experience in the ages of barbarism" (1871:vi). What did the mode of classification show? How did it come about? What did it reflect? Morgan's answer was: The actual biological facts as these were known or knowable, given the state of knowledge of the group on the one hand and the state of marriage and sexual relationship on the other, at the time the classification first was established. In Morgan's own words: The family relationships are as ancient as the family. They exist in virtue of the law of derivation, which is expressed by the perpetuation of the species through the marriage relation. A system of consanguinity which is founded upon a community of blood, is but the formal expression and recognition of these relationships. Around every person there is a circle or group of kindred of which such person is the center, and the Ego, from whom the degree of the relationship is reckoned and to whom the relationship itself returns. Above him are his father and mother and their ascendants, below him are his children and their descendants; while on either side are his brothers and sisters and their descendants and the brothers and sisters of his father and of his mother and their descendants as well as a much greater number of collateral relatives descended from common ancestors still more remote. To him they are nearer in degree than other individuals of the nation at large. A formal arrangement of 258 DAVID M. SCHNEIDER WHAT IS KINSHIP ALL ABOUT? 259 the more immediate blood kindred into lines of descent, with the adoption of some method to distinguish one relative from another and to express the value of the relationship, would be one of the earliest acts of human intelligence [1871:10]. And so Morgan called the descriptive system a "natural system" precisely "because it is founded upon a correct appreciation of the distinction between the lineal and several collateral lines and of the perpetual divergence of the latter from the former. Each relationship is thus specialized and separated from every other in such a manner as to decrease its nearness and diminish its value according to the degree of distance of each person from the central Ego" (1871;142-143). Conversely, the classificatory system, as it is used among American Indians and others, Morgan said, is contrary to the nature of descents, confounding relationships which are distinct, separating those which are similar, and diverting the streams of the blood from the collateral channels into the lineal. Where, for the descriptive system, knowledge of the lines of parentage is necessary to determine the classification, just the opposite is true for the classificatory system; a knowledge of parentage is quite unnecessary. It is impossible to explain its origin on the assumption of the existence of the family founded upon marriage between single pairs; but it may be explained with some degree of probability on the assumption of the antecedent existence of a series of customs and institutions, one reformatory of the other, commencing with promiscuous intercourse and ending with the establishment of the family, as now constituted, resting on marriage between single pairs {paraphrased from 1871:143). It will prove useful for us to keep two parts of Morgan's paradigm distinct from each other. One is the mode of classification itself. The other is the manner in which the mode of classification can be established, that is, by means of the analysis of the kinship terminology. Morgan's paradigm states that the mode of classification of kinsmen derives from the knowledge of how people are actually genetically or biologically related to each other. This knowledge in turn depends on their form of marriage. Hence for Morgan, as for others since, marriage is the central institw kinship. Implicit in this part of Morg: digm is the premise that marriage consists of " sexual relationship between male anil fn It is the processes of biological reproduce that make the married pair the parents 0f their biological offspring and the offspring 0f such a mated pair are siblings. The links which are recognized or marked in the mode of sification of kinsmen are the biological or etic links among these people as these mi. known, which in turn depend on the mode of marriage. Thus, by taking one male and ■.' female in the abstract and tracing their sibl. their parents, their offspring, and the pat-siblings, offspring, and spouse of each of these it is possible to create a genealogical grid, a\ it is called today; the particular classificatic ■ ■■ kin which a particular people use cat . mapped on this grid and compared with <:' ■ classifications which other people use by comparing the differently partitioned grids, classification, in turn, can be derived frmn which positions on the genealogy are groi: .. together and which positions are distingui; ... [I]t seemed obvious to Morgan thai mode of classification could be read directly from the kinship terminology; that is, f positions on the genealogical grid which . grouped together under one kinship term could be distinguished from those positions on ■ genealogical grid grouped under a diffi rent kinship term and so on. Hence kinship terminology was the key to the mode of classification and in fact, practically the only key, since rht kinship terms meant (either only or prim;: specific relationships of blood or marriage. '1 Iw taxonomy, then, was derived from no oi ■■ source than the kinship terminology. What was kinship all about for Morgan, \ then? Kinship was about the way in which a people grouped and classified themselves as. r; compared with the real, true, biological facts of consanguinity and affinity. The facts of con- ■'■ sanguinity mean those persons who are related "■■/'■ by biological descent from the same ancestor. The facts of affinity are the facts of marriage, and marriage means the sexual, reproductive ■ ■ relationship between male and female. McLennan took issue with Morgan on one specific score, and his argument is easy to mis- ,r.-;-understand if one does not observe it closely.. . I If... mm . mistake, or rather I should say error, have so lightly assumed the system to be pern of blood-ties" (1886:269). for the following reasons I think that assump-■ (jjjrt was an error:-(l) it is apparent, on the ijjghtest inspection of Mr. Morgan's tables, rfiat 'son' and 'daughter', in the classificatory i-'system, do not mean son or daughter 'begotten Sby' or 'born to'; that 'brother' and 'sister' are f-.--terras which do not imply connection by des-cer). from the same mother or father; and that 'jnitthw' does not mean the bearing mother. j.-rDm the analogies of the case, we must be-■ : lieve that 'father' does mean the begetting .^father. ...These facts surely ought to have ":■ strongly suggested that the classificatory .. .system cannot be a system of blood-ties at -fi'all... [1886:270], {2) That the classificatory i; .'system is a system of mutual salutations merely, appears from many of its peculiar features. For one thing, the names for relationship j. are framed for use in address. They want generality [McLennan 1886:270, 273]. This, then, is what McLennan said KINSHIP ■TERMS were all about; they were courtesies ./and modes of address, of mutual salutation. But did McLennan differ with Morgan on what 'KINSHIP' was all about? Not a bit! all, or almost all, the peoples using a form of :the classificatory system have, besides, some well-defined system of blood-ties - the system .■ which traces blood-ties through women only, : or some other. It is inconceivable that any , people should have at the same time two en-.. tirely different systems of blood relationship. And it may be confidently affirmed that in > every case it is the system which is unquestion-. i ably a system of blood-ties, and not the classi-: iicatory system, that alone is of practical force - which regulated succession, for instance, to honours or estates. ...What duties or rights are affected by the 'relationships' comprised of the classificatory system? Absolutely none. : They are barren of consequences, except . indeed as comprising a code of courtesies and ... ceremonial addresses in social intercourse [McLennan 1886:270-273]. For McLennan as well as for Morgan, 'kinship' was about marriage, about the facts of procreation and conception, about blood-ties and genetic relationships as they could be known or were knowable, about the ties that arise out of the biological facts of human reproduction; for McLennan, rights and duties, succession and estates followed blood-ties, not kinship terms. For both Morgan and McLennan, marriage meant a sexual relationship between male and female; consanguinity meant descent from the same ancestor. These are the only two components that are necessary for the construction of a genealogy, that is, for the construction of the analytic apparatus needed to describe any particular mode of classification or kinship system and to compare it with any other system.2 Ever since Morgan's Systems, anthropologists have followed this basic paradigm in its essential outline and have continued to argue about the meaning of kinship terms as well. For many since then, like Durkheim and Rivers, the notions of paternity and maternity and blood connection had to be taken in their social and not in their biological meanings; indeed, their social and their actual biological senses did not always accord with each other too well. Sometimes these biological relationships are either presumptive, fictive, errors of fact based on ignorance, or putative rather than empirically demonstrable. But this hardly alters the fact that it is the system of what are socially