Printer Friendly

Unilineal descent and the house--again: the Ngadha, eastern Indonesia.

The phenomenon in its specificity should be explored before declaring the relevance or not of already accepted understanding. Bruce Kapferer 1997:11

Discussion among anthropologists about the criteria for differentiating between and thus the criteria for classifying various kinds of kin-based group composition has recently lost much of the intensity it once had. However, even if work on kinship has now moved in directions unforeseen until the 1960s, the debates of the mid-twentieth century have not lost their relevance. Thus while we now have a far more sophisticated understanding of Euro-American kinship in particular and cognatic kinship more generally, we are also in a position to reappraise some remaining conundrums beyond the Euro-American sphere, thanks to an insightful intervention by Claude Levi-Strauss (1982, 1987). I refer to the notion of 'house-based societies' (see also the contributions in Carsten and Hugh-Jones 1995).

In this article I wish to backtrack a little from important work on 'the house' (although I take it up towards the end) because it seems to me that the lessons of the earlier discussions I alluded to have not quite sunk in. I have in mind the debate between descent theorists and alliance theorists over, to put it simply, the question of which phenomenon is more fundamental, descent or alliance. More precisely, it has to do with a separate aspect, namely the nature of group recruitment, especially as regards descent.

What has occasioned the present exercise is that there are conflicting accounts of the fundamental principles of the social organization of the Ngadha in eastern Indonesia. In an earlier work (Smedal 2002), I proposed that the Ngadha are best described as cognatic. According to Susanne Schroter (2005) they are matrilineal. Schroter, consequently, employs terms such as 'clan', 'sub-clan' and 'lineage' in her analyses; while I opt for 'House' and 'House coalition network' in mine. In my view, the vocabulary of descent theory is misplaced in any analysis of Ngadha social organization; Schroter, on her part, has reservations about the Levi-Straussian 'House' terminology. (1) Since it is unlikely that both of us 'got it right', it would seem prudent to reconsider not only the evidence, but also the theories employed. A broader reason for doing so concerns the very question of comparison. If Schroter's account is accepted, then the Ngadha might most fruitfully be compared with other matrilineal social formations. If mine is, then a more reasonable comparison would be with other 'house-based societies.' Each solution has its own problems, for different reasons. To anticipate part of the argument, I seek to demonstrate that in Schroter's case, the central problem consists in the veracity and interpretation of the ethnographic data; while in mine it consists in how Levi-Strauss defined 'house-based societies.'


I submit two disclaimers. Firstly, it is beyond my purpose here to review the debates, puzzles, critical reviews, and counterattacks that the concept of descent has generated--attributable to misapprehensions, misreadings, and entrenched theoretical commitments, not to mention biases--and sometimes, of course, to ethnographic errors, inadequate information, and to the plainly incompatible opinions of the protagonists.

Secondly, I shall not discuss the various modes of descent: patrilineal, matrilineal, bilineal, multilineal, alternating, parallel and cognatic. To clarify what the terms might mean and to illustrate how each mode may regulate the 'transmission of rights' to various kin statuses would require an altogether different paper, especially if one should also consider what happens when several modes are assumed to operate simultaneously. Rodney Needham once calculated, with characteristic enthusiasm, that a combination of six of the seven modes just listed would yield a typology of 63 descent systems, and that if the calculations are based on permutations--instead of combinations the total, depending on the assumptions and methods adopted, could amount to 325, 1,956, 3,125, or 46,656 types. He comments laconically that whether this is worth doing would depend on the results (Needham 1974:48, note 1).

The problem revisited

Since to my knowledge the results Needham hinted at have yet to be produced, we must proceed without them. My point of departure is circa 1950. Following Edward Evan Evans-Pritchard's The Nuer (1940) and Meyer Fortes's two books on the Tallensi (1945, 1949), the fundamental challenge in the early 1950s was to understand how so-called stateless societies work--and by this I mean quite specifically how rights and duties are distributed among statuses, and how political alignments are generated. The question was how societies are organized in the absence of kings and chiefs, centralized government, administrative machinery and judicial institutions. But this challenge also pertained directly to persons; to their fundamental identities and allegiances as social beings. In the smallest societies, the problem virtually solves itself. Here, most members are related to each other, and the political and kinship structures largely overlap. In larger societies of hundreds of thousands, even millions of people, the problem of political affiliation--as outlined by Evans-Pritchard and Fortes--is solved by unilineal descent groups of varying range and magnitude. First-year anthropology students are taught to identify 'segmentary lineage systems' in which small groups called 'lineages' may be connected to each other in what Fortes called 'fields of clanship' (Kuper 1988:202)--or if viewed from 'above', society can be segmented into clans and clans into lineages. The grand appellation for the whole scheme, when predicated exclusively on agnatic links, is 'a patrilineal poly-segmentary stateless society' (J.A. Barnes 1962:9).

Under given conditions, smaller lineages would unite with lineages they considered themselves one with, because their members had the same ancestor --and when conditions changed they would segment. Integration would follow differentiation. A built-in feature is therefore what one might call structural flexibility. As Fortes (1969:75) put it, late in his career, lineages are seen to be implicated in 'structurally counterposed processes of "fission and fusion"'. The social structure itself, the morphology, is 'exhibited as present in process, and process is shown to shape morphology' (Fortes 1969:75).

The units in these systems, then, are unilineal descent groups. As a first approximation we may say that a 'unilineal group' is a group whose membership is always ascribed by exclusive reference to one parent, never to the other. In many cases, these groups are also corporate; they control an estate--land, for example--and are able to act as 'one person' with respect to this estate in their external affairs. The underlying assumption is that only through groups defined unilineally can members be precisely and incontrovertibly placed in unambiguous groups, without rival claims and split allegiances ripping groups apart. For all that descent theory has been made to stand for, it is this assumption that has attracted the most critical attention from its adversaries.

Having introduced the word 'corporate', I should also say that this is the point where Evans-Pritchard and Fortes part company. The Nuer patrilineages that Evans-Pritchard wrote about have little or no corporate identity; they are, as it were, 'pure' lineages with no commonly held economically relevant landed estate. But while estates are often tangibly and materially economic they may well be primarily symbolic. Evans-Pritchard always insisted that among the Nuer, ceremonial relations are the very essence of lineage affiliation (Scheffler 1986:348, note 6). The Tallensi patrilineages that Fortes wrote about, on the other hand, are corporate groups, and although the difference has important implications, notably with regard to residence and the role of cults, I shall not dwell on them here.

