Rethinking Navajo Social Theory

Hal Schef*ler, in arguing that native concepts about procreation provide the basis for kin reckoning universally, presented considerable evidence for his argument, in addition to the extension rules for which he is best known, This essay applies this evidence to the Navajo materials and shows that a Schef*lerian analysis is correct. By contrast, the analysis of Navajo kinship by Gary Witherspoon, indebted to David Schneider’s ideas, is shown to be wide of the mark. Schef*ler also argued, in much the same logical vein, that gender classi*ication around the world is bipartite, that claims of a “third sex” are without merit. The argument is applied to “third sex” claims by Wesley Thomas, which claims are shown to be baseless.


Introduction
The Navajo are easily the most ethnographically studied people in the world. Their ideas 1 about kinship were relatively well known (see references) even before the publication of Gary Witherspoon's (1975) Navajo Kinship and Marriage and his other publications (see below).Witherspoon's analysis rejects the idea that Navajo kinship is grounded in the procreative relations within the nuclear family. As such it is self-consciously indebted to David Schneider's writings. Here it needs to be recalled that these writings, especially A Critique of the Study of Kinship (Schneider 1984), alleged that Western ethnographers applied to non-Western forms of kin-reckoning ideas based only on their own ethnic background, which, he maintained, are largely inapplicable to these exogenous materials. I shall call this "the Schneider Narrative," repeated, seemingly endlessly, by Witherspoon (see below) and Schneider's other admirers (e.g. Carsten 2004;DeMallie 1994;Schweitzer 2000). It there-This being so, I should add that a complete coverage of what has been published on the Navajo is virtually 1 impossible. I have tried to read most of the more frequently cited books and articles, but I still welcome correctives to (or con*irmation of) my analyses from those better acquainted with the pertinent literature. fore should come as no surprise that Schneider provided an Introduction to Navajo Kinship and Marriage, calling it "the de*initive statement" on its subject (1975:vii).
By contrast, Harold Schef*ler is best-known for a theory of kinship that grounds kinship phenomena in nuclear family concepts and relations. This grounding was made explicit early in Hal Schef*ler's career when he noted that different societies perceive "the facts of procreation" differently, … but they differ only within certain clear limits …. All such theories provide for the existence of a "genitor" and "genetrix" (parents), their offspring, who are related to one another (as "siblings") … and … for the existence of … genealogical relations [among] such persons. Kinship as a cultural phenomenon has to do *irst and foremost with any particular person's … relationships with other persons as these are … conceived to result from what his culture takes to be "the facts of procreation." … [I]t is the triad self-genitor-genetrix that should be considered to be the "atom" of kinship. For it is that unit which "generates" the elements "brother" and "sister." … Clearly, then, the elementary relations of a kinship system are parent/child, husband/wife, and sibling/sibling … These are of course the constituent relationships of the nuclear family (Schef*ler 1966:83-84).
The main purpose of this article is to show that Schef*ler's ideas are in fact closer to those of the Navajo than Witherspoon's are, despite the latter's repeated assertions that he has captured what may be dubbed "the native's point of view." I hasten to add that I have great respect for Witherspoon's painstaking analysis of Navajo metaphysics (see esp. Witherspoon 1977;Witherspoon and Peterson 1995), and even more respect for his personal devotion to the Navajo. His analysis of Navajo kinship, by contrast, is, seriously *lawed -an assertion I intend to justify in most of the rest of this article. From there I proceed to Navajo notions of gender, because -and this is less well-known -Schef*ler broached gender classi-*ication, and because his analysis here is of a piece with his arguments about human kinship. Finally, I shall have something to say about scholarly standards, which, I shall argue, the many admirers of David Schneider have violated with a vengeance.

Prolegomenal Matters
But *irst I need to go into more detail on Schef*ler's arguments. His central idea is focality, which I shall mostly render as semantic centrality. Although he is best known for applying this idea to genealogically-based extension rules, about which I shall have more to say later, there are at least six more applications of this concept in his published corpus.
(1) In his analysis of kin classi*ication on Choiseul Island, now part of the independent Solomon Islands (Schef*ler 1965), he points out that although more distant kin are super*icially classed with one's parents, siblings, children, and other close kin, these latter are said, in native parlance, to be the 'true' members of their respective kin classes (Schef*ler 1965:75, 81). In other words, close procreative kin provide the foci of the pertinent kin classes, a set of models for these classes, to which other members of the classes are likened as their extensions. These conclusions are underscored by the Choiseulese expression lavala. Here, Schef*ler (1965:750) instances one's actual father, to whom this expression can never apply. Instead, the father can be referred to as mama, i.e. by the lexically unmarked 'father' term -much as when an English speaker speaks of my father he/she is not referring to a Catholic priest. Schef*ler (1965: 75) adds: "That mamae denotes true father … ("blood" in Choiseulese terms) is further attested to by the fact that subsequent husbands of the mother (step-fathers), even though they may say that their stepson is 'just like a true son,' are nevertheless always quali*ied by the lavala designation …." Lavala signals diminution from a focus. Usage of this term indicates someone other than the Real McCoy. It 2 seems reasonable to suggest that lavala means "classi*icatory." It should be emphasized that this analysis is based upon native semantic distinctions: it is by no means an exogenous imposition on such distinctions.
(2) Further evidence for semantic centrality is presented in Schef*ler's re-analysis of some of Louis Dumont's materials on the Tamil speakers of South India (Schef*ler 1977). Here he -Schef*ler -points out that the kin-terms applicable to other members of pertinent classes are locally rendered by reference to the closest procreative kin of those classes. Thus e.g. the designation of classi*icatory 'fathers' bear pre*ixes which subclass them as 'small fathers' or 'great fathers', depending on their ages relative to one's genitor, the latter natively designated, as on Choiseul, as one's 'true father.' 3 (3) Still another piece of evidence for the semantically central status of native procreative notions in kin-reckoning stems from David Schneider's argument that American kinship has two "distinctive features" -"common substance" and "code for conduct" (Schneider 1968). Schef*ler (1976) argued instead that normative prescriptions pertinent to kin categories -in America and elsewhere -are dependent upon criteria which de*ine those categories in the *irst place, and that there is no reason to question the conclusion that, again throughout the world, these criteria are founded on local ideas having to do with procreation.
(4) In his Choiseul analysis, Schef*ler (1965:69, 91) refers to native expressions which he translates as 'father's side' and 'mother's side.' This indicates that, although kin terms are applied more widely, their use without lexical marking indicates one's actual parents -much as when I say my mother I am not referring to my Cub Scout den mother when I was a boy, or when I say my father I do not have in mind the man who superintended the funeral service I attended not long ago for an Irish-American friend.