defined as the biological facts of reproduction that 'kinship' is all about. That there are rights and duties, statuses and roles, and interpersonal relationships of different complexions associated with the genealogically defined 'kinship' relationships has always been, agreed; but the two have been kept quite distinct and held to be inherently distinguishable so that the defining feature, the definition of a 'kinship' relationship as against any other kind of relationship, has always been the biological aspect, whether treated as pure biology or as the social definition of what biology is. Indeed, the prevailing view since Morgan has been that the fictive or presumptive or undemonstrable biological relationship, the social aspect itself, is modeled after, or is a metaphorical extension of, or is a social accretion to, the defining and fundamental biological relationship. Thus for instance adoption is not ruled outside the 'kinship' system but is understandable as a kind of 'kinship' relationship precisely in terms of the fact that it is modeled after the biological relationship. Without the biological relationship, in this 260 DAVID M. SCHNEIDER view, adoption makes absolutely no sense. Hence even if it is in its social aspects, and even if it is as a social relationship that anthropologists are concerned with it, the real, actual, and true facts of biology as they concern human reproduction remain the base and the defining feature of 'kinship.'3 A variant of this view, which is not fundamentally different from it is the position that the genealogical grid itself, can be treated as the defining feature, regardless of how the specific genealogical relationships themselves may be defined and even when they are not defined in biological terms at all. Thus whatever the theory of procreation may be for a given people, it is the fact that a system of genealogical-like relations can be mapped out and partitioned into categories, each systematically related to the other which is the crucial and defining feature of 'kinship.' Yet however different this position may seem to be at first glance it boils down to the fact that a parent-child relationship - however that may be defined procreatively - obtains which implies a sibling relationship which implies a sexual relationship between parents, and so on, which creates the genealogy. By definition, of course, no genealogy is formed or can be formed from the exchange of morning greetings or salutations, nor can a genealogical grid be constructed from material other than some set of premises about the nature of human reproduction. The two sides of 'kinship,' the biological model (whether real or presumptive, putative or Active) and the social relationship (the rights, duties, privileges, roles, and statuses) stand in a hierarchical relationship to each other, for the biological defines the system to which the social is attached, and is thus logically prior to the latter. If two relationships are precisely the same except for one single feature, that one is the 'kinship' relationship where some biological relationship prevails or is presumed to prevail, and the other one, lacking this feature, is not. It is possible to hold that the genealogical grid can be distinct from all other aspects of 'kinship' and that the boundaries of the system are those defined by the grid. These boundaries, for some but not for all, include the putative or metaphoric extensions of the genealogical grid. [... ] The position I have argued both in prinr in person is that Morgan's paradigm k wr^q and that no matter how elegantlj n v, ^ been revised, amended, altered, embelliskT; or tightened up, it does not do what it purpo to do. I take it that Lounsbury and Goode" nough in the United States follow that para" digm, as well as Levi-Strauss in France. Leach and Needham in England, and many others { do not mean this as an exhaustive li>,t (,t fQ| lowers of Morgan's kinship work, but mily t0 suggest that it holds a position of preeminence in the anthropological world today. Neithet do I mean to suggest that the work of ( enough and Lounsbury or any of these men is. in any sense identical except that they all follow Morgan's use of the genealogical grid as the basic analytic too! and they all remain wedded to Morgan's definition of what 'kinship' is all about. My criticism of Morgan's paradigm is plainly contained in the alternative strateg\ I have followed. I have tried here to show its utility and productivity given my aims, objectives, and analytic procedures. 1 There is general agreement among all of us, followers of Morgan as well as others, that a classification of kinsmen does not exhaust the 'kinship' system by any means. Where we differ ' is in how we handle this fact. The position''' which follows from Morgan's paradigm; : which Lounsbury, Goodenough, Levi-Stranss,:-" Leach, Needham, and the many others whom! . should mention take, is that the kin classification can be treated as a distinct, separate, and autonomous part of the 'kinship' system, however it may be related to a larger system. Just as some anthropologists believe that the phon- ..■ emic system can be analyzed apart from-., grammar and syntax in language, so these an- ■ thropologists also feel that the kin classification can be analyzed apart from the rest of llie 'kinship' system. My own position is that an accurate account of the kin classification m a cultural sense (see below) cannot be given, without taking the-whole 'kinship' system into account. The second major part of the strategy I have followed is to ask what, in each and every i WHAT IS KINSHIP ALL ABOUT? 261 Ilillillil nce the definition of the domain of 'kin-, may be for each and every culture I study, do noi assume that this domain is defined a ' °ri bytne bio-genetic premises of the genea-f ideally defined grid. In other words, where ^followers of Morgan take it as a matter of Jefiniaori that the invariant points of reference ovided by the facts of sexual intercourse, conception pregnancy, and parturition constitute die domain of 'kinship,' I treat this as an open, empirical question. Of what primitive e\ejfients, I ask in each and every case, is the cultural system composed? It is this question which on the one hand enables me to ask what 'kinship' is all about, while on the other hand it seems to deprive me of an externally based, systematically usable comparative frame. I shall return to this point below. The second major aspect of the strategy I have used follows directly from, and is required by, the third, which is the use of a different, narrower, and I think sharper and more powerful concept of culture than has been traditional in anthropology. Briefly, I start with concrete, observable patterns of behavior and abstract from it a level of material which has usually been called 'norms.' The normative system consists in the rules and regulations which an actor should follow if his behavior is to be accepted by his community or his society as proper. These are the "how-to-do-it" rules, as Goodenough has recently put it (1970). They should on no account be confused with the patterns of behavior which people actually perform. It is the rule "thou shalt not steal" that is the norm, not the fact that many people do not steal; it is the rule that a middle-class father should earn the money to support his family, not the fact that many actually do. The next step is to abstract from the normative system what, following Parsons, I have called the 'cultural system' {Parsons 1966, 1971). This consists in the system of symbols and meanings embedded in the normative system but which is a quite distinct aspect of it and can easily be abstracted from it. By symbols and meanings I mean the basic premises which a culture posits for life: what its units consist in; how those units are defined and differentiated; how they form an integrated order or classification; how the world is structured; in what parts it consists and on what premises it is conceived to exist, the categories and classifications of the various domains of the world of man and how they relate one with another, and the world that man sees himself living in. "Where the normative system, the how-to-do-it rules and regulations, is Ego-centered and particularly appropriate to decision-making or interaction models of analysis, culture is system-centered and appears to be more static and 'given,' and far less processual (but only in contrast with the normative system of course; culture has its own processes and its own rules of change and movement). Culture takes man's position vis-a-vis the world rather than a man's position on how to get along in the world as it is given; it asks "Of what does this world consist?" where the normative level asks, "Given the world to be made up in the way it is, how does a man proceed to act in it?" Culture concerns the. stage, the stage setting, and the cast of characters; the normative system consists in the stage directions for the actors and how the actors should play their parts on the stage that is so set. This is not to say that the cultural and normative level are unconnected. The cultural level contains implications for the general directions in which normative patterns of action ought to take place, but it does not spell them out in the detail which the normative patterns themselves provide. The cultural premise that "there are two kinds of relatives, relatives by blood and relatives by marriage," does not tell how a man should treat his relatives by marriage. Yet once it is known that there are these two categories, how each is defined, and the values attached to each, general directions of action are laid out already even if they are not sufficient to provide a precise template for how-to-do-it. By the same token, the cultural premises allow a wide range of possibilities and alternatives in the normative rules,4 This conception of culture is far more narrow and, I think, far more precise than those generally in use in anthropology today. Furthermore, it is explicitly tied into a wider social theory rather than linked in a loose, ad hoc way to a variety of eclectically given and not always internally consistent theories. This conception of culture and the social theory of which it is a part yields a considerably smaller, more concentrated, and homogeneous body of 262 DAVID M. SCHNEIDER WHAT IS KINSHIP ALL ABOUT? 