As many later commentators have pointed out, the elegance of the African-derived segmentary lineage model was seductive. It did not take long before freshly trained post-World War II graduates began doing fieldwork. The model was made to bear on new findings, and unilineal descent groups were soon found elsewhere. The best known of these 'new discoveries' is probably the great number of 'patrilineal descent groups' recorded in the New Guinea Highlands in the 1950s and 1960s. In the Highlands, it was reported that there were neither powerful authorities nor centralized institutions; patrilineal descent was emphasized, and there was evidence of group segmentation. To make many long stories short, analysts applied the African model with much ardour and little modification (Kuper 1988:205). But it soon became clear that despite resemblances, discrepancies were abundant, in fact so numerous that the question arose whether the model--which was a rather idealized version of the African prototype--was applicable after all.

John A. Barnes (1962) pointed out three areas in particular in which New Guinea Highlands societies seemed to diverge from the African type. Firstly, if descent groups use descent as the only criterion of membership, then 'it is hard to discover descent groups' in the Highlands where groups usually have shallow, weak genealogies and contain many non-agnates (J.A. Barnes 1962:6). Secondly, unlike the case according to the African paradigm, men in the Highlands societies have a considerable range of optation at their disposal; it is possible to have multiple group membership, and there is a strong emphasis on individual choice and initiative (J.A. Barnes 1962:6). Thirdly, group segmentation does not follow the 'chronic' structurally determined African pattern but seems to be 'catastrophic' and 'arbitrary' (J.A. Barnes 1962:9). The activities of individuals, and the assorted factors that condition individual choice, seem more important than genealogical principles per se (A.H. Carrier and J.G. Carrier 1991:16).

Furthermore, many of the non-agnates are quite powerful, the difference in status between agnates and non-agnates is often unclear, and an agnatic ancestor cult either does not exist or does not provide a context for bringing together agnates who reside elsewhere. Now since all of this is contrary to what was considered typical of the patrilineages in Africa, John A. Barnes (1962:6) concludes--carefully--that 'it seems prudent to think twice before cataloguing the New Guinea Highlands as characterized by patrilineal descent'.

Anthropological literature grew rapidly, and the documentation of the variety in social forms in the Highlands improved correspondingly. We now know that Melanesian groups are constituted in a wide variety of ways. Principles of recruitment can focus on common substance, territory, marital alliance, ritual participation, and especially on the crucial distinction between sharing and exchanging. Even if the groups happen to be relatively 'stable' qua groups, they are composed by people who affiliate and disassociate themselves in relative fluidity. This is not to say that most people are not members of the groups of their fathers, but--and this is what I intend to make clear now--to be a member of one's father's group and to be a member of a patrilineage are two very different things.

Important distinctions

Much of the confusion over the difference I just noted stems from what is meant by 'rules of recruitment'. But I must first backtrack a little. An essential insight informing William Halse Rivers Rivers's classical work on descent is that one must never mistake principles of inheritance and succession for principles of descent. Although these principles have to do with the acquisition of rights, only the last of these concerns 'group recruitment'. A fundamental difference as far as inheritance and succession are concerned is that rights must be relinquished before the recipient 'takes over'. These rights must either be given away by someone who holds them, or be transferred at death. On the other hand, when rights descend, 'the recipient is added [...] to a collectivity of bearers of the same rights' (Scheffler 1986:340). This is a crucial sociological point because by acquiring a right by descent, the person joins a collectivity or group of persons, all of whom have this right. They are, in this sense, equals.

Virtually all authors after Rivers seem to agree that the distinction he drew is fundamental. But even when we leave the issues of inheritance and succession to one side, and concentrate exclusively on descent--which we can now say constitutes a rule of recruitment--other distinctions are equally relevant.

The first is Fortes's classic distinction between filiation and descent. Filiation, Fortes (1959:207) says, 'is the relation that exists between a person and his parents only', whereas 'descent refers to a relation mediated by a parent between himself and an ancestor', and Fortes defines an ancestor as 'any genealogical predecessor of the grandparental or earlier generation'. In other words, 'filiation' pertains to one direct genealogical link between persons of two generations; 'descent' pertains at the very least to two genealogical links between persons of three generations.

A second distinction, central to the argument that follows, pertains to the word 'rule' itself. It is most commonly taken in a regulative sense. This means that under certain conditions rules are 'followed', while at other times they can be 'bent', 'broken' or 'violated'. As long as 'rules' of 'descent' are taken in this sense, the discussion cannot be very fruitful. For what is at issue here is the notion of 'rule' in its constitutive sense--in the very sense that it defines who is included in a group, and thereby who by definition is excluded from it.

The third sort of distinction to be drawn requires more elaboration and brings me directly to the argument. I should acknowledge that my understanding of the issues at stake here is much indebted to two articles by Harold W. Scheffler (1985, 1986) where, among other things, he criticizes his own previous work (for an extensive discussion, see Scheffler 2001).

The 'mature' Scheffler's principal point, drawing on Fortes, is that in order to know whether a group is a unilineal descent group, and not merely something that happens to resemble such a group, we must be absolutely clear about whether patrifiliation (or matrifiliation) is the necessary and sufficient condition for inclusion. In the event that it is, we can speak of a principle of unilineal descent, and only then would contributions from 'lineage theorists' be relevant to our understanding of the dynamics of the internal and external relations of the group we are interested in. An invented example will illustrate this point.

Let us assume a group of people who call themselves B. We observe that the group consists of members of either sex. We are told that by virtue of being B, people have certain rights and duties--some, perhaps, pertaining to men, others to women, and many to everyone. We are also told that people who are not B do not have these rights and duties. We then want to find out how one becomes 'a B'; we want to understand the principle whereby said rights and duties are acquired. In other words, we want to establish the rules of recruitment of this group. That is easy, we are told: all Bs have a B father. What the mother is, is irrelevant. We note that what is involved here is what Fortes calls patrifiliation. Next, we conclude that the Bs must constitute a 'patrilineal descent group.' For if all Bs are B, and have all the rights that all Bs have, simply by virtue of having B fathers, surely what is involved is a principle of patrilineal 'descent'? To examine whether this conclusion is warranted, we must consider the difference between three possibilities.