(5) Several more examples can be found in Schef*ler's re-analysis, with Floyd Lounsbury, of concepts of parental connection among the Siriono of the Bolivian Amazon (Schef-*ler and Lounsbury 1971), initially brought to anthropological attention through a wellknown monograph authored by Allan Holmberg (1969). Thus, Schef*ler andLounsbury (1971:40) tell us that the Siriono Research in semantic theory shows that categories are graded rather than dichotomous (D' Andrade 2 1995:104-21). But this can be ignored for present purposes since a simple focal/nonfocal distinction will suf-*ice.
The South Indian terminologies are especially apt partly because Dumont (e.g. 1957) claims that they dis 3 play "inherited af*inity" in addition to consanguinity. In contrast, Schef*ler argues that all of the terms are fundamentally consanguineal. Moreover, Dumont's position was accepted by Schneider, as evidenced by the latter's high opinion of what he dubbed "alliance theory" (Schneider 1965) and by his written communications with me about a dozen years before his passing in 1995. It seems reasonable to suggest that this opinion owed a very great deal to Schneider's anti-genealogical approach to kinship. It is well worth noting, *inally, that, more recent ethnography on the South Indian area provides overwhelming support for Schef*ler's position (e.g. Busby 1997;Clark-Deces 2014:39-41;Good 1980). impose a number of food taboos on pregnant women and … their husbands, the presumed genitors of their unborn children. They also practice a form of [post-natal ritual] and … rationalize it on the ground[s] that there is some sort of "intimate connection" between both parents and the child, such that an event affecting them will affect it, too … Such beliefs and customs are virtually incomprehensible unless Siriono presume some sort of substantial and enduring connection … between the child and its presumed genitor and genetrix. Because of their nature, these connections cannot be undone, and for each person there can be one and only one set of such relationships. They come into being with him [or her] and cannot be created at any other time (e.g. through adoption).
Employing a dictionary written by a German missionary to these people, Schef*ler and Lounsbury (1971:47) perform a morphological analysis of a native kin term, edidi which Holmberg glosses as 'child' … meaning "son" or "daughter." … [Although this term is employed in a "classi*icatory" manner it] *igures in utterances such as se eriiri atea, "to conceive (or engender) a child"; … girri teakula,"to give birth"; erirri ubua he or eriiringi, "fetus"; tiri, "one who was joined with her (in the stomach) …"; and tiri-mbae, "one who has or can have no children, viz., sterile." Thus, although the 'child' term is used more widely, its semantic associations make it clear that its procreative signi*icance is semantically central.
(6) Schef*ler's initial employment of extension rules was almost certainly derived from Lounsbury, at one time his senior colleague at Yale (Shapiro and Read 2018). Critics (e.g. D' Andrade 1970;Tyler 1966;Wallace 1965) were quick to point out that these rules presupposed a universal genealogical grid for which there was no real evidence, and that they were, in the parlance of the day, not "psychologically valid." Thus, as discussed by Read (2018), who employs an example from my own *ieldwork among the Aboriginal people of northeast Arnhem Land, Australia, it is grammatically possible, assuming the focal status of close procreative kin and employing appropriate suf*ices, to formulate, in one of the local dialects, a construct translatable as 'mother's brother's son or daughter.' But when I presented this construct to my informants, asking them for the pertinent kin-term, they were utterly *lummoxed. By contrast, they were entirely at home when I asked about 'someone sired by a man I call 'mother's brother' (Shapiro 1982:274-76). More generally, beyond the closest procreative kin they plainly reckoned kin-term assignment on the basis of relatively clearly stated matri-and patri-*iliative rules (Shapiro 1981:34-38), which perforce are psychologically valid. As discussed by Read (2018), there are clear English parallels. Thus, the son or daughter of someone I call "uncle" I call "cousin," but it is not necessary to know whether this "uncle" is my MB, FB, FZH, MZH, or "just a friend of the family." Schef*ler, for his part, was aware of such matters of psychological validity. The key publication in this regard is his analysis of kin classi*ication of the Baniata, who, like the Choiseulese, inhabit the Solomon Islands (Schef*ler 1972). This is such an important contribution that it merits attention in some detail. Here he notes that the [extension] rules posited [here] are not merely my analytical constructs, for the people … use many similar rules, both of de*inition and extension, and can readily state them. Of course, they do not express them in exactly the same way as I have.
For the most part, their rule-like statements … are more speci*ic and less general … than the rules [I state]. For the parent, child, and sibling terms the tendency is to specify their [focal] senses discursively by statements such as "my ['father'] is the man who engendered me," my ['mother'] is the woman who gave birth to me," and my ['sibling'] and I were born together; we have the same mother and father" (Schef-*ler 1972:350).
Note that the unmarked use of native kin terms signals the closest procreative kin: item (4) above. Such kin may be marked by a term which Schef*ler (1972:354) renders as 'real, true, proper, or par excellence': item (1) above.
Schef*ler (1972:367-68) further tells us that kin term product statements "may be used … to specify extended ranges of terms. For example, the people say "the ['brother'] of my ['father'] is my ['father'] also …." Moreover, some Baniata made statements of the following sort: "my father's brother is like my father and my mother's ['sister'] is like my mother, so their children are like my brothers and sisters" (Schef*ler 1972:369; emphasis in original). All this is comparable to his Choiseul analysis -see item (1) above.
The foregoing is no more than an abbreviated selection from Schef*ler's publications of his attention to native notions of semantic primacy. It bears emphasis that none of this entails Lounsbury-inspired extension rules. Was he the *irst scholar to attend to these notions? Not at all! With the exception of the "distinctive features" contention (item 3 above), all of Schef*ler's arguments for semantic centrality have substantial precedent. The rendition of close procreative members of kin classes as, in local parlance, the 'true' members of these classes was noted by Alexander Goldenweiser more than eight decades ago in the second edition of his introductory text. To wit: [I]n many primitive tribes the terms used for the immediate members of the family are either distinguished from the same terms in the extended sense by the addition of some particle, or terms corresponding to "own" are used … Family is family, whatever the system of relationship … (Goldenweiser 1937:301).
Often this is associated with lexical markers translatable as 'false' or some other indicator of secondary semantic status, reserved for collateral kin. One or both sets of markers can be found in the systems of kin classi*ication of such well-studied societies as Aboriginal Hawaii (Handy and Pukui 1972:65-68), the Cree (Speck 1918:151), and the Mae Enga (Meggitt 1964).
Similarly, classi*ication of collateral kin by reference to the closest procreative kin (item 2 above) is exempli*ied in many systems of kin classi*ication, such as Bengali (Inden and Nicholas 1977), Malay (Banks 1974) and Tamil (Beck 1972:216). Cover expressions, like 'mother's side' and 'father's side'(item 4 above) or teknonymous usage, which imply that, when kin terms are used without lexical marking they refer to primary kin, have been reported for Fiji (Ravuvu 1971), Tikopia (Firth 1936), and the Aboriginal Australian people of northeast Arnhem Land (Shapiro 1981), among many others. The procreative signi*icance of kin terms, emphasized by Schef*ler and Lounsbury for the Siriono (item 5 above), was *irst noted more widely by Westermarck (1894:88 et seq.). Finally, the sort of native extension rules which Schef*ler noted for the Baniata have been called "kin term products" by Dwight Read (e.g. 2018), who provides several examples of their ubiquity in the ethnographic record.