263 materials abstracted as culture than many other definitions. This leads to the final point in this introductory section. What 'kinship' is al! about is considered here only in its cultural aspects; it is 'kinship' at a cultural level as here defined. I am explicitly not speaking of 'kinship' at a psychological level. Nor am I speaking of it as a system for the organization of social groups, that is, not at the social system, social organizational or social structural level, for these are, by my definitions, not the same as the cultural system. The cultural level is focused on the fundamental system of symbols and meanings which inform and give shape to the normative level of action. This theory, like every other theory, is easily transformable into a series of questions which are put to the data. It assumes that every concrete act, or system of action, has a cultural component, a social system component, a psychological component, and so on. Thus the question I am asking, which follows directly from this theory, is: "What are the underlying symbols and their meanings in this particular segment of concrete action and how do they form a single, coherent, interrelated system of symbols and meanings? If that question cannot be answered satisfactorily, there must be something wrong with my theory. If I can answer the question, it may at least show that the theory can be applied, even if it is not enlightening. I have followed this theory, however, not because it is merely applicable but because I think that it is enlightening and that we learn much from it. Many other fruitful and useful questions can be asked about the same segment of concrete action using the same social theory. For instance, one can ask about the motivations entailed in that action. Or one may ask the history of that segment of action. A relatively common question in anthropology is sometimes called a comparative, functional question.5 The crux of the issue, then, is what is being compared. If we ask how any particular cultural system is constructed, for instance, we ask what units it contains, how they are defined in that culture, how they are differentiated and articulated as symbols and meanings; but if we pose a question taken from outside any particular socio-cultural system, this is different from the cultural question. For instance, it js '-' functional prerequisite to the maintenance 0f any society to regulate sexual behavior, unregulated sexual behavior is a source of jjs ruption. We can then ask of each socio-ctilu1Ea| system or society, "How do they do it?" Thi -boundaries of sex and of the different regulatory mechanisms are defined in terms of their?:' relevance to the question and are related only loosely to the boundaries which the so : itself embodies in its cultural constructs. We"'-may ask, for whatever reason, how the pt0. cesses of human reproduction are ordered '.in'-, different societies, and a study of certain-" aspects of their 'kinship' system will of course-be included; but the particular cultural con- -structs which obtain within that society are' cut off or are included at points determined by the relevance of that material to the ques-. '■' tion being posed and asked in a comparative-framework from outside the bounds of the par-'. ticular culture. A cultural question is by definition a question of from what units this particular socio-cultural ■ system is constructed, of how those units ate defined and articulated, and of how those units form a meaningful whole. It is not true that such a question necessarily yields material which is unique, distinctive, and cannot, be compared from one society to another. Quite.' the contrary. The systems of symbols and mean- : ings of different cultures can be compared as ■ easily as systems of human reproduction can be-compared from one society to another. I do not mean to play semantic games here or. to beg fundamental questions; but if culture consists in the system of symbols and meanings of a particular society, and if a social system consists in the manner in which social units are organized for various social purposes, then comparative operations of the' former are cross-cultural comparisons while, by definition, comparative operations of the latter arc not cross-cultural comparisons but rather cross-social comparisons, that is, comparisons of social organization, social systems, or social structures. The key definitional difference lies between the concepts of society and culture, between modes "of organization of action systems and systems of symbols and meanings. I am concerned with questions of cross-cultural comparison and questions having to j0 with the analysis of particular cultures, not social systems. "wliat happens if this analytic strategy is used on a particular 'kinship' system? Does it tell us anything usefully new or different about that system or about the nature of 'kinship'? I have tried to do this for the American 'kinship' system. Since much of this material already has been published (1968, 1969: 116-125, 1970:88-90), I will merely touch on the points which bear directly on the task at hand. What anthropologists have heretofore regarded as THE domain of 'kinship' in American culture turns out to be only one part of a larger domain, made up of two different parts. The domain we have traditionally called 'kingship' is Ego-centered, consisting of a network of related persons, such as mother, father, brother, etc. It is not hard to see that ^ this domain is constructed out of many different kinds of components from many different -systems. Thus each unit in the system, such as . 'mother' or 'father,' is defined first by whatmight be called a pure 'kinship' component, second by an age or generation component, third by a : sex-role component, fourth by a class component, and by other components of other kinds as '■:'■■ well. Hence I have called this the 'conglomer-;: ate' system or the 'conglomerate' level of the system. To understand the second part we must go a . step further. We can, by using a level of contrast ^vhich is not generally employed in 'kinship' analysis, abstract the 'kinship' component ■ alone and in its pure form from the conglomerate system. We do so by asking what the dis-: tinctive features are which define the person as a kinsman as against a non-kinsman.6 When we do this, it becomes apparent at the level of the 'pure' system that the distinctive features or the defining features (1968:22, 41 ff.) or the primitive and irreducible elements are, first, shared bio-genetic substance and, second, a code for conduct which I have characterized as diffuse, enduring solidarity. These two elements combine to yield three major categories of kin; when both elements occur together the category of blood relative is formed; when the code for conduct element occurs alone and without the shared bio-genetic substance element the category of relatives-in-law or relatives by marriage is formed; and, finally, when the shared bio-genetic substance is present alone the category of relatives in nature is formed. Hence at the pure kinship level the so-cailed 'kinship terms' do not play a classificatory role.7 If we consider the pure 'kinship' system alone, we can see that the distinctive features by which it is defined are parts of two much wider and more general categories of American culture. That is easy to see when we remember that blood relatives are considered to be related in nature and that they are parts of the natural order of things as defined in American culture. Their second distinctive feature, the code for conduct, is simply part of that much wider category called the order of law, defined in opposition to the order of nature. This is the order imposed by man on nature, the order defined in American culture as consisting in rules, regulations, customs, traditions, and so forth which man, with the aid of human reason, creates. The limited domain of law in the juridical sense is only one part of this wider domain; and when we understand how much a part of the same domain they are, we have explained in some significant degree why relatives 'by marriage' are also called relatives 'in law.' At the pure level, then, part of what anthropologists have traditionally been calling 'kinship' seems to be defined in American culture as an indistinguishable part of these two much wider and more general cultural categories, the order of nature and the order of law. If we now consider the domains of religion and nationality8 and analyze them as we have analyzed 'kinship,' a rather interesting fact emerges. We again distinguish the pure system from the conglomerate system. The conglomerate system of nationality consists in the entire federal and state systems; the legislative, judicial, and executive branches of government; the two Houses of Congress; the different states and their organization, and so forth. But to abstract the