The first is that the principle of patrifiliation is the necessary and sufficient condition for membership. The second is that patrifiliation is necessary but not sufficient. The third is that patrifiliation is sufficient but not necessary.

I begin by considering the last possibility. What this means is that the group B is, in principle, open to recruitment from the outside. For, although each person 'is' a B because he/she is born into the group of Bs--of which his/ her father is a member--the father himself need not have been a B by birth. The rule being what it is, each person has a choice of whether or not to retain his/her natal membership or, if the available groups are mutually exclusive, to join another one. If the groups are not mutually exclusive, a person may even acquire more than one such membership and so, in turn, may his children. It is self-evident, therefore, that groups of this kind are not exclusively composed according to a principle of descent; they are not patrilineal groups. In fact, apart from fathers and their children, their members need not be related in any way at all (Scheffler 1986:343). In other words, although people are entirely truthful when they say that someone is a B because his or her father is a B, the father may well have been born an A. Perhaps it can be verified statistically that most Bs have fathers who are also B by birth. But this is entirely irrelevant to our question: are 'the Bs' a patrilineal descent group? In this case, the rights that our Bs have do not descend.

Clearly, the same conclusion must be drawn with regard to the often-reported quasi-patrilineal groups in the New Guinea Highlands mentioned earlier. These groups may well exercise the principle that patrifiliation is sufficient for inclusion, but that does not mean that they constitute patrilineal descent groups. 'Although it is a matter of principle that each and every person [in the Highlands] belongs to his or her father's group, it is not a matter of principle that each and every person remains identified for life with that group.' (Scheffler 1985:3.)

What about the second possibility: that patrifiliation is a necessary, but not sufficient condition for inclusion? This, in our hypothetical case, would mean that all Bs are agnatically related--that is to say, related genealogically 'upwards' (in a diagram) or 'backwards' (in time) through paternal links but not all persons agnatically related to Bs need to be members of the group with that name. Thus, in order to qualify for the rights that membership entails, certain additional stipulations must exist. Those with the necessary patrifilial credentials are doubtless potential members, but they must attain further credentials in order to be granted group membership. Those without such credentials simply do not become members. The nature of these credentials is not obvious. Scheffler (1986:342) notes, however, that if the credentials depend on a person's choice, then the person may choose whether or not to become or to remain a member of a particular group, but cannot choose between groups. Alternatively, one could assume that choice here rests not with the individual member but with the group as a whole, or with persons acting on its behalf who hold powers of excommunication. Whatever the case may be, it is clear that we do not, in this case either, have before us a patrilineal descent group, because the rights in question do not descend automatically and indiscriminately--the decisive point being that these rights are allocated. Someone born as the child of a male member of this group satisfies only the first of a potential series of conditions for membership; birth is a prerequisite for membership, but it is not enough. (2)

I turn now to the last alternative: when patrifiliation is the necessary and sufficient condition for membership. This is the African model in its pure form and the criterion of membership can be stated very briefly. The principle defines a group as one that includes all and only children of all male members, and all and only descendants of males of all male members. In other words, the group is defined in such a way that 'leakage' is impossible. No males can be admitted, and none can withdraw or be expelled. The only choice a person has is between being an active or an inactive member. Membership, in short, is for life, and there is no escape hatch.

To conclude, the factually correct statement that 'all Bs have a B father' certainly tells us something about how the group is composed. We cannot from this knowledge alone, however, determine that it is composed according to the principle of descent.

Some readers might think, impatiently, that such talk about necessary and sufficient conditions is formalism at its worst. It enlightens very little, and it contradicts the empirical evidence of the classical African unilineal descent systems which have been analysed hundreds of times. It has long been accepted in anthropological circles that African descent group membership is far more fluid than the African model allows. One might want to assert next that there is in fact nothing substantially different between the (very) loosely agnatic groups of the New Guinea Highlands and the African lineages. Should we then complete the U-turn, and take the Highlands models to Africa? (3) Well, if we take a few points into account, it would seem unwise to do so. It is true that African patrilineages often consist of a high number of non-agnates but this single fact can be attributed to two separate issues. The first pertains to residence, and the other to external recruitment. It appears that on both counts, Africa and the Highlands differ profoundly.

Concerning residence, Evans-Pritchard made it quite plain that the Nuer 'communities and lineages are distinctly different kinds of group and have distinctly different kinds of rights and duties' (Scheffler 1986:345). They are different, because Nuer lineages are dispersed groups; lineage membership, besides its political implications, is first of all about ritual and ceremonial matters, and about exogamy. Residential groups, on the other hand, are in principle open to virtually anyone. Any Nuer 'is free to live anywhere--without regard to, and without effect on his lineage affiliation' (Scheffler 1986:346). In contrast, the so-called 'quasi-descent' groups of the New Guinea Highlands are at the same time residential groups; one author, significantly avoiding descent idioms completely, calls them 'parish groups' (de Lepervanche 1967:157; Scheffler 1985:11).

The second point, concerning extra-lineage recruitment, is vital, because the 'genuinely' patrilineal descent groups as just outlined would seem closed to immigrants. How, then, does assimilation take place? This is precisely where the full force of the dogmatic descent principle, as a constitutive (not regulative) rule, becomes apparent. In the Highlands, as we saw, adult men easily become members of other groups. Moreover, the fact that non-agnates are hardly concerned about their being immigrants indicates just how weak the agnatic principle is. In the paradigmatic cases of unilineal descent groups in Africa, however, immigration occasions much legal fiction making, and lineage membership requires generations before it is acknowledged. 'The assimilated parties must have patrilineal status imputed to them.' (Scheffler 1985:14.) Otherwise, they cannot claim the rights that accompany membership. But in the Highlands there is simply no need 'to indulge in the legal fiction that certain members are agnatic descendants of the founding member when it is known that they are not. Not only is there no need to do it, but it makes no sense.' (Scheffler 1985:14.) This is only what we should expect, given what we already know about the Highlands. For if there is no descent dogma, no constitutive rule of descent, there is nothing to appeal to.