Before the Navajo materials can be tackled directly, a *inal prolegomenal matter needs to be stressed. In all of the ethnographic cases with which he deals, Schef*ler is con-cerned with native notions of biological paternity, not what has been called "social paternity." He was particularly critical of the claims that in some societies paternity is assigned without question to the mother's husband, for example, by the payment of bridewealth, or there being no knowledge of the relation between sex and conception, such that paternity is ipso facto assigned the husband. In what is, by my lights and to this day the best general statement on human kinship (Schef*ler 1973), he points out that [i]t is often taken as a legal presumption … that a woman's husband … is the genitor of her offspring. But this presumption … is widely regarded as rebuttable. … [T]he ethnographic record offers little if any reliable evidence that [parental] relations established out of wedlock are ever totally discounted … (Schef*ler 1973:753-54) In a nutshell, native concepts of fatherhood are elements sui generis: they are not derived from relationships with mothers. This is germane to the Navajo case, as we shall see. In the same encyclopedic article, and referring to the "ignorance of paternity" claims for certain societies, Schef*ler (1973:749-51) refers to earlier ethnographic reports which make it plain that the alleged entry into the mother by a spirit-being is not held to cause conception, as the "ignorance" theory maintains, but rather is posited to occur at fetal quickening: it is a native theory of viviEication of an already formed fetus.
All this being so, Schef*ler is led to the following seminal conclusion: "The elementary relations of any kinship system are … those of genitor-offspring and genetrix-offspring … " (Schef*ler 1973:755). This of course is a re-statement of the conclusion Schef*ler reached seven years earlier (see above).

Navajo Kinship: A SchefDlerian Analysis
I can now turn directly to the Navajo materials. I begin with the sort of analysis which I think Schef*ler would make, focusing on the nuclear family as a residential unit and as a native concept, and to wider spheres of kin according to these concepts. From there I proceed to Witherspoon's analysis of these materials, which is in crucial respects different, and which, I shall argue, is very wide of the mark.
Some of my sources (Aberle 1980(Aberle :124, 1981Adams 1963:59-60;Shepardson and Hammond 1970:221, 5 236-37) speak of "lineages" among the Navajo, but these appear to be no more than contingent clusters of uterine kin. There are no native notions of "lineation" (see esp. Lamphere 1977:90-93) such as we *ind among the Somali (Lewis 1994) and some other systems in northeast Africa and adjoining parts of southwest Asia. The matter has been theorized by Adams (1983:411), Murphy (1979) and Shapiro 1971). over, not long after birth, the infant is placed on a cradle board which may be used by subsequent children "but not out of the family" (Bailey 1950:79; see also Lamphere 1977:70;Schwarz 1997:114).
Conception in the *irst place is said to be caused by "[w]ater from a man and (menstrual) blood from a woman" (Bailey 1950:18; see also Leighton and Kluckhohn 1947:13;Schwarz 1997Schwarz :69, 117, 259, 2007Wright 1982:388;Wright et al. 1993:785). Another -perhaps contrastive or complementary -account has both parents contributing "water" to form the fetus (Bailey 1950:19). Finally, it is pertinent to note that Navajo paternity is in no wise based on marriage: thus, in instances of adultery the child is af*iliated with its genitor's matri-clan, which may or may not the matri-clan of the woman's husband (Aberle 1961:129-30; see also Shepardson and Hammond 1970:169-70, 189-91). This should be kept in mind in considering Witherspoon's argument (see below) that the Navajo father is construed as an af*ine.
There is considerable variation in Navajo kin classi*ication (see esp. Freed and Freed 1970;Landar 1962), but most of my sources indicate that, although parents' same-sex siblings are super*icially classed with the pertinent parent, they can be distinguished as 'older mother,' 'younger mother,' 'older father,' and 'younger father,' depending on age relative to the linking parent (Aberle 1961:197;Chisholm 1983:68;Downs 1972:23, 41;Freed andFreed 1970:1441;Shepardson and Hammond 197:88-89, 207-15, 219). We saw this pattern earlier in Schef*ler's analysis of Tamil kin classi*ication, and here again it indicates that biological parents enjoy semantically central status. Reciprocally, a man or a woman may distinguish his or her own children from those of his or her parallel siblings, real or classi*icatory (Aberle 1961:198). A variant of this pattern involves rendering all parental same-sex (and sometimes opposite sex) siblings as 'little mother' or 'little father,' with a comparable argument regarding semantic centrality: indeed this is made even plainer by Herbert Landar's translation of 'little father' as "one who is a minor replica of my father" and 'little mother' as "one who is a minor replica of my mother" (Landar 1962:989). As if all this were not enough, Landar (1962:997) notes an instance in which one's own sister is rendered in the Navajo language as the 'real' member of her kin class -this being established by the native query "Are you both of the same mother?" There is also teknonymous usage, such that (for example) an individual may be referred to as "X's father," in which case one's biological father is intended and not just any member of his kin class (Reichard 1928:100-01, 105;Shepardson and Hammond 1970:119; see also Landar 1962:993-94). This too implies focal status for close procreative kin, much as, once again, when I say, my father, I refer to the man who engendered me and not to a Catholic priest.
Step-parents, for their part, do not merit the focal parental terms: they are classed either by age relative to the linking parent or as af*ines (Aberle 1961:200, but compare Shepardson andHammond 1970:221).
Relative age *igures as well in Navajo kin classi*ication. For example, a (real or classi-*icatory?) mother's brother who is younger than Ego may be called 'sister's son' (Shepardson and Hammond 1970:217-18, 223). This suggests that the semantically central member of the 'mother's brother' kin class is older than Ego, very likely his actual MB (ibid.:220), and, conversely, that the conceptually central member of the 'sisters' son' class, is younger than Ego, likely his actual ZS (Ibid:223). 6 Outside Ego's own matri-clan and his father's matriclan but inside the sphere of close procreative kin, kin terms are applied on the basis of the kin classi*ication of the linking relative: thus, e.g. male members of one's FF's matri-clan are called 'father's father' (Shepardson and Hammond 1970:223-26). There is thus a Crow-like character to this sphere, suggesting that the FF is the semantically central member of his kin class (compare Schef*ler 1973:767). Comparable considerations apply to the MF kin class. And since, as already noted, Ego's mother and father are the foci of their respective kin classes, it follows that semantic centrality in Navajo kin classi*ication is provided by close procreative kin. 7 There is evidence for performance-based kinship among the Navajo, albeit this seems to be modeled on native notions of procreation. Thus, the Franciscan Fathers note that "it is reprehensible to marry a woman with whom one has concluded a friendly relationship by frequent visits and endearing terms …, as this is equivalent to consanguinity with her" (1910:433, emphasis added).