pure system we simply ask, What makes a person a citizen? What are the distinctive features which define a person's nationality? He is either born an American or he is - and the word is of course quite 264 DAVID M. SCHNEIDER significant - naturalized. Once again we find that the distinctive features are shared substance (being born American} and a code for conduct which enjoins diffuse, enduring solidarity: being a loyal American, loving one's country, and, in President Kennedy's felicitous phrase, "Ask not what your country can do for you. Ask rather what you can do for your country." The same is true, as I have tried to show elsewhere, for religion, where the conglomerate level includes the organization of the church and so forth; but what makes a member of the church is, once again, shared substance and a special code for conduct which can be characterized as diffuse, enduring solidarity (1969). [...] If the conglomerate level consists in units made up of elements from different pure systems, then the question arises of how the different components relate with each other in the conglomerate unit. Are they simply added together? Do they form some special configuration? The answer to these questions seems to be that at the level of the person, each pure component receives further specification of its content, defined now with reference to how a person should act, and this further specification derives from the total context or the interaction of ali of the defining components. Let me try to explain this by going back to the so-called 'kinship' component once more. The 'kinship' component says that persons are related either by shared bio-genetic substance, a shared code for conduct, or by both. But at the level of the person something called 'distance' comes into play, so that the question is no longer shared or not shared but of how much is shared. If the shared elements are now conceived in terms of magnitudes, then class factors, personal factors, and a variety of other considerations permit it to be cut off at various points and at various times and under various circumstances and for various purposes at the option of the actor himself. Hence the common observation in both America and England that some people will actually trace blood connections to people, whom they then say they do not count among their relatives or kinsmen. They simply say, "Yes, they are my blood relatives, I suppose, but I don't consider them relatives; they are too far away" or words to that effect. pure In sum, the difference between the system and the conglomerate system ljgr their orientations. The conglomerate system ^ oriented toward action, toward telling p60J* how to behave, toward telling people how-to6 do-it under ideal circumstances. It is thus mud, closer to the normative system. Tbf r», system, however, is oriented toward the state of-being, toward How Things Are. It is jn the transition from How Things Are and Hoy/ Things Ought To Be to the domain of If That Is So, How Then Should One Act that the pure systems come together to form the conglo'mer-ate systems for action. At this point the question of just what is " 'kinship' all about or how the domain of 'kinship' is to be defined must be raised. If, on tho one hand, the broad categories of the order of' nature and the order of law contain as special instances the two major components which ate distinctive features out of which the categories of kin are formed, and if, on the other hand, at the level of the pure system, the 'kinship' system, the nationality system, and the religious system cannot be distinguished from one another in terms of their defining features, what justification is there for calling this system either a 'kinship' or a 'religious' or a 'nationality' system? They are, culturally, speaking or with respect to their distinctive features, all exactly the same thing. And if it is true that at the level of the conglomerate system it is not possible to say that; the 'kinship' component is dominant and modifies the sex-role component, or the other way around, but only that each retains its integrity in the configuration, while a new, emergent level is formed, then it is equally problematic as to what, for comparative ana--lytic purposes, any particular conglomerate system should be called. That is, if it is a 'kinship' plus sex-role plus age-role plus class, system, why call it a 'kinship' system? Or, for that matter, why call it a sex-role system? Is there one good reason why a particular bundle of components should be characterized by only one of its components rather than by another? There is ONE good reason and that is when, in the particular culture we are studying, it is done that way. I can think of no other good ■ WHAT IS KINSHIP ALL ABOUT? 265 ■jliis turns out to be the case in American |ture. As in modern Western European cul-in general, there are clear-cut, formal div-;-joiis which are called in that culture itself 'institutions.' These institutions refer precisely t0 the conglomerate level - the 'family' is one, jjje 'church' is another (it may also be called 'ielig'on')> the 'state' is a third, and so on. Hence if our term 'kinship' is synonymous wjth that institution as it is defined in American culture, sometimes called 'family,' then 'kin-sl,ip' is indeed a valid cultural unit which is actually found in American culture, and it is found so that its defining features are at the cultural level to be identical with those of religion and nationality while it is found to be very clearly differentiated from those other units at the conglomerate level and in its normative aspects. Nor should it be forgotten that, however 'kinship,' nationality, and religion are differentiated at the conglomerate, organizational level, the very same distinctive features which define ali three as cultural domains are themselves an integral part of the orders of nature and law. That is, we have simply not explored the entire universe of American culture and so Vre cannot as yet say what other units should ■fall into the same cultural category with 'kinship,' nationality, and religion or, to put it in the other way, whatever other categories ex-: haust the domains of the order of nature and Taw. Thus there are grounds for accepting Parsons' suggestion that education ought to be considered along with kinship, religion, and the moral community (nationality) as part of a single cultural entity. My purpose here, however, is not to introduce a new element but to remind the reader that we have approached .American culture rather as the blind men approached the elephant. Even if we have discovered that a leg is linked to the body, we have not gone much further and cannot claim to have examined American cultural categories exhaustively. This is a very important point to which I shall return. In introducing the terms pure and conglomerate I confine their meaning to the cultural level and speak only of cultural components. It is clear that the pure level is confined to the cultural level alone. The conglomerate level should be understood as the cultural elements embedded in the normative system and the way in which they are embedded in that system as well as the way in which they are articulated to the social system components at the normative level. The normative level thus includes more than those cultural elements in it. It follows therefore that the conglomerate level and the normative system are at the same level of abstraction, but that the notion of conglomerate is simply the identification of the cultural elements in their matrix of the normative system. To move to the pure cultural level, then, is to abstract distinct cultural domains apart from and regardless of the normative matrix in which they are found. Thus one normative matrix containing certain cultural elements may be an institutionally distinct family system in modern Western European society, but the pure cultural domain is quite different as I have tried to show, and the same cultural elements can be found in a variety of other differentiated normarive systems as well (such as religion, the moral community, etc.). To distinguish the conglomerate level, then, is simply to locate the cultural elements in their place in the normative system and to be able to analyze them in relation to each other and to the normative system which contains them. Let me conclude this section by repeating that as anthropologists we can study different cultures or we can study different social systems. If we study different cultures we do not do the same thing as when we study different social systems. When we study different cultures we study different conceptual schemes for whar life is and how it should be lived, we study different symbolic and meaningful systems. We do not study the different ways in which different theoretically defined functions are actually or ideally carried out. There is thus a major difference between cultural anthropology and what has been called, following British usage (and quite correctly, too) social anthropology or comparative sociology. I take my task to be the study of culture and identify myself as a cultural anthropologist (although I will be the first to admit that this has not always been so). Given this definition of the task, we can proceed. Even if 'kinship' is culturally segregated as a domain at one level of American culture - the conglomerate or normative level - it is not culturally segregated as a distinct 266 DAVID M. SCHNEIDER WHAT iS KINSHIP ALL ABOUT? 