The Ngadha

The foregoing account may come across as a blast from the past. But, trusting that a dense exposition without nuance has brought across the basic idea of what 'descent' entails--at least its unilineal mode--I now turn to Schroter's account (2005) of the Ngadha. (4) My remarks are addressed primarily at what is approximately the first half of her article (the sections entitled 'Social structure', 'Gendered symbols', and 'Authority'). It is in these sections that she seeks to establish the ethnographic basis for her subsequent argument (which I shall not discuss in detail; it deals with the gendered roles of women and men especially in terms of 'power' and 'knowledge', and the ways in which these roles may now be changing). Of course, the soundness of her argument necessarily hinges on--and must be judged by--the accuracy and reliability of the ethnographic evidence as she presents it. Since that argument in turn largely draws on comparative studies of 'matrilineal' societies, it is obvious that the strength of Schroter's conclusions depends on whether or not the Ngadha--according to accepted criteria--can be said to be matrilineal.

Before continuing, I should make it clear that while Schroter's fieldwork sites and my own are located some 10 to 15 kilometres apart, both are found in what I shall refer to--for want of a better designation--as 'greater Ngadhaland' where the vast majority routinely practise uxorilocal postmarital residence. (5) In the following, occasional mention will be made of a contrasting pattern of residence in what I refer to as 'the Were region' in Ngadhaland's eastern highlands where I conducted intermittent investigations in the 1990s in order to ascertain precisely the rationale behind a residence pattern so at odds with the one practised everywhere else.

Social organization

Schroter (2005:321) states:
   Central to their system of social organization, which unites the
   principles of alliance and descent, are unilineal kinship groups
   (woe) and the house (sa'o).

      Descent groups are divided into three levels: clan, sub-clan, and
   lineage. A clan is a named entity that descends from a mythical
   pair of ancestors. It has no political or social function and
   serves merely as a symbolic point of reference. Members of a clan
   do not necessarily know each other nor do they interact. In
   contrast to the clan, the sub-clan is an important social unit
   whose members are obliged to provide mutual support. They share
   ceremonial obligations, particularly in case of death, during the
   annual cycle, and in ancestor worship. Each sub-clan consists of
   three lineages related hierarchically, each of which possesses
   particular ritual titles. Each of these lineages is comprised of a
   female and a male half, thus forming six different lineages. The
   highest-ranking title is called saka and is divided into saka pu'u,
   the female component, and saka lobo, the male component. Pu'u means
   trunk or source, and refers to the idea that women are the source
   of everything and that human life originates from the female body.
   Lobo is the tip, the offspring, the younger one. The lower titles
   are kaka and dai, split into a binary pair: kaka pu'u and kaka lobo
   and dai pu'u and dai lobo. Kaka and dai are classified as followers
   or younger siblings of saka.

Without providing any ethnographic particulars in support of her claim, Schroter simply asserts that Ngadha kinship groups (woe and sa'o) are 'unilineal'--a possibility that prior analysis (Smedal 2002) has considered and rejected. In the quoted extract, Schroter states that each 'sub-clan' consists of three 'lineages' and that they--since they are 'comprised of a female and a male half'--form 'six different lineages'. In my view, this is inaccurate. There is no predefined limit to the number of (what I prefer to label) Houses (what Schroter calls 'lineages') that a woe may contain. While the maximum number of levels of named Houses is usually three it is sometimes the case that the maximum has not been reached, or that it was reached in the past, but following demographic decline the lesser cadet Houses no longer exist. Furthermore, it is also the case that--depending on a combination of favourable demographics, fortunate economic developments, and even immigration--lower levels may be represented by more than one House. To be specific (unlike Schroter), in Gurusina, the largest traditional village (nua) in the subdistrict (kecamatan) of Jerebuu, there were four resident woe with a total of 32 named Houses in 2007; one woe, for example, consisted of 10 Houses, another of 8.

Schroter's statement that woe and sa'o unite 'the principles of alliance and descent' is largely uncontroversial, although the woe as such is never directly involved in specific marriages; that is the prerogative of the sa'o. (6) But when she adds that the Ngadha 'clearly differentiate between alliance and descent groups' (Schroter 2005:321, note 6), one wonders if she means a. that the Ngadha differentiate between two principles (alliance and descent) or b. that they differentiate between alliance groups and descent groups. The first possibility is simply as self-evident to the Ngadha as it would be to everyone else. The second possibility makes little sense, because it is precisely the sa'o, as 'kin group' (I largely avoid, for reasons already given, the concept 'descent group') that is activated whenever a marriage is contemplated.

Schroter's account with regard to the nature of Ngadha group formation is inherently confusing. She writes of 'three levels: clan, sub-clan, and lineage' but provides indigenous terms only for what she calls 'sub-clan' (woe) and 'lineage' (sa'o). She states that 'Arndt (1954:189) uses cili bhou for clan, a term that is largely unknown and no longer in use today' (Schroter 2005:321, note 7). But Arndt does nothing of the kind. On the page Schroter directs us to, Arndt employs the term cili-bhou for what he calls a sub-clan (Klanteil; literally 'clan part'). (7) What is really puzzling, however, is Schroter's insistence that in Ngadhaland a delimitable social unit ('clan') exists, which encompasses presumably numerous woe (her 'sub-clan'). But she offers no clues as to what a 'clan' might be and might do; she gives no Ngadha term for it; and she does not inform us whether it carries any symbolic representation. One wonders, therefore, how it is manifested. (8)

This puzzle only grows when Arndt is consulted again. The first sentence under his subheading '4. The clan' (Arndt 1954:204) begins as follows: 'The native term for clan is voce [...]'. (9) Plainly, Schroter's account of the fundamental units in Ngadha social structure conflicts not only with what Arndt had to say about them and with what I have taken them to be, but also with what has been reported by others who have recently published on the topic. (10)

In my view, the Ngadha have two main levels of social organization--woe and sa'o--not three.