Gary Witherspoon's Schneiderian Analysis of Navajo Kinship and its Defects
It is, as I shall argue later, pertinent to point out that Witherspoon's writings have a decided "postmodern" *lavor. Both his *irst book, Navajo Kinship and Marriage ( Witherspoon 1975), and his second, Language and Art in the Navajo Universe (Witherspoon 1977) are replete with references to the Schneider Narrative (see below), and in his co-authored volume comparing Navajo art and science with their Western counterparts he maintain s that Western science is "just another way of knowing" (Witherspoon and Peterson 1995:109). Navajo Kinship and Marriage, despite its title, takes no account whatsoever of any of the evidence for semantic centrality in Navajo kinship. Moreover, and, remarkably enough, the book has virtually no data on kin classi*ication. Yet its author insists that his analysis is respectful of Navajo understandings, pedantic statements such as the following: Many anthropologists make studies of systems of kinship and marriage with too many questions answered before they get to the *ield. They already know that kinship is the apprehension or misapprehension of genealogical relationship, that Shephardson and Hammond (1970:223) indicate that kin terms are "extended … by analogy" from ego's fa 6 ther's "matrilineage" to less closely related members of his matri-clan, but exactly what "lineage" is to be construed as being here is unclear (see previous note), as also is what Navajo concept is entailed by "analogy." For a fuller treatment of the principles employed in Navajo kin classi*ication see Landar 1962;Reichard 7 1928:74-87; Shepardson and Hammond (1970:204-40). Carr et al... (1939) proffer an "alliance" rendition of Navajo marriage, wherein matri-clans are supposedly "allied" by individual unions, but they present no evidence for such an "alliance" other than clan exogamy. Such a rendition is of course common in the theoretical literature, which sees no need to inquire as to what rights are held by the allegedly "allied" units by virtue of an individual marriage. In fact, Navajo marriages are arranged by close kin of the woman and, to a lesser extent, of the man (Aberle 1961:124-26;Lamphere 1977:40;Reichard 1928:68-69;Shepardson 1995:166;Shepardson and Hammond 1970:86-87, 91, 170-76), and there is no evidence whatsoever that such unions entail rights and obligations for entire clans. The matter has been raised -and to my mind settled -for Aboriginal Australia, the locus classicus of "alliance theory," by Hiatt (e.g. 1968) and myself (Shapiro e.g. 1968). kinsmen are not af*ines and that af*ines are not kinsmen, … that residence is either patrilocal, matrilocal, neolocal, or some combination of these, … [and] that fatherchild and mother-child are both relationships of the *irst order … [By contrast] I try to avoid molding the Navajo system of kinship and marriage into the standard anthropological categories. Instead, I attempt to … reveal the Navajo system according to its own features … (Witherspoon 1975:xi). This is of course vintage Schneider. And, like its inspiration, it utterly misunderstands both the gist of previous studies (see above) and the nature of the comparative endeavor. Thus the utility of established categories was questioned two decades earlier in Goodenough's seminal article on residential programs on the Micronesian island of Truk (Goodenough 1956). Schef*ler argued as early as the mid-1960s (Schef*ler 1966) that father-child and mother-child are indeed "both relationships of the *irst order" -something already noted. And it was established years earlier that kinsmen can be af*ines as well, as evidenced e.g. by parallel cousin marriage in most of the Islamic world (e.g. Murphy and Kasdan 1959). Witherspoon thus evidences a highly truncated appreciation of the history of kinship studies. In fact Navajo Kinship and Marriage utterly ignores several decades of previous ethnographic theory: its only comparative reference is to Schneider's American Kinship. Even worse, it fails utterly to appreciate the nature of ethnographic comparison. Thus Goodenough's most important contribution was his argument that the comparative enterprise requires an external schedule of primitive elements by virtue of which comparison can be effected, much as the catalogue of phonemes in linguistics. Witherspoon tacitly 8 employs such a list by his insistence that it is Navajo kinship and marriage which is his concern: otherwise, he is left only with what he takes to be native understandings, unmoored to the logically prior notions signaled by the expression "kinship and marriage." In short, if such systems vary so profoundly; what justi*ies labeling them this way?
Most of Witherspoon's claims about Navajo conceptualizations are either self-contradictory, or unrelated to speci*iable ethnographic operations, or entirely false. Thus, despite his claim that cultures differ virtually kaleidoscopily in their kinship notions, we learn that Navajo render the sibling relationship as "those with whom one came up out of the same womb" (Witherspoon 1975:16). Now this particular idiom seems to me rather close to my own, allegedly biased Euro-American sense of siblingship. Admittedly, the Navajo conception of this relation may embrace more people than mine does, but if they apply this idiom to classi*icatory siblings too -it is not clear that they do -, then this is plain evidence that they construe it primarily in procreative terms. The point, alas, is entirely lost on Witherspoon.
Then there is his assertion that among the Navajo "[t]he acts of giving birth and sustenance are imbued with meaning …, and this meaning can be described as … diffuse, enduring solidarity " (1975:20). This too is of course straight from Schneider. Utterly ignored in this impressionistic rendition are the *ield operations to which it might conceivably pertain. It is not, presumably, a translation of a Navajo expression, but then what is it ethnographically? We are left without an answer.
Goodenough's argument here is relatively implicit in his Truk essay but more expressly stated in his later 8 publications (e.g. Goodenough 1970Goodenough , 1981. It should also be noted here that there are important connections between Goodenough's writings on ethnographic description and comparison and Schef*ler's thought (Shapiro and Read 2018).
The lack of rigor in Witherspoon's analysis is compounded by his treatment of the mythical *igure of Changing Woman, noted by other ethnographers as a sort of Primeval Female, associated with the earth (see esp. Schwarz 1997) and said to be in the Navajo language 'our mother' (Witherspoon 1975:20). About this apparently *igurative usage Witherspoon has this to say: Such references to kinship with the earth have previously been categorized as metaphorical. I believe that this reference to the earth as our mother is based on more than simply a metaphorical extension of the concepts of kinship. In Navajo mythology, the earth was the source of life for all beings through their emergence from the underworlds. In this act of emergence, the earth gave birth to all living creatures.… The emotional … tie between the Navajo and the earth is also strong and intense. Thus by every Navajo concept and de*inition, the earth is a … true kinswoman (1975:20; see also Witherspoon 1970:58).
No it's not! That the earth is said to "give birth to all living creatures" suggests strongly that it is metaphorically modeled on the birthing activity of *lesh-and-blood Navajo women (Franciscan Fathers 1910:354;Schwarz 1997:15). Elsewhere, Witherspoon (1970:58) seems to recognize this when he writes that "[a]ll beings had their origin in the underworlds and have their existence here because they symbolically came up out of the womb of the earth." In any case, there is no indication in the pertinent literature -not even rhetorically -that the earth is said to be a 'true' mother or otherwise ascribed semantically central status, which status is apparently enjoyed, as we have seen, only by one's biological mother.