267 domain in most of the cultures we encounter outside the Western European culture area. Quite the contrary. The vast majority of other cultures we know do not have culturally differentiated domains of the sort which occur in Western European or American culture. But note clearly what I am trying to say. I do not mean that we must cease to observe domestic arrangements in different societies; I do not mean that we cannot ask how a people order their relationships between men and women or between a woman and the child she bears; I do not mean that we cannot go out and ask for the theories of procreation which a particular people hold. We can do all of these things and more and have learned much from such questions. I mean only that such questions of organization, or social organization, or social structure should not be confused with or identified with questions about the nature, the structure, or the content of either particular cultures or of culture in general. Because domestic arrangements can be an analytic category which may not correspond to anything as it is defined as a cultural category in a particular culture, the relationship between a woman and the child she bears may be an analytic category which we erect for various reasons, but it may or it may not correspond to any particular cultural category in a particular culture; theories of procreation may be an analytic or functional category which we invent but which may or may not have one or another cultural counterpart in a particular culture, or be incorporated indistinguishably into one or another cultural scheme in a particular culture. It may indeed be true that some culture does have, as a cultural category, 'domestic units,' but that needs to be shown empirically, not assumed so simply on one theoretical ground or another. There is one final point. I have said that the American 'kinship' system has two distinctive features, shared bio-genetic substance and diffuse, enduring solidarity. I have said elsewhere that these derive from the master symbol of coitus and that each is a facet of this act. The last few pages of my book, American Kinship, make the point that the biological elements have symbolic significance. They constitute an integrated set of symbols in the sense that they are a model for how life, in certain of its aspects, is constituted and should be h>. . The symbols are 'biological' in the senst : ..' the culturally given definition of the syinh0i system is that it is derived from the fact biology as a process of nature itself. t4i?i-fundamenta! to our understanding that we ajw preciate that these biological elements are'i .:. bols and that their symbolic referents are biology as a natural process at all.9 Now one may well ask if in a somev. ■ roundabout way I am not saying here \. ..■ Morgan and his followers have always sa>d, for they too stressed genealogy as a biologically defined network, and descent can easily be seen as shared bio-genetic substance, the whole -being treated in its social rather than in its biological aspects. I think that I am saying -something quite different. First, although what appear to be biological elements seei-be present in both Morgan's analysis and mine, we treat those elements in very different wa\ s [ insist that these 'biological' elements have primarily symbolic significance and that their meaning is not biology at all. Morgan and his followers have insisted that it is the biological elements of human reproduction as they are scientifically demonstrable in nature w . ..■ are directly reflected in 'kinship' and that it: is these facts which people have slowly, ovei time, learned to recognize more or less accurately and then give further social value. Tor Morgan the matter stopped there; but for many of his followers, like Rivers, Malinow ski, and some of our contemporaries, the biological elements need not rest on the scientifically determined or actual facts of nature but merely on whatever the natives believe to be the facts of human reproduction. Thus whatever their theory of procreation, it is the fact that I are believed to be true facts of nature and it.is therefore in terms of these biological, or hypothetical facts of nature, that 'kinship' is defined. Whether it is the true facts of human reproduction or only those which the natives happen to believe to be true, human reproduction in its biological aspects plays the fundamental role for Morgan as well as for the functionalists who follow him. For both, the socio-cultunif position of 'kinship' is similar. For Morgan, it was an achievement of great evolutionary significance when die classificatory system gave :^ay to the descriptive system, for it showed not 0I1ly that man had achieved the most advanced ,/i focni of the family but had also achieved an advanced level of knowledge, for the descriptive system was "founded upon correct appre-, j ciatton of the distinction between lineal and several collateral lines of perpetual divergence 0f the latter from the former" (1871:142). "A j formal arrangement of the more immediate jj]ood kindred into lines of descent, with the < adoption of some method to distinguish one : -■ relative from another and to express die value of the relationship, would be one of the earliest aCts of human intelligence" (1871:10). For the 'functionalists' of Malinowski's school the situation was not much different except that the evolutionary material was excised with gusto. 'Kinship' was the social recognition of biological facts, and the presence and function of a socio-cultural system of 'kinship' was explicable and understandable precisely on the ground that these facts constitute elements in the external environment with which man must cope directly as well as indirectly and to which he must adapt. His way of coping with them and adapting to them is, by definition, the 'kinship' system. The family, a part of the 'kinship' system, was seen by ; Malinowski as, among other things, one way of maintaining order in the sexual sphere, for it provided rules and regulations governing sexual relations and these, when obeyed, were orderly and permitted man to proceed with his life in an orderly fashion and without disruptions and the chaos that would be attendant on unregulated sexuality. As I have already indicated, I too am a functionalist and I too have a functional explanation to offer, though it is somewhat different from Malinowski's or his co-workers'. No one can disagree that man must cope with the facts of life and the facts of nature, whether he knows what these facts are scientifically or has only erroneous beliefs. It can be demonstrated easily that the question of how man copes with the facts of human reproduc-tion is answered only in part, and in very small part at that, by the 'kinship' system. But that is not the main point here at all. The main point here is that that is a social system question, a sociological question. It is a question of how roles are defined and articulated into a set of patterns for action which adapt man to the facts of his environment. A different functional question centers on the cultural rather than the social system level. That question has to do with the system of symbols and meanings which the so-called 'kinship' system embodies, with what the boundaries of that sub-system of symbols and meanings are and whether they stretch beyond the 'kinship' system or only fall within a portion of it. The functional question at the cultural level, then, is what that system of symbols and meanings consists in and, once that is answered, what part it plays in the total socio-cultural system. [... ] III In section I, I drew a distinction between culture and social or normative system and said that this distinction had an important place in a wider social theory, essentially Parsonian in conception. The fundamental point of section II was that at the cultural or symbolic level, 'kinship,' religion, nationality (pending a full clarification and revision of this term), and possibly education as well are identical, although they are quite different in their social system or social organizational aspects. In sections I and II, I emphasized that the question asked of the data is different, depending on whether it is a cultural question or a social system or social organizational question. The next problem, and the problem of this section, is the old one of how comparison can be conducted on a cultural level if it is assumed that each and every culture may be uniquely constituted. How can one compare wholly different things? In part, the answer to this question has been given in the discussion of the differences between culture and social system. The units of any particular culture are defined distinctively within that culture. By definition, they cannot be imposed from outside. It follows, therefore, from the definitions and the theory used here, that there is and can be only one cultural question, the question of what its particular system of symbols and meanings consist in. We must start, of course, as adults who have lived in our own society and been socialized in 268 DAVID M. SCHNEIDER WHAT IS KINSHIP ALL ABOUT? 