Marriage, bridewealth, offspring, and the question of descent

Schroter (2005:322) states that 'Membership in a clan, a lineage, or a house is always determined through the uterine (maternal) line. Genealogical continuity is transmitted only through women.' She is positive that the Ngadha are matrilineal. My own view is different. Based on ethnographic evidence, as opposed to Schroter's unsupported assertion, I have shown that while it appears to be the case that 'membership' is determined through 'the maternal line', this is only, and only nearly, true statistically, not as a matter of principle. The reason for it being nearly true is found in the logic of the bridewealth economy, as explained elsewhere. (11) Moreover, Schroter does not consider marriage and house affiliation practices in the Were region. Here, again as a consequence of the logic of the bridewealth economy, the continuity of the House is routinely embodied by men, not by women. (12)

In a footnote to the passage just quoted, Schroter (2005:322, note 11) writes:

Despite this, the Ngada have seldom been considered a matrifocal people and there is remarkable confusion in the anthropological literature regarding Ngada kinship, their system sometimes being defined as patrilineal (Arndt 1954), sometimes as matrilineal (Bader 1951 [sic: 1953]:135) or cognatic (Barnes 1972:85).

I note that Schroter refers to Arndt and Bader as anthropologists but whereas both undoubtedly received ethnographic training at the Catholic Societas Verbi Divini Institute before they were posted to Flores, they were first of all missionaries/priests. Of the three authors, only Robert H. Barnes qualifies as an anthropologist and his conclusion, that Ngadha kinship qualifies as cognatic should, in my view, be accorded somewhat more weight than the earlier and patently less theoretically informed assessments by Arndt and Bader. (13) Note, incidentally, that Barnes's assessment is based on his reading of Arndt's 1954 work, a reading that conflicts sharply with Schroter's. Note, too, that Arndt (1954:167), if anything, says the opposite of what Schroter (providing no page reference) imputes to him:
   The most frequent expressions for blood relationship are tuka mogo,
   tuka yi, cura mogo [...]. With tuka mogo, tuka yi, cura mogo is
   meant the entirety of all those who can trace their descent from a
   common ancestor or ancestress or from both; because kinship is
   reckoned on the paternal as well as on the maternal side, although
   the expressions tuka mogo, tuka yi suggests a priority for maternal

Here Arndt speaks of kinship and descent as reckoned bilaterally, even if linguistic locutions (tuka can mean belly, stomach or uterus; mogo and yi mean together, common, while cura means sinews) would appear to suggest a preference for the maternal side. More importantly, Robert H. Barnes (1980) later published a more thorough account of Ngadha marriage and bridewealth practices--arriving at the same conclusion as in 1972--a work that Schroter does not consider.

Schroter (2005:322) continues:
   Children, whether born in or out of wedlock, are regarded as
   members of the house, lineage, and clan of their mother, not of
   their father. As members of their mother's kin groups, they have
   specific rights and obligations which are distinct from rights and
   obligations connected with their father's kin groups. In cases of
   divorce or extramarital birth, fathers have no means of gaining
   custody of their children, as this would separate the children from
   their mothers and matrilineal relatives. A corollary of this
   matrilineal rule of group membership is that children born outside
   wedlock are accorded rights equal to those born in wedlock and are
   in no way subject to discrimination. Moreover, they are full
   members of their mother's various kin groups, and it is tacitly
   known that many siblings have different genitors.

The first thing to note here is Schroter's reference to what happens in cases of divorce. Now, given that very nearly 100 per cent of all Ngadha claim 'to be Catholic' (Schroter 2005:320), one cannot but question the frequency of divorce, if it takes place at all. Of course, it is possible that Schroter has in mind cases when a marital union has been officially sanctioned in adat terms but has not (not yet, that is) been consecrated, as a sacrament, in church. But in recent decades, the time window between the two ceremonies is usually extremely small. Another possibility is that Schroter refers to what once was the case, in Arndt's time, for example. (14) A third is that she speaks of abandonment. At any rate, it would have been helpful if Schroter had been more precise about this point.

But there is more in the passage just quoted that invites comment. That it is as a 'corollary of [the] matrilineal rule of group membership' that 'children born outside wedlock are accorded rights equal to those born in wedlock' contradicts the fact that children born out of wedlock in the Were region--where House (sa'o) affiliation routinely follows male links--would also be accorded rights in their mother's sa'o. This is simply because in such a case the genitor (and his kin and affines) will not have negotiated bridewealth and the genetrix will still be attached to her father's sa'o. As such, no pasa ('to remove', 'to transfer', 'to bring home' [to the husband's House], Arndt 1961:403) payment has been made. The essential point is that rights--hak--in a sa'o (such as rights to cultivate sa'o land, to represent the sa'o in rituals, and to 'speak', that is, to voice one's opinion on sa'o matters) and obligations inhere in any of its affiliates and that these same rights devolve to their own children. Whether these rights devolve to the children of a daughter qua affiliates to her natal House or qua affiliates to the natal House of her husband (if indeed these are two separate Houses, which they may or may not be) depends entirely on the amount of bridewealth that the husband's group has agreed to pay to her natal House.

Thus, the fact that a child of an unwed woman remains with her and that the father is barred from obtaining 'custody' is neither here nor there with respect to whether or not the social organization can be described as 'unilineal'. The third point in the quoted extract warranting comment is that 'it is tacitly known that many siblings have different genitors'. Perhaps sometimes it is, but in my experience such cases are generally common knowledge. Schroter (2005:335; see also Smedal 1996:50) herself draws attention to the well-known waja institution of compensatory payment which ensures that any premarital sexual relation is announced in public. But she omits mentioning the equally important kati institution which by way of a similar payment produces the same effect: public knowledge of adulterous relations. In both cases, a fine (usually a good horse or its equivalent in cash--twice the amount if the union produces progeny) must be presented to the House of the woman. Such awareness of illicit sexual unions is important because without it, the likelihood of inadvertent incestuous relationships would increase. It is only in the improbable event that the male and the female in question somehow manage to keep the nature of their relationship secret indefinitely that the knowledge that siblings have different genitors would be 'tacit'.