Still, upon *irst reading Witherspoon's contention, I wondered if I could *ind a plainer statement of such modeling in the extant ethnographic literature. My search was only partly successful. I could *ind nothing of this kind regarding Changing Woman, but the Navajo residential edi*ice, the so-called "hogan" (an Anglicized version of a Navajo lexeme) is also said to be a 'mother,' and Maureen Schwarz (1997:44) quotes an informant's remark as follows: "This home provides a place for you to sleep and rest and eat and enjoy yourself, just like the way your … mother takes care of you" (emphasis added). In short, the mother provides a model for a rendition of the hogan, which is likened to a mother. Note too that when this informant employed the 'mother' lexeme, it was clear that the reference was to the biological mother, something we have come across before.
Presumably, the metaphorical rendition of Changing Woman as 'mother' is related to the equally *ictive notion that all Navajo are kin ( Witherspoon 1970:58) -something which is surely comparable to the Judeo-Christian idea that, as children of a single Creative Being, all people are brothers and sisters. In neither case do such embracive assertions gainsay kin/nonkin distinctions.
Then there is Witherspoon's assertion that "Navajo de*ine kinship in terms of action or behavior, not in terms of substance " (1975:21). This can hardly be the case because behavioral norms pertaining to kinsmen depend upon those criteria by which kin are separated from nonkin in the *irst place, and there is, as we have seen, no reason to argue that these criteria are not procreation-based. It is worth noting that this is much the same error that Schef*ler pointed out in Schneider's study of American kinship (see above).
There is more. Witherspoon tells us that "[t]he primary bond in the Navajo kinship system is the mother-child bond …" (1975:21). It is unclear that "primary" here has any op-erational meaning, and if we allow it intuitively, it is also unclear how Navajo kinship differs from kinship anywhere else. Mothers seem everywhere to be motherly.
A chapter of Navaho Kinship and Marriage is devoted to af*inal ties. Here Witherspoon asserts that "the primary symbol of af*inal solidarity is found in sexual intercourse " (1975:23). Thus once again there is indebtedness to Schneider on American kinship, and once again "primary" is used entirely intuitively. Moreover, in what sense can sexual intercourse be construed as a symbol? The symbols that at least some of us employ in everyday life, like the Christian cross or the red/green contrast in traf*ic lights, have a public character, but there is evidence that the Navajo, like most other people, regard sex as a decidedly private matter (Csordas 1989:478;Kluckhohn and Leighton 1974:91;Leighton and Kluckhohn 1947:88;Schwarz 2007:89). And Navajo men are enjoined markedly to avoid their mothers-in-law, which avoidance seems especially to entail profound restrictions on sexual activity (Aberle 1961:150-51;Bailey 1950:56;Downs 1972:38;Franciscan Fathers 1910:447, 449;Reichard 1928:71-72, 88, 140;Shepardson 1995:167;Shepardson and Hammond 1970:175;Spencer 1947:51).
In this regard, it is pertinent to consider Witherspoon's suggestion "that the Navajo father-child relationship might be best characterized by af*inity " (1975:30). This rests on the following considerations: First, "the intensity or closeness of the father-child relationship varies proportionally to the intensity or closeness of the father-mother relationship" (ibid.). Second, "Navajo children often refer to and address their father as an in-law. The term [translatable as] 'male in-law of a proximate generation' is often used in a joking context by a Navajo child in referring to his father " (1975:31). Third, at least according to some informants, a man may marry a woman of his father's matri-clan (ibid). Fourth, and *inally, marriage of a man to his wife's daughter by another man is permitted and, apparently, fairly common (ibid.) None of these arguments, I submit, is decisive. The emotional closeness of the mother-child bond, even if supported by non-impressionistic evidence, is irrelevant to the nature of any other relationship. If it were otherwise, and given the same reasoning, then (e.g.) the sibling tie, it could be argued, is one of af*inity. In fact Witherspoon's contention rests on his impression, already noted, that the mother-child bond is "primary." That a man or woman may, under certain conditions, marry someone of his/her father's matri-clan has no necessary bearing on the af*inal status of that clan with regard to Ego: it may simply signify that certain consanguineal relatives, as locally construed, may marry. In this regard It is pertinent to repeat that individuals of one's father's matri-clan are held to be 'kin.' Much the same argument applies to a marriage between a man and his wife's daughter by another man.
More interesting is address and reference to one's father as 'male in-law of a proximate generation.' This, as noted, has a ludic character, which is to say that it is not taken seriously. James Downs (1972:40) reports what appears to be the same playful usage: Joking with parents … usually refers to membership in clans. A boy can for instance refer to his mother as 'woman who married into my father's clan,' and thus speak of her as an in-law … Such joking might be called a play on kinship just as we make plays on words (Downs 1972:40; emphasis added; see also Aberle 1961:152-53).
If taken in earnest -as it should not be -, and using Witherspoon's reasoning, we might conclude from this that a Navajo regards his/her mother as an af*ine -a conclusion entirely at odds with his own analysis. He seems to recognize its shortcomings when he writes that "the Navajo father is related to his children by both kinship and af*inity " (1975:34). But even this modi*ication is unwarranted: we have seen that all the evidence for the af*inal status of the father is questionable, and that all ethnographers report that he is reckoned as 'kin. ' Finally, there is a Navajo myth in which an archetypal personage entertains "doubts about [his] paternity" (Mindeleff 1898:33) --which doubts would be unlikely if this "paternity" were established by marriage negotiations, the marriage ceremony, and the ensuing cohabitation, all of which are quite public (e.g. Aberle 1961:124-28;Franciscan Fathers 1910:422-23, 446-49;Reichard 1928:68-70;Shepardson and Hammond 1970:170-76).
Several scholars in addition to Schef*ler have pointed out the scholarly sloppiness of Schneider's American Kinship (Fogelson 2001;Kuper 1999;Wallace 1969). Yet Witherspoon is undeterred: Navajo Kinship and Marriage, I would suggest, is indebted at least as much to Schneider's undisciplined analysis as it is to anything reportable from the Navajo *ield.
In the Introduction to his second book, Language and Art in the Navajo Universe, Witherspoon states that "[a] kinship system is ultimately based on a theory of reproduction …" (1977:4). This is a surprisingly Schef*lerian conclusion: one can *ind several statements to this effect in Schef*ler's corpus (e.g. see above and Schef*ler 1966:83-84). Moreover, this volume has better data on Navajo kinship conceptions. Here Witherspoon provides the pattern, already noted, wherein the 'mother' and 'father' terms are modi*ied for collateral kin, who are rendered with reference to the linking parents. But he fails to see the implications of this *inding for his analysis, which otherwise has much the same relativistic and didactic quality as Navaho Kinship and Marriage. Thus, after an abbreviated coverage of Navajo creation myths, which involve cerebral and oral generation, we are told that "[t]hinking and singing the world into existence attributes a de*inite kind of power to thought and song to which most Westerners are not accustomed. It is rather obvious that the Navajo ontological conception of thought and speech is very different from our own" (1977:17; see also Witherspoon and Peterson 1995:16). But this is surely not the case. In the Book of Genesis God creates by speaking, and the Book of John begins with Jesus being rendered as The Word; in both instances there is an at least implied contrast with sexual generation (Shapiro 1989:and references therein). This has Navajo parallels: thus, Changing Woman is held to create by externalizing her non-sexual bodily emissions (Franciscan Fathers 1910:356, 427;Schwarz 1997Schwarz :62-67, 2007compare Munn 1970).There is, I suggest, more than a super-*icial parallel between Navajo ritual chants, which employ the bright colors of dry paintings (aka sandpaintings). and this presentation, which I dare to hope will be illuminating, perhaps even seminal.