269 our own culture before we even imagine any others. Thus we start by asking that question and answering it for our own culture, which always serves as a base-line for cross-cultural comparison. Without some comprehension, however botched, distorted, biased, and infused with value judgments and wishful thinking, both good and bad, our own culture always remains the base-line for all other questions and comparisons. In part, this is because the experience of our own culture is the only experience which is deep and subtle enough to comprehend in cultural terms, for the cultural aspects of action are particularly subtle, sometimes particularly difficult to comprehend partly because they are symbols not treated usually as symbols but as true facts. So it is difficult at times to convince an American that blood as a fluid has nothing in it which causes ties to be deep and strong. So, too, many aspects of culture are unconscious and are not part of an explicit scheme of things. A more fundamental reason for the fact that our own culture is always implicitly or explicitly, immediately or remotely, the base-line for comparison and comprehension is that that is what anthropology is all about. It is an attempt to understand ourselves as human beings by using anthropology as the mirror for man. By seeing ourselves against the contrast of others and by seeing others in contrast with ourselves, we learn about ourselves and about mankind. It is unnecessary to raise the old problem of how it is possible for two things to be compared as wholes when each is wholly unique. We are spared this burden by the fact that the basis for comparison is given by our definition of culture as a system of symbols and meanings. Symbols and meanings can be compared just as easily as modes of family organization, the roles of seniors to juniors, or the methods of agriculture. The comparative base is given therefore by the theoretical stipulation of what it is we are trying to abstract from each system and from the fact that we can indeed systematically abstract the system of symbols and meanings for each society. Hence the key to the comparative problem is in locating the symbolic elements from a careful analysis of the units which the culture itself defines. We do not say, "Let's look at the lineages," we ask instead what units this culture postulates, and the answer may have nothing whatever r<> jj with lineages. We must then follow these bolic elements throughout the particular eta] ture, wherever they may lead and in whatever forms they may be found. In short, framing a question is the first step. It must thenar-answered for our own culture as an hypothesis'' One then takes those cultural constructs and asks if any other culture has anything like it0r not, how they differ, where and in what way and where they appear to be the same. [.,■.]'■■■■■■■■ Let me conclude and summarize by returning to Morgan and company, I think it is quite clear'■ that this is not in fact what Morgan and his'-1 followers have actually done. Their cultural ~ categories do not come from a previously analyzed culture at all, but are composed of ad hoc-' elements which derive from social system quej. -■■ tions, functional questions, and from (in Morgan's case especially) evolutionary comicl--,. erations, all of which are quite foreign to any. particular culture. Morgan did not use the cul-' tural system of American 'kinship' as the model for his comparative analysis because as I have: * shown in American Kinship, the genealogical grid which Morgan used is not part of thai;: system. Morgan is quite clear that what he.:, took to be the comparative model as the many "■-quotes cited at the outset of this paper show, is:-; the genealogically defined or biologically, defined network. By using the genealogically dc- ■! fined grid Morgan and his followers have-: protected themselves from finding out-., what the true units of American 'kinship' in a cultural sense are and what their distinctive-, features actually are. In other words, they have not dealt with American 'kinship' as a cultural system but have simply assumed that their model caught or contained some part of it. I have affirmed repeatedly that the genealogically defined grid is not appropriately ap-; plied to American 'kinship' for three reasons. First, it does not in fact correspond to the cultural units of which the American 'kinship' system is actually made up, nor to the d i si i active features in terms of which these cultural:' units are defined, unless, of course, the results of the research presented in American Kinship are largely in error. Second, the genealogically defined grid is tied to the false assumption that it is possible to discover the classification of;. ([iosiW'11 without taking into account the rest of "the 'kinship' system, particularly the system of ' tQ]es and patterns for behavior as well as the '"'ffider cultural context in which the 'kinship' ■ system is situated. Third, as McLennan was first to point out and as only a few since have maintained, is the false assumption that the so-called 'kinship terms' are either the sole avenue through which the classification of kin can be . established or constitute a major or decisive ■ body of evidence on that problem. One might raise the question of whether, perhaps, the American 'kinship' system is unique in that it is the only one in the world where the genealogical grid is inappropriate for cultural comparison. I am sure that you will agree that this does not seem to be so. I can assure the ■ reader that from my own work on Yap, the Mescalero Apache, and the Zuni... the genealogically defined grid does not fit these cultures either. I would suggest that the Nuer cannot be fitted to a genealogical grid, nor most of the Eskimo systems we have adequate information on. The important point is that the genealogically defined grid is the only analytic device that has been applied to most of the systems which anthropologists have studied. There has been almost no systematic attempt to study the question without employing this device. To put it .:simply, it is about time that we tested some other hypotheses. [... ] IV We are ready now to deal with the question which is the title of this paper: What Is Kinship All About? . The answer is simple and self-evident by now. : 'Kinship' is an analytic category which has been "prevalent in anthropology since Morgan first invented it. In the way in which Morgan and his followers have used it, it does not correspond to any cultural category known to man. The closest thing to it is the Western European category of 'family,' but if I am correct in my analysis even that is not close. From the beginning of this paper I have put the word 'kinship' in quotes, in order to affirm that it is a theoretical notion in the mind of the anthropologist which has no discernible cultural referent in fact. I have consciously misused the term 'kinship' simply as a way of beginning the discussion. But it is no longer necessary to misuse the words; now we can use them correctly. 'Kinship' is what Morgan's, Goodenough's, Lounsbury's, Levi-Strauss', Leach's and Need-ham's {among others) analytic schemes are all about, but they have no known referent in any known culture except at the conglomerate level of Western European culture, as in America. To speak precisely, the title of my book, American Kinship, is a misnomer. I really did not deal extensively with 'kinship' at the conglomerate level nor did I intend to; in the pure culture level there is no such thing as 'kinship.' Hence my use of the term 'pure kinship level' is wrong, too, which I have tried to suggest by the use of quotes around the word. The level is the pure culture level as defined by certain symbols. Let me conclude this section on a simple note. For a while anthropologists used to write papers about Totemism as if it were a concrete or conceptual entity that had an actual, existential counterpart in the cultures of the Australian aborigines, among others. Goldenweiser and others then demolished that notion and showed that totemism simply did not exist as a useful analytic category precisely because it had no corresponding referent in any of the cultures with which it was alleged to be associated. It became, then, a non-subject. In due course Levi-Strauss wrote a book about that non-subject, in which he first explained that it was a non-subject and therefore could not be the subject of the book, for it did not exist outside the minds of those who invented it and believed it, and these were anthropologists, not the natives they wrote about. The 'matrilineal complex' suffered the same fate in the hands of Lowie. In my view, 'kinship' is like totemism, matriarchy, and the 'matrilineal complex.' It is a non-subject. It exists in the minds of anthropologists but not in the cultures they study. If you like to think that I have devoted a good part of my intellectual life to the industrious study of a non-subject, you are more than welcome to do so. If you think that I have now talked myself out of a subject for study you are quite right, too. But that is not the whole story. I have talked myself out of studying 'kinship' as 270 DAVID M, SCHNEIDER if it were a distinct, discrete, isolable subsystem of every and any culture. What I have learned and have tried to convey here is that in the study of culture one must proceed in a very different way. When I started to study American 'kinship' I went to households to talk with the inhabitants about how those who were living there were related to each other and to others who were not living there. I systematically collected genealogies at the very outset. When I began to discover that their concepts were somewhat different from those which traditional 'kinship' studies led me to expect, I followed their concepts and their definitions and the formulation of the cultural domains of their actions, depending as well on my own experience over the past years here in America, on Yap, and among the Mescalero Apache. Once that was done, and it was not easy for me to do it systematically, I could see that there was no such thing as 'kinship,' except as it existed as a set of a priori theoretical assumptions in the mind of the anthropologist. One must take the native's own categories, the native's units, the native's organization, and articulation of those categories and follow their definitions, their symbolic and meaningful divisions wherever they may lead. When they lead across the lines of 'kinship' into politics, economics, education, ritual, and religion, one must follow them there and include those areas within the domains which the particular culture has iaid out. One does not stop at the anthropologist's arbitrarily defined domains of 'kinship,' 'religion,' 'ritual,' and 'age-sets,' etc., but instead draws a picture of the structure of a culture by means of the categories and congeries of units which the culture defines as its parts; one interrelates these in terms which, in that particular culture are symbolically defined as identical, drawing distinctions among parts which that culture itself defines as different by their different symbolic definition and designation. Proceeding this way, a somewhat different analysis emerges than when one asks questions about the social system or the social organization. Yet the two systems, as I have said all along, articulate and are related to each other. Ultimately the study of culture can no more be isolated from all other sub-systems of a society than the study of its social system, althou this is the way we have been proceeding |n [t" past, largely neglecting or omitting the sti,< ■' J culture or relegating the idea of culture to the kinds of hats the natives wear or, correspond ingly, to the level of arts and crafts it nas achieved. [... ] V I will try to briefly summarize this paper and '' what I take to be its major points, and add one : point in conclusion. Theory suggests that it may be useful to systematically and rigorously distinguish evil---* ture from social system, defining culture rather narrowly as a system of symbols and meanings "■-When this view of culture is applied to what have ordinarily been treated as 'kinship' systems, new material emerges because a-new question has been asked about it. Instead'1 of the classic question which is at the socials system level of How Does This Society Organ- ■ ize to Accomplish Certain Tasks (establish, alliances, maintain control over territory,'-pro-.-' vide for inheritance and succession, hold arid-transmit property, etc.), a cultural question1-; is asked: namely, what are the units, how are they defined in the native culture itself, how does it postulate their interconnections, their mode of differentiation, by what symbolic devices do they define the units and their relationship, and what meanings do these have? I tried to give an example, briefly condensed-.-' from published literature, of what happens when a particular 'kinship' system is analyzed in this way, using my own work on American 'kinship,' and I think I was able to show that some rather new and different results emerged; ' One of the lessons derived from this study uf American 'kinship' was that the very same symbols defined 'kinship,' nationality, and religion at the cultural level and that, if this were:' so, then all of these - with the possible addition;: of the educational system in American culture:;. - could be included in one single cultural unit or domain. Hence there need be - there could be - no grounds for distinguishing the 'kinship' system from the 'religious' system, from the . 'nationality' system, from the 'educational'.-system at the cultural level. WHAT IS KINSHIP ALL ABOUT? 271 Further, where the bio-genetic elements, the ■' jcinerrt5 of conception and parturition were ■'^icen simply as defining elements or were ^e3xed as states of affairs with which every socit-t>' must cope in one way or another, the jtcrmif- strategy of study which I commended ■ |JL,J the suggestion that these defining elem-j'lpfsof 'blood,' of one flesh and blood, of bio-netjc identity could be understood as syni-'■■bbis which stood for social relationships of jjifute, enduring solidarity. That is, the bio-:: jpjjical elements which previous theories took s merely defining features, 'givens' in the state of affairs, could be understood better as sym-■: hols f°r kinds of social relationships, and probably these did not derive from, not stand ?for, the biological material they purported to : "order functionally. Indeed, at many points the ■/scientific facts sharply contradicted the cultural ■facts about biology; but the fact that the scientific facts had little or no discernible effect ,-,n changing the cultural facts seemed good evidence for concluding that the bio-genetic ^elements in American kinship were primarily symbolic of something else and hardly 1 relevant to biology as a natural or actual state of affairs. ■ The next step in the argument was simply to generalize from that fact. 'Kinship,' from the time of Morgan, had been defined in terms of consanguinity and affinity, that is, by an a . priori set of criteria, and studied with respect to the organization of its elements for discharging certain functions. If 'kinship' is studied at the cultural level, however, then it is apparent that 'kinship' is an artifact of the anthropologists' analytic apparatus and has no concrete counterpart in the cultures of any of the societies we studied. Hence the conclusion that 'kinship,' like to-temism, the matrilineal complex, and matriarchy, is a non-subject, since it does not exist in any culture known to man. I then tried briefly to show that even those who seemed to have broken with the Morgan tradition - Rivers, Leach, Needham, and Levi-Strauss - were still ensnared in that tradition either by their commitment to genealogical criteria in the definition of kinship or by their commitment to the positing of questions purely at the social system or social organizational level, or both. I used Lounsbury and Good- enough as examples of those who were without question squarely in the tradition of Morgan. And finally, embedded here and there in the paper is the plea to try, for a change, another approach to kinship, another set of hypotheses, to ask another question and see what the payoff might be. We have asked these functionally defined, social organizational questions of kinship exclusively since the 1870's. There is no need to stop asking those questions for they are good, productive, legitimate questions. I only urge that we ask a different kind of question, a cultural question, as well. In conclusion, if the argument of this paper has any merit, it follows that it will no longer be possible to study 'kinship' or religion or economics or politics, etc., as distinct cultural systems, for in each case the definition of each of these domains has been in social system or sociological and not cultural terms. This has been the classic Weberian approach,10 where the basic frame of reference is the institution, socially or sociologically defined, and the two different questions, one organizational and the other cultural, are then put to the data. (Indeed, one of the favorite Weberian questions of recent times has been of the effect of religious organization and its cultural aspects on the development of the economic system.) The result of this Weberian approach is a fragmentation of the cultural material into artificial segments which remain unlinked and unlink-able. It is not possible to relate the cultural aspects of the religious system to the cultural aspects of the 'kinship,' political, or economic system without extraordinary skill and good luck, if it is possible at all. If the argument I have presented here is followed out logically it will be necessary to treat the whole culture as a single cultural system, following out its different segments and its different divisions and domains as these are defined and differentiated by the symbolic system itself. It follows from the irreducibility of the cultural to the social systems, or vice versa, that this examination of the cultural system as a whole, apart from its social system aspects, is necessarily undertaken independently of any examination of the social system, at least in its initial phases. Ultimately, of course, as the Parsonian theory of action makes so clear, 272 DAVID M. SCHNEIDER these all come together and are mutually interdependent parts of any concrete system of action, but they are analytically distinct. As no one system can be reduced to any other each system therefore has an integrity of its own which must be respected by the analytic procedures used. It is precisely this failure to distinguish the social system aspects from the cultural aspects and the primary analytic emphasis on the social system to the neglect of the cultural that has led us - the descendants and followers of Morgan -into such untenable conclusions as I have tried to deal with here - that because in some sense genealogy and procreation and conception are 'really out there as indisputable and unavoidable facts of life' it is and it must be the material out of which kinship systems are made. To my mind it will no longer be acceptable to consider 'religion as a cultural system' any more than it would be acceptable to consider 'kinship as a cultural system' or 'politics as a cultural system.' Each culture must be approached apart from its institutional segments, its social organizational segments, or its social structural segments, and from a purely cultural perspective. Once the cultural system as a whole is outlined - at least in its more or less broad outline, with its major symbolic features defined and the links between them roughly established - then, and only then, can such questions be usefully raised as, for instance, the role of the culture of a given society on its economic development, its religious organization, or its political system. But I would stress the importance of undertaking cultural analyses which are truly and clearly independent of the sociological analyses and uncontamin-ated by them. This is not the place to elaborate this last point but only to make clear that if the analysis of this paper has any merit, then the independent study of the culture of a society as a whole culture must be undertaken apart from and uncontaminated by the study of its social system. NOTES 1 I would like to acknowledge with gratitude the helpfulness and the many useful suggestions of Dr. Priscilla Reining, the fact that Mr. James A. Boon was kind enough to :-''^ the first draft of this paper when my heaTf prevented me from doing so, for Mr! g0 , • many useful suggestions and criticising*"-well as those of Mr. Carlos Dabezies, aiKj ^ the long and useful letters on the first draft; this manuscript written me by Mr. Micha Silverstein, Dr. Paul Kay, Dr. Roy d'Artdrade Dr. Edmund Leach, Professor Claude I gv;' Strauss, Dr. Ward H. Goodenougti, and Dr. Judith Shapiro. In addition, I would tnen tion again my seminar on Culture Theory 0f spring, 1971, the students in the Departmerr of Anthropology who heard and discussed the' first version of this paper, making many hclp-ful suggestions, and the students at the Uni-':: versify of Minnesota who also listened patiently and made acute and perceptive observations and suggestions which 1 have corporated without further, more specific, acknowledgment. 2 Unless, of course, one takes the position that ■ marriage is necessarily entailed in the notion1 of descent and therefore all that is needed is-one single component, parenthood. 3 One of the best contemporary statements, of this position is contained in two papers by-Ernest Gellner, 1957 and 1960. 4 Parenthetically, I should note that the cultural, system can be abstracted either from the nop mative system or directly from the level.of., observable behavior, for it is a distinct aspect of each. Methodologically the situation may, be such that it is easier to abstract the cultural' material from the normative system, and I think that this is often true. Furthermore, it is a useful methodological device to treat the: normative system and the system of observed-behavior as relatively independent sources:of material - they cannot be completely independent, of course — so that the cultural ma'-.' terial abstracted from the normative system-can be checked against the cultural material' abstracted from the patterns of concrete .behavior. If these two sources do not ) ield the same cultural material, the analyst is alerted to the fact that he has a problem on his hands; for if every cultural premise is embedded in the normative system, and the normative system plays a role, though by no means die, only or' even the decisive role, in concrete action, then the cultural aspect ought to appear in both and not only in one area,. Finally, it should be noted that some parts of WHAT IS KINSHIP ALL ABOUT? 273 jj,e cultural system are constructs of the observer which deal with implicit, covert, or unconscious categories while others can be formulated directly from the natives' own, expli1^1 model itself. For further discussion 0f these points see my "American Kinship" (1968), especially Chapter One. $ AH questions are really functional. When the ■ question deals with the relations between the parts under given conditions it is a structural question. When it deals with the relations between parts taken over a period of time and with reference to their change and interaction, then the question is processual. Hence the popular term 'structural-functional' is a fundamental misunderstanding as well as a misnomer. Since all questions are functional, some structural, and some processual, then all questions are either functional-structural or functional-processual, and it is a mistake to call a kind of theory 'structural-functional.' '-.'/ See Parsons (1970) on this point. I cannot think of a single work on 'kinship' ■■:■-.> which has systematically done this. Instead, vi . the assumption is made that consanguinity '■'{. and affinity define 'kinship' and, therefore, if a bond of either sort can be shown to obtain, .:■ ■ then by definition those are kinsmen. This is a good example of the difference between asking a social from a cultural question. The . .: externally given criterion, a definition of 'kin-ship' taken from outside the culture, is used rather than a definition of 'kinship' elicited ;:. from inside the culture itself. :;\7 The fundamental reference here is Schneider ■■ (1970); see also Schneider (1965). Note the ■K-. discussion of McLennan above as well. The -.'.- point is fundamental, since the assumption has been widespread if not universal since Morgan that the mode of classification of kin is embodied in the kinship terminology :•■■} and can be derived from no other source. As I have suggested (Schneider 1968, ch. 2, 5), ::: there are other, more reliable as well as valid v ways of deriving the classification of kin than ;:- by the use of kinship terms. I am obliged to : ■ Michael Silverstein for pointing out to me that I had failed to make this point clear in earlier drafts of this paper. 8 I am roughly summarizing Schneider (1969) here. There is, however, a fundamental error in that paper which I want to acknowledge here but which I cannot correct fully since there is hardly space in which to do so. First, let me acknowledge that my seminar in Culture Theory in the spring of 1971 drove home to me the fact that there was an error involved in this formulation; second, that Talcott Parsons also pointed out the error and that the solution emerged in conversations with him during that same period. The problem is that birth in a country is not quite comparable to birth by a mother in American culture. The word birth is the same but two different meanings seem to be implied. Second, nationality is really a modern invention and the presumption implicit in most of the work I have done on American kinship is that these are fundamental cultural categories of long standing and considerable stability. That the concept of nationality seems to fit so easily does suggest I am not far off the target. The solution seems to be to treat the third element in the triad not as nationality but as something like Durkheim's 'moral community,' the group sharing the same cultural system constituting one society. This may at certain levels be a nation, at others but a region of a nation or a smaller unit, or at certain levels supra-national with an ethnic or racial or religious reference, as the Jews and Moslems and the Buddhists or Christians at certain levels constitute such a moral community. The second aspect of the solution centers on series of symbolic equations between birth, blood, and land or place or locality. It is the analog in some instances of the "Where Ya From?" questions which strangers sometimes ask of each other; but I must forego spelling out the ways in which common blood and common soil or land are treated as equivalent under certain conditions in America. I hope to be able to publish a correction of this in the near future but until then will leave things stand here as originally presented. 9 It is even a moot question as to whether the symbols derive from the facts of nature and the facts of biology as these can be demonstrated scientifically. What is indisputable is that the symbols are formed of elements which in native culture are defined as biological, particularly as aspects of the reproductive process. What is disputable is whether they in fact derive from, or mirror, or are models formed after the scientific facts of biology. I do not think that they are, but this is a subject best left to another time. 274 DAVID M. SCHNEIDER 10 Of which C. Geertz (1966) is a clear example. Schneider (1968) also starts from such an institutional beginning but does not attempt to relate the cultural and social system aspects, only to isolate the cultural aspects. References Cited Geertz, C. 1966 Religion as a Cultural System. In Anthropological Approaches to the Study of Religion. ASA Monographs 3:1-49. Gellner, E. 1957 Ideal Language and Kinship Structure. Philosophy of Science 24:235-242. -1960 The Concept of Kinship. Philosophy of Science 27:187-204. Goodenough, W. H. 1970 Description and Comparison in Cultural Anthropology. Chicago: Aldine. McLennan, J. F. 1886 Studies in Ancient History. Comprising a Reprint of Primitive Marriage. London and New York: Macmillan. Morgan, L. H. 1871 Systems of Consanguinity and Affinity of the Human Family. Washington: Smithsonian Institution. Vol. 17 of the Smithsonian Contributions to Knowledge. (First printed 1870.) Parsons, T. 1966 Societies. Englewood Cliffs: Prentice-Hall. -1970 Some Problems of General Theory in Sociology. In Theoretical Sociology. J. C. McKinney and E. A. Tiryakiarj New York: Crofts. -1971 The System of Modern Societies Englewood Cliffs: Prentice-Hall. ..; ' Schneider, D. M. 1965a Kinship and Biology \n Aspects of the Analysis of Family Stuictute. A. J. Coale et al., Eds. Princeton: Princeton University Press, pp. 83-101. -1965b American Kin Terms and Terrns fw Kinsmen: A Critique of Goodenoughs Componential Analysis of Yankee Kinship TCJ._ '■ minology. In Formal Semantic Analysis E. A. Hammel, Ed. American Anthropologic:'. 67(5, pt. 2):288-308. -1965c The Content of Kinship. Man log. 122-123. ' -1965d Some Muddles in the Models: Or How the System Really Works. In The Relevance of Models for Social Anthropology.': London: Tavistock, pp. 25-85. -1968 American Kinship: A Cultural A