Next, Schroter (2005:323) discusses the consequences of the absence of female offspring:
   A house or lineage cannot survive without female offspring. In
   cases where there are no female children, genealogical continuity
   is formally broken and the whole system is threatened. To avert
   such genealogical crises that would result in disappearance [sic]
   as a defined group, the Ngada have only one option: they must make
   an exception to the marriage rules, thus allowing one of their male
   members to marry patrilineally and reside virilocally with his
   wife. This kind of marriage is called pasa. It is an exceptional
   arrangement that entails an extensive exchange of valuables. In
   'ordinary' marriages, which follow the principles of matrilineality
   and uxorilocality, the bride's house receives a nominal
   bridewealth: a horse, a pig, or a little money. Pasa is different.
   The bride's lineage demands large amounts of cash and several
   animals as compensation, thus turning this kind of marriage into a
   ruinous enterprise. Consequently, it happens rather seldom and is
   regarded as a measure of last resort. These precarious
   circumstances and their exceptional nature underscore and
   strengthen the matrilineal order.

Disregarding the infelicitous phrase 'to marry patrilineally' it is true that outside the Were region, a lack of female offspring gives House members cause for concern. But as Schroter surely knows there is another option, besides a pasa marriage, namely the institution of 'temporary residence' (pagu raru dhutu nu: 'light the fire and keep the smoke going'). Installing a 'live-in' couple is a stopgap solution, conveniently evading the need for large-scale bridewealth transfers that a pasa marriage would entail and securing an inhabited House while waiting for a candidate for House stewardship to mature (or to be born). It is important to realize that the eventual (future) female House steward is not sought from among the offspring of the couple in pagu raru dhutu nu residence (whose genealogical credentials are likely to be tenuous); all that is necessary is that she is genealogically linked to a living or dead House affiliate, male or female. Incidentally, pagu raru dhutu nu can sometimes generate its own problems in that the offspring--male as well as female--of the couple in residence may 'forget' (an ironic euphemism) the very reason for their parents' temporary installation and hence contest any attempts at eviction when their presence is no longer required.

It is correct, as Schroter suggests, that in 'ordinary' (outside the Were region, that is) marriages (known as dheke, 'to ascend, to climb' (Smedal 2002:502); the locution signifies that the groom enters his wife's House--it is the conceptual opposite to a pasa marriage) the economic value of bridewealth is modest, though to call it 'nominal'--as in 'negligible'--is to diminish it unnecessarily. But she is in my view mistaken in stating that pigs are among Ngadha bridewealth items. As elsewhere in Flores and much of eastern Indonesia generally, pigs are instead offered by the bride's group as 'counter-prestations' to reciprocate the bridewealth items presented by the groom's group15--but among the Ngadha, as mentioned already, these pigs are immediately slaughtered and most of the pork is consumed on that very occasion. It is more striking, however, that Schroter omits mentioning the water buffalo, the bridewealth item par excellence. (16) As I see it, the significance of the buffalo in Ngadha bridewealth practices is crucial and cannot be ignored.

To say, as Schroter does, that the institution of pasa underscores and strengthens 'the matrilineal order' is but another uncorroborated assertion. As should be clear by now, in my view it would be impossible to demonstrate the existence of any such order because there is no evidence for it. On the contrary, what pasa strengthens is the more fundamental principle that a child's sa'o affiliation depends on whether or not its parents paid a large bridewealth (as is common in the Were region) or a small one (as is common elsewhere); Arndt (1954:42-6) is exceedingly clear about this. The gist of the matter is, as Robert H. Barnes (1980:112) put it, 'Arndt speaks of bridewealth as the basis of the whole of Ngada marriage and family law', a point echoed by Molnar (2004:124) in her study of the neighbouring Hoga Sara: 'The payment of bridewealth thus has a strong bearing on house membership'.

Rank, hypergamy, residence and knowledge

Schroter (2005:324-5) goes on to outline the problems in Ngadha social organization represented by marriage across divisions of rank:
   Ideally members of these groups marry endogamously; a man marrying
   up seldom comes in for criticism, but there are sharp reprisals for
   a woman marrying down. This is because, according to the Ngada, it
   is women that are charged with keeping the blood pure, so that a
   transgression of group boundaries on their part is regarded as la'a
   sala, heading down the wrong path. In the past such a transgression
   was punishable by death, and even today women suffer public
   humiliation and degradation.

The 'groups' (or rather, categories) referred to here are nobles, commoners, and (descendants of former) slaves. But Schroter is not entirely clear about the practice--hypergamic marriage--which she tries to explain. The fundamental point of the ritual whereby a noble woman is demoted after having had sex with (or indeed having married) a man from a lower stratum of Ngadha society is that not only is the woman (nowadays symbolically) killed (see also Smedal 2009b:218), but also the man: he, too, has transgressed and is regarded as having la'a sala. (17) To state that 'a man marrying up seldom comes in for criticism' is therefore to deny the evidence also reported by Arndt (1954:334-5) from various domains in Ngadhaland. (18) For the sake of clarity it is worth adding that it is also not the case that a man marrying down seldom comes in for criticism; it is integral to the rank system as described by Arndt (and myself) that noble men are in principle free to choose a spouse from any stratum, just as, conversely, women with slave status are. Schroter (2005:244) is justified in lamenting the restrictions noble women must live with as regards the paucity of eligible men, but fails to make the point that marriage restrictions pertaining to men of slave descent are equally severe--they, too, can only legitimately marry someone of their own rank (see Smedal 2009b:218-20, especially Table 1).

There is another issue relevant to any discussion of the social practices predicated on rank distinction that I wish to take up briefly before moving on. As was noted by Arndt (1954:334-5) and which is even truer today, nobles, relative to commoners, are decreasing in number. This demographic trend is of course not lost on the nobles who, mindful of realities mentioned earlier (bridewealth expenses and the possibility that a House without daughters is a House under threat) sometimes contract marriage with the explicit purpose of populating a (noble) House. Interestingly, the general concern with bridewealth prestations, particularly if a woman is to reside virilocally, is of little import in such cases. Recognizing that the problems epitomized by 'old maidens' in otherwise empty noble Houses are widespread, noble couples do on occasion settle virilocally but do not thereby consider it a pasa marriage of the Were type. Whenever a woman is recruited in these marriages it is simply known as dhepo haki (to follow the husband) and no or only a token bridewealth is requested by the bride's side. In such cases, the very survival of the House as a noble House takes precedence over bridewealth demands.