Closer to the ground, there is the same contention that the mother/child bond is "primary" (1975:85), without any operational speci*ication as to what "primary" might mean, and the same charge that the attribution of kinship to those other than humans is metaphorical is simply a Eurocentric projection (1977:86, 91-94). We are told as well that "Navajo relations of kinship are … de*ined by actions and not substances" (1977:85), the latter, presumably, being the case among Euro-Americans. But one of these "actions" is held to be giving birth, which is surely signi*icant outside Navajo country. The other is "sharing sustenance," similarly widespread and in both cases, for reasons already noted, logically dependent upon prior local de*initions of kinship.
Witherspoon here provides an extensive discussion of the Navajo 'kin' term noted by previous ethnographers, telling us in the process that it allows kinship to extend beyond "those who share common substance" so as to apply to "any persons who act according to the behavior code of kinship," as well as, tactically, to anyone who an individual is trying to engage (1977:88; see also Landar 1962:988;Witherspoon and Peterson 1995:10). Once again, he fails to appreciate that such a code stems from a speci*ication of the criteria which posit kinship in the *irst place, and to see that many Euro-American institutions are closely comparable. Brother, can you spare a dollar?
At one point, Witherspoon seems to recognize that Navajo behavioral norms for kin are logically dependent on a prior statement of what kinship is in the *irst instance. Thus he notes that kin classes are "constructed and differentiated with regard to … giving birth and the relatively de*inite … attributes of sex, generation, lineality, and age. … [T]hese classes are strictly and solely de*ined by who gave [an individual] birth and for whom the birth was given [i.e. the idiom of paternal connection].… Ongoing behavioral relationships or the lack of them have absolutely nothing to do with one's membership in any of these classes … " (1977:95). Once again, we have a decidedly Schef*lerian statement, and one followed by an admirable analysis of the data pertinent to it. But this in turn is followed by the assertion that, for the Navajo, there are only two primary or unmediated relationships. These exist between husband and wife and between mother and child. The mother is the focal point in both of these relationships, for both involve a different relationship to her womb. It is she who provides the link between the father and the children … (1977:105-06).
In short, and again, the mother is "primary," the father not: elsewhere in Witherspoon's corpus the latter is rendered as "somewhat secondary, because [the relationship] is traced through another person" (Witherspoon 1970:59)." "The father-child bond," we are asked to believe, "is really just another extension of the husband-wife bond. It ties an inmarrying man with children who have the same cultural identity as the mother" (Witherspoon 1959:59). All this, alas, runs counter to the Navajo notion that a man contributes 9 something irreducible to the generation of his children, something on an ontological par with the mother's contribution (see above). Witherspoon (1977:196-97) notes this, but is too intent on minimizing Navajo fatherhood to see its implication.
More broadly, Witherspoon seems intent on representing the Navajo as exemplars of some of the primitivist fantasies of our times: as in so-called "matriarchy theory": mothers are somehow "primary" and there is kinship with nature, and among all people -or at least all Navajo. By contrast, we are asked to accept the following: In the West, theories that explain the origin and nature of society assume the individual as the fundamental reality. Societies arose during a hypothetical time of war of individuals, all against all. The warring individuals at some point came together and decided to form a society to govern themselves. This is known … as the social All this is reminiscent of Malinowski's claim that the Trobriand father is said to be a 'stranger' or an 'out 9 sider' with regard to his children. I have elsewhere (Shapiro 2018b) argued that this is a metaphor, and that he is in fact regarded as a kinsman. The cited article should be consulted for a detailed argument and pertinent references.
contract. … Society is [thus] a contingent arrangement of individuals that partially alters the natural state of individuality (Witherspoon and Peterson 1995:9). This Hobbesian imagery is hardly all there is to Western social theory. A more comprehensive statement would have to include patriotic celebrations, Holiday dinners, and at least the more staid religious services. Also necessary is a consideration of Navajo witchcraft (Kluckhohn 1944), adultery, and, in former times, warfare (Hill 1936). But Witherspoon is less concerned with reasoned comparison than with writing a morality play.

Addendum
After this essay was largely completed, I was able to secure the *irst two volumes of lessons on the Navajo language by Father Berard Haile (1941), a Catholic missionary who spent most of his adult life with the Navajo. I append some of its *indings here, partly for the reason just noted but, more importantly, because it supplies information not obtainable (so far as I can discern) from any other source. Additionally, as will be seen, it bears heavily on the analyses so far considered.
Father Berard distinguishes between the "family" kinship terminology of the Navajo and their "clan" terminology. It is not entirely clear that this distinction is achieved lexically by the Navajo, but in any case, we learn that actual siblings may be referred to by terms which distinguish them from classi*icatory 'siblings' (Haile 1941:58). This in itself does not allow a conclusion concerning focality: there is no indication, for example, that the terms can be translated as 'real' or 'true' (see above). But, as we have seen, the 'sister' kin-class is divided into 'real' and (by default) 'unreal' subclasses, so it seems a relatively safe bet that the 'brother' and other kin-classes are so divided. In short, it would seem that procreatively close kin are indeed lexically distinguished from others of their kin-classes. And if this were not enough, we learn from Father Berard that the "family" 'brother' term is translatable as 'he was produced or born with him or her' and that a pluralizing version of this lexeme translates as 'he was produced (or born) with them' (Haile 1941:57, 62). This is an unequivocally procreative idiom, and as such it provides the strongest evidence for a Schef-*lerian interpretation of Navajo kinship.
Witherspoon's contrastive analysis can be faulted on yet another front. Father Berard reports a term which he translates as 'parents' (Haile 1941:60), which suggests that one's father, far from being an af*ine, as Witherspoon contends, is conceptualized as being on a logical par with one's mother. Another term, which Father Berard tells us "is applicable to both parents," is translatable as those 'who gave us birth' and is used among actual siblings (Haile 1941:79), thus underscoring both the logical parity of one's parents and his/ her conceptualization of them as co-procreators.

Navajo Gender ClassiDication
Schef*ler's contribution to gender studies is to be found in a contribution to feminist scholarship. His words: Consider [the] claim that in some North American Indian societies "gender role" rather than genital anatomy determined … classi*ication as a male or female ….
[This] argument is intended to liberate gender from any biological basis … [and] to show that a system of two genders is by no means inevitable. [Such scholars] ac-knowledge, however, that "gender role" is de*inable only as … behavior normative for a member of one or the other genital-sex class … and that assignment to one or the other sex class is typically at birth and not dependent on any … behavior on the part of the person being classi*ied. Because, logically, categories must be de*ined … by criteria independent of the normative implications of inclusion in those categories, certain forms of conduct cannot be both criteria for and normative implications of inclusion in … the same category. It must be that [we] are dealing with situations in which some men (less often women) are permitted to act, in some degree, as though they were women (or men), or as an anomalous 'he-she' or 'she-he.' … [One] cannot … produce any linguistic data to demonstrate that [members of the alleged "third sex"] are treated in any language as a genuine third gender (Schef*ler 1991:377-78, emphases in original).
This needs unpacking. The contention that gender class can be independent of genital anatomy or any other "biological basis" dovetails with currently faddish claims, in the academy and outside it. Schef*ler's counter-claim is that gender, like kinship, has as its semantically central form, certain traits that might be dubbed "biological," or, better, "folk biological;" i.e. native appreciations of biological difference. The further counter-claim -that a stipulation of the folk-biological criteria for membership in a gender class is logically prior to the behavioral criteria pertinent to that class -is a transform, in the sphere of gender, of Schef*ler's charge against the "alternate distinctive features" argument put forward by David Schneider (see above). Finally, though this counter-claim is not developed, the positing of 'he-she' and 'she-he' gender subclasses indicates nonfocal status, these categories being logically dependent on 'he' and 'she' (i.e. 'male' and 'female') classes.
These points are pertinent to the Navajo case, and entirely contrary to the argument put forward by Wesley Thomas (1997), a native Navajo speaker. Here we are told that 10 'woman' constitutes "[t]he primary gender in Navajo culture" (Thomas 1997:158): this recalls Witherspoon's assertion (see above), and, like it, it presents no corroborating semantic evidence. But the key concern of Thomas' essay is with a Navajo gender category which, following Hill (1935), I shall gloss as 'hermaphrodite,' a label which, as we shall see, is quite accurate. Thus Thomas (1997:157) tells us that '[t]he traditional gender system" of the Navajo is "based initially on biological sex." "Initially "here, it would seem, is meant, however unwittingly, to indicate focal status from which 'hermaphrodites' "subsequently" depart; i.e., temporal imagery is here used to express differential semantic status. Thomas further notes (ibid.) that there are "sex-linked … roles." These involve the considerations that "women are generally weavers, men are generally hunters." Moreover: Women's sex-linked activities include those associated with childrearing, cooking and serving meals, making pottery and baskets, and doing … other work associated with … the domestic sphere. For men, getting wood, preparing cooking *ires, building homes, hunting , … and doing or overseeing work associated with the ceremonial After the title of Thomas's essay, identi*ication as Navajo appears. This is also the case in his identi*ication 10 in the co-authored introduction to the volume (Jacobs et al.... 1997). I *ind this odd. In the latter case, his coauthors are not identi*ied ethnically, nor is it common to do so in scholarly publications. This may seem innocuous, but as soon as I began to read the singly-authored piece I predicted I would be in store for an unwelcome lesson in identity politics and not -not primarily, anyway -a scholarly essay. I was all too correct. (Thomas 1997:158, emphasis added; see also Aberle 1961:141-42;Bailey 1950:74, 84;Kluckhohn and Leighton 1974:94-95;Lamphere 1977:73, 83, 122, 147;Shepardson 1995:165-66;Spencer 1947:24-25, 28, 99)).

aspects of everyday life are appropriate. A['hermaphrodite'] mixes various aspects of the behaviors… of both females and males
In fact, this unremarkable division of labor is by no means hard-and-fast: there is abundant evidence, for example, that women can be active in Navajo ritual life, though some regard this as deviant (see esp. Schwarz 2007). But the italicized portion of the quote is key, for it makes it plain that the 'hermaphrodite' category is conceptualized as a blend of male and female qualities. It is not, as Thomas (1997:160) contends, a "third gender" on a logical par with the focal pair. It is, in Schef*ler's parlance, a "he-she" or a "she-he." This conclusion is reinforced by Thomas' translation of 'hermaphrodite' as someone in "a constant state of change" (1997:171), implying as it does non-membership in any *ixed category. Even more telling, *inally, is Katherine Spencer's report that in Navajo mythology the 'hermaphrodite' is someone "who will know women's work and live like a woman, who will know the ways of both men and women" (Spencer 1947:98).
In another essay on Navajo gender classi*ication, one which appears in the same volume as Thomas' contribution, Carolyn Epple (1997:185) provides much the same translation, though she is far more explicit on the liminal position of 'hermaphrodites.' She interviewed several such individuals, three of whom I quote: I am different from a straight male … and from a homosexual male… I'd rather have a woman [as a sexual partner] than a queen. I mean I do have some kind of maleness.
[T]here's an in-between type of person … I don't want to be a drag queen, and I don't want to be a girl.
A queen is identi*ied with a female. But I don't consider myself a girl. I'm a man and am attracted to men (Epple 1997:181).
The *irst two informants occupy a nonfocal gender category. One claims "some kind of maleness," the other identi*ies as an "in-between type of person." The third informant unambiguously self-identi*ies as male. One, as if to underscore the point, remarks "There are two [gender classes] -male and female -no others" (Epple 1997:184). Epple (1997:188) concludes: "My … work with ['hermaphrodites'], which includes the question, Are ['hermaphrodites'] a third gender? indicates that, in the explanations of my teachers, they are not an alternate gender status." This conclusion is supported by some other *indings. Thus, Navajo seem to posit a pervasive bipartite plan of the sort brought to the attention of anthropologists by Robert Hertz (1960) and, later, Rodney Needham (e.g. 1973), and in the realm of gender classi*i 11 cation this is exempli*ied by the categories "male" and "female." Hermaphrodites as such do not *igure at all. Moreover, in kin classi*ication gender in its male and female forms is sometimes an important component, but I have been able to uncover only a single reference to how 'hermaphrodites' *igure in such reckoning: thus Katherine Spencer (1947:99) notes a myth in which "double kinship terms" -'my grandfather' comingled with 'my grandmother' The pertinent literature here is especially extensive, so I hesitate to cite it at all. The single best source is 11 Schwarz 1997. See also the tabulation in Reichard 1950:596-97. -are applied, thus indicative of derived gender status (see also Shepardson 1995:166). Finally, Navajo mythology posits a primeval separation of females from males, with the latter taking 'hermaphrodites' -described in these narratives as "neither entirely male nor entirely female" -acting as women (Schwarz 2007: 17, 20-21;Spencer 1947:25-26).
It is well worth noting that the 'hermaphrodite' category is, itself, subdivided. Epple (1997:183) suggests this when she writes that "a hermaphrodite, that from which the nonhermaphroditic ['hermaphrodites'] are generalized, is … understood in terms of femaleness and maleness and not necessarily as a third sex." In other words, the semantically central member of the 'hermaphrodite' class is an individual with mixed genitalia; such an individual provides a conceptual basis for extension to other members of this class. And this is just what W.W. Hill (1935) found over eight decades ago. More speci*ically, he noted that Navajo distinguish between 'real' "hermaphrodites" -i.e. individuals with genitals of both sexesand 'those who pretend to be "hermaphrodites"' -individuals with unambiguous genitalia (Hill 1935:273). It should immediately be noted that this is virtually identical structurally to the Choiseulese distinction between a 'true' 'father' and a 'classi*icatory' one -a dichotomy which, as we have seen, has numerous ethnographic parallels. But even more pertinent here is that, in Navajo conceptualization, semantic centrality in gender classi*ication is dependent on genital anatomy, contrary to Wesley Thomas and other marketers of intellectual faddishness, but much as Schef*ler would have predicted. 12 To the extent that we can generalize from the Navajo case, it seems clear that people categorize gender primarily on the basis of observable biological characteristics. More certain is that a comparable calculation is at work in the kinship sphere among the Navajo and virtually everywhere else: kinship has primarily to do with the procreation-derived relationships within the nuclear family. From these bases there is extension in both domains to other people, to mythical beings -'hermaphrodites' *igure in Navajo cosmology -, and, in the kinship domain, to artifacts and environmental features. Schef*ler, I am quite certain, would have reached much the same conclusion.

Concluding remarks
The foregoing analysis, I submit, is closer to Navajo understandings than Witherspoon's is, largely because, as I have noted, Witherspoon is less dependent on these understandings than he is on what I have termed "the Schneider Narrative." This Narrative has in subsequent kinship studies been connected with an embrace of "postmodern"/"deconstructionist" argument (Feinberg 2001:11).and inadequate attention to previous scholarship, which, as we learn from Janet Carsten's overview of these studies, are just old hat (Carsten 2004). Yet, trendiness aside, it should be clear from my prolegominal remarks that the Schneider Narrative is plainly and simply false: far from being entrapped in Eurocentric assumptions, Schef*ler and ethnographers before him relied on discerning analyses of the pertinent languages, attending especially to the phenomena of lexical marking (see esp. Schef*ler 1985esp. Schef*ler , 1987. I am quite certain that Schneider was familiar with at least some of these analyses, and equally certain that he lacked the analytical skills to grasp their import. This criticism may seem unwarranted in view of the enormous prestige he enjoys in current kinship studies. But to see it all one has to do is to consider his argument in American Kinship that his In an endnote to his essay, Thomas (1997:171) seems to support Hill's analysis. For an outstanding critique 12 of scholarly faddishness in "third gender" studies, see Besnier (1994). employment of the expression "distinctive features" is "taken directly from linguistics" (Schneider 1968: 32). This is preposterous, for two reasons. First, the alleged "distinctive feature" of "enduring, diffuse solidarity" has no clear empirical referent, something already noted; rather, it stems from Schneider's undisciplined intuition. Second, and related, the expression "distinctive features" as employed by linguists refers to empirically discernible sound properties, and these properties are independent of any particular language. The elucidation of these properties in linguistic theory is a result of careful analyses of particular languages already studied, which analyses have generated an etic grid to be used in the future study of particular languages: Ward Goodenough (1970Goodenough ( , 1981 was especially important in applying all this to a theory of ethnographic description. The result is a discipline based on cumulative and replicable knowledge, not the analytical sloppiness and unchecked guesses of a messianic leader. Whatever the faults of their analyses, most of Schneider's early admirers knew enough linguistics to come up with valuable studies of kinship constructs in the communities in which they worked (e.g. Inden and Nicholas 1977;Leaf 1972;Silverman 1971). For the most part it is only with the scholarship of the last four decades that kinship studies have taken a decided downturn, with primitivist, Marxist and especially so-called "radical" feminist fantasies trumping reasoned analysis. What makes Witherspoon remarkable is that, almost alone among David Schneider's *irst cohort of graduate students, his rendition of his *ield materials, lacking as it does anything resembling expertise in linguistic methods, is an unintended harbinger of the present retrograde state of the sub-discipline. Nowadays, "expertise" in kinship studies is bestowed on the basis of participation in a New Establishment, one which presents itself as "cutting-edge" on the grounds that it has, supposedly, "deconstructed" more intellectually demanding and more responsible analytic methods.
The shoe really should be on the other foot: the "deconstructionists" in kinship studies themselves warrant deconstruction. Like most Establishments, they refuse reasoned debate: when they address scholarly opponents at all, it is with ad hominem charges of ethnocentrism, "biological reductionism," or subscription to unfashionable political agendas (e.g. Franklin 2001;Sahlins 2012;Schneider 1989). Schef*ler, by contrast, was keen to respond to critics: his Baniata essay, already considered, is key here, but it is supplemented by a very considerable corpus of intellectual engagement (see references).
Instead of an empirically unsustainable relativism, the work of the man honored in this article supports the contention of the Psychic Unity of Humankind on which our discipline was founded. Early in his career, Hal Schef*ler maintained that although the nuclear family … may not be present in every human society -it is present … in better than ninety-nine per cent -recent advances in our understanding of [our] nature as a primate have established that the nuclear family has its roots in that nature. The pervasive recurrence and durability of the nuclear family … is the product of several factors: the extreme and prolonged dependency of human infants and children upon both male and female adults; the division of labor between the adult sexes which arises out of behavioral differences related to the sexual differences; and the enduring unions that tend to form between adults of the opposite sex in order to rear children, to subsist, and because of the continual sexual receptivity of the human female (Schef*ler 1966:12).
More than a half century later, with massive increases in what have been called "alternate family forms," these remarks may seem dated. But if they do, we ought to ponder why, for example, in some of the Nayar groups studied by Kathleen Gough a decade earlier, and in spite of the absence of a localized nuclear family, not only is individual paternity recognized but, as well, genitors have special and locally recognized relationship with their children (Gough 1961); why, in the best-studied case of polyandry, individual genitorship is comparably respected (Levine 1988:167-68); and why, in at least some African-American ghettos, women evidence a decided preference for men who are prepared to stay with them (Stack 1975).
Equally important, Schef*ler was much concerned with replicability, with the responsibility of a scholar to other intellectuals and to a wider public. In the selfsame essay he insisted that "a naïve imaginative apprehension of other people's models will not do as an anthropological method; our apprehension of other people's models must be established by means of some systematic, replicable method" (Schef*ler 1966:68). In today's anthropological climate this too may seem outdated, which invites us to ponder the political implications of relying instead on arbitrarily selected bits of evidence, undisciplined intuition, and -worst of all in a free society -an unwillingness to engage in genuine debate (see esp. Shapiro 2018a). In case we need reminding, given today's Executive leadership in Washington, our political system is founded upon more universalistic principles.