Anyone with even the most cursory acquaintance with social organization in eastern Indonesia will be familiar with the expressions 'wife-giver' and 'wife-taker' (or 'wife-receiver'). In convenient if sometimes criticized anthropological shorthand (Allerton 2004:343), these expressions cover a range of indigenous terms for the two categories of affine that are recognized when marital alliance is of the asymmetric type, so typical in the region, and when alliance is of the far less commonly encountered symmetric type as well (Molnar 2000:132, 136). What has long been clear is that the Ngadha differ from virtually every other eastern Indonesian social formation in that their marriage practices are not predicated on the categorical distinction between wife-givers and wife-takers. (19) Moreover, where the wife-giver/wife-taker distinction is operative it follows that the units so labelled are exogamous and that they are predicated on unilineal descent--in eastern Indonesia, in the vast majority of cases, they are predicated on patrilineal descent.

Thus Schroter is vaguely correct in stating that 'among the contemporary Ngada one can hardly speak of the giving or receiving of wives since it is the men who move'. But when she continues, 'Cecilia Ng (1993:136), analysing the Minangkabau, describes the marriage system as an exchange of men, and the same can be said of the Ngada' (Schroter 2005:328), her account in my view veers off in an unfortunate direction. Here, Schroter confuses the logic of one system (of the Minangkabau) where

--cross-cousin marriage is viewed favourably

--parallel-cousin marriage is considered incestuous, and

--the matrilineage is exogamous (Fischer 1964:105)

with the entirely different logic of another system (of the Ngadha). All over Ngadhaland

--any opposite-sex cousin is a potential spouse fai weta, haki nara 'the female cousin is a wife, the male cousin is a husband', see Smedal 2002:500) and

--stipulations of exogamy between units such as sa'o and woe simply do not apply (see Arndt 1954:19; Smedal 2002:499).

As I see it, therefore, Schroter's comparative exercise serves no purpose and cannot be taken seriously. Her subsequent appeals to works by Evelyn Blackwood (1995, 1999), Michael G. Peletz (1995a, 1995b), Joke van Reenen (1996) and Alice Schlegel (1972) are hence equally misplaced.

While Schroter is vaguely correct in the first part of the statement just quoted, she again leaves out essential information. She consistently disregards marriage practices in the Were region where, to repeat, women routinely move. But she fails to mention that elsewhere in Ngadhaland many men do not, or hardly at all, move--simply because a significant proportion of matrimonial unions takes place between women and men who are affiliates of the same House or of the same woe and the distance newlywed men actually move can be measured in metres. In the two villages from which I have reliable information on this topic, woe endogamous unions accounted for more than 40 per cent in each village; the number of House endogamous unions was about 10 and 15 per cent (n = 126), respectively (see Table 1). If these data were analysed with respect to village affiliation, not woe affiliation, village endogamous unions would certainly account for an even greater proportion of the total.

This is crucial with respect to the argument Schroter subsequently seeks to advance. She asserts that men--since upon marriage they move elsewhere--cannot know what women know. She holds that men's much-touted 'expert knowledge' or '"high" knowledge' is exaggerated, and that women, because they remain in their natal village all their lives, have much deeper knowledge than do men about village affairs, genealogies, sexual scandals, and so on. Women, according to Schroter (2005:330-1), 'have no choice but to be experts' and 'neither a brother's nor a husband's expertise can compare with that of a woman'. I would not dream of suggesting that women are in any respect ignorant of village affairs, and it goes without saying that brothers residing far away cannot know what locals know. But it is unlikely that all of the women's married nara (women's brothers and, it must be added, male 'cousins') live elsewhere. It is equally unlikely that all these nara--many of whom routinely conduct sa'o and woe rituals of every kind and direct marriage and land rights negotiations whenever required--are less knowledgeable than their sisters. That in Schroter's experience (2005:330) such nara were often absent from the village during her visit tells us nothing about the state and extent of their knowledge.

'The house'

As noted already, we owe the term 'house-based societies' or 'house societies' (societes a maison) to Levi-Strauss's well-known reanalysis (1982) of the social organization of the Kwakiutl.

The kinship systems of the Kwakiutl of southwestern Canada, as well as the Yurok of northern California--both of them coastal populations--have been notoriously difficult to analyse. Franz Boas (1897, 1920) and Alfred Kroeber, who studied these societies for decades, were unable to decide whether they were organized according to principles of matrilineal or patrilineal descent. Boas speculated over whether the Kwakiutl were in the midst of a complex process of transformation from one to the other, or whether they could best be described as bilateral. He vacillated because the unit (or group) in question appeared to have features that anthropologists at the time normally associated exclusively with either form of unilineality. Boas, on his part, ended up rejecting classification altogether because none of the analytical categories available to him were relevant, resigning himself to describing the Kwakiutl tribal subdivision as a unique type of structure without equivalent. Instead he referred to it simply by its indigenous name, numaym, which means 'house'. (20)

Levi-Strauss compared traits of the Kwakiutl numayma (plural of numaym) with those of medieval European noble houses and with certain social forms of feudal Japan. He suggests (Levi-Strauss 1982:174) that had Boas and Kroeber been more familiar with European and Japanese history--or even with ancient Greece--the issues might have been resolved, for in these temporally remote societies 'the house' is a social group to which membership is not determined according to any single incontrovertible principle which is either present or absent. Rather, a person may be included according to a number of principles that can be invoked according to circumstance. It is neither possible nor necessary to go into the details of Levi-Strauss's reanalysis here. But I want to emphasize that it is informed not only by the particulars of Kwakiutl ethnography and of feudal social forms in Europe and Japan, but to a significant degree by readings and reappraisals of documents on African, Indonesian, Melanesian and Oceanic societies.

As has been suggested elsewhere (Carsten and Hugh-Jones 1995; Smedal 2002), Levi-Strauss's aim is to understand forms of social organization which, in his terminology, lie between the two poles of 'elementary' structures and 'complex' ones. At one pole are the constrictions of unilineally ascribed group membership already discussed and positive marriage rules, implying group exogamy, which often go with it. This is what characterizes the vast majority of social formations in eastern Indonesia. At the other pole are the open-ended and undifferentiated social formations where, apart from a small number of negative marriage rules, notably the prohibition on incest, marriages are contracted according to what are basically market principles: common tastes and interests, mutual attraction, beauty, prowess, wealth, class, and so on. In this ultimately evolutionist scheme (Carsten and Hugh-Jones 1995:9), 'the house', as Levi-Strauss sees it, represents a compromise. It contains a multitude of cross-cutting rules, rights, obligations, principles, preferences, inclinations, and possibilities which inform actions pertaining to marriage and descent, often in ways that may surprise the analyst.

The Ngadha House and gendered symbols

Schroter (2005:322) stresses the importance of 'the house' in Ngadhaland:
   The focus of each of these groups [saka, kaka, and dai, OHS] is the
   great lineage house (sa'o meze), the lineage's central sacred
   place, serving both as an ancestral shrine and a place of assembly.
   It is the house, even more than the sub-clan, that represents an
   individual's social and emotional ties; at the same time, it
   constitutes the most important economic and ritual collective
   entity. The members of any given house cultivate its land, share in
   its harvest, and shoulder the primary responsibility for raising
   money to fund a child's education, to finance a wedding ceremony,
   or to cover costly burial expenses.

Here, Schroter describes it well; the crucial position occupied by 'the house'--including its status as corporate group--cannot be missed from these remarks. It is odd, therefore, that she declares her 'reservations about' the societes a maison model proposed by Levi-Strauss 'since it is the unilineal kinship which is the most important for [the Ngadha]'. (21) Again, it is asserted (not shown or argued) that unilineality is the basic Ngadha principle. But there is also something curious here. Schroter writes as if oblivious to the fact that Levi-Strauss was motivated to develop the concept of 'the house' precisely because, as just noted, 'the house' indeed unites such principles as affinity and descent as well as other principles.

However, it is even odder that she writes, 'Women not only have access to the means of production but also own both land and houses. Their economic base is extraordinarily strong, and so they are not dependent on men.' (Schroter 2005:342.) In my view, this statement is simply erroneous if by 'land and houses' Schroter refers to sa'o which also, almost invariably, constitute landed estates and thus corporate groups. These estates are controlled by the House affiliates as a collective. Thus, even relatively minor repair work on a sa'o implicates all its affiliates (not merely its current residents), partly because such work always involves expenses (to be covered collectively) and partly because it always involves rituals which by definition implicates other (and sometimes all) sa'o affiliates. Likewise, every sa'o affiliate's access to land is routinely reviewed and renegotiated during the yearly reba festival, especially following events such as affiliates' marriage, death, or out-migration. That women can and do voice their interests at such venues is something Schroter and I can agree on. But based on my experience, it is highly unlikely that the interests of a woman--including the resident sa'o steward (a House 'daughter', weta)--would prevail in the event that they were contested by the woe and sa'o 'council of elders' (known as the mosa laki: adult, accomplished men). These men would be no other than her 'brothers' (nara), her 'uncles' (pame) and, at an older age, her sons. What Schroter calls 'ownership' of 'land and houses' is simply not an individual matter in Ngadhaland, nor is it specifically related to gender.

Schroter (2005:327) also describes 'the house' and its wood carvings: 'The one sa'o is separated from the inner veranda by a door and a little winged stairway, both beautifully adorned with carvings that show buffaloes, horses, and chickens'. This is not terribly incorrect, as far as the winged stairway is concerned, though from what I have seen stairways are decorated with carved 'animal' representations not of buffaloes, horses, or chickens, but of serpents or elephants. According to my own observations, however, 'the door itself is never subject to decoration'; instead 'the two boards on either side of it are' (Smedal 2000:117, Figure 4 [p. 118], Figure 6 [p. 120]). Moreover, the buffalo is never, unlike the horse and the chicken, represented in its corporeal outline, only metonymically, by its horns. Schroter overlooks, however, the arguably even more important carvings on the various horizontal and vertical beams (stylized representations of gendered gold jewellery, for example). Their importance lies in indexing precisely the very status of the (mature) House, which is discernible to the familiar eye. Notable, too, for their symbolism are the non-figurative carvings that adorn these House beams. A comprehensive analysis of Ngadha House carvings is presented in Smedal (2000), a work that Schroter evidently has not consulted.

Schroter comments also on what she calls the 'sub-clan shrines'--the ngadhu sacrificial pole and the bhaga miniature house--conceptually gendered representations of the woe's founding couple prominently displayed in the village plaza (see illustrations in Schroter 2005:324, 326), which the Ngadha say are 'husband and wife'. Schroter (2005:325) notes, correctly, that during important rituals the ngadhu is fed with, among other things, sacrificial blood, but she is mistaken when she states that 'chickens, pigs, and water buffaloes [...] are sacrificed at the pole'. As Schroter (2005:334) herself remarks in another connection, 'The reality is different'. Pigs and chickens are never sacrificed at the ngadhu. The water buffalo always is (Smedal 1996:48-9, note13, 2009a:280). Or quoting Molnar (2000:234) again, 'at the sacrificial post [...] pigs and chickens, which are symbolically valued as feminine animals, are forbidden to be killed. Only the symbolically masculine water buffalo [...] can be sacrificed at the post.'

It is uncontroversial to suggest, as Schroter does, that the Ngadha House (sa'o) displays gendered symbolism, much of it female or feminine. She writes that inside the sa'o, 'a representation of the uterus serves as the symbolic locus for female reproductive capacities and matrilineal descent'. (22) Presumably (she does not specify), she refers to the kobho, which is usually suspended at the ritual centre of the House (the mata raga 'rack' on its interior back wall). It is a small gourd container for palm gin, the ubiquitous libation during any Ngadha ritual. (23) But what Schroter does not tell her readers is that the mata raga rack also routinely holds 'the House's sacred sword (laja sue or sau ga'e) and lance (bhuja kawa), [and] its su'a sa'o, the "male" digging stick symbolizing the land over which the house has rights' (Smedal 2000:110). These objects two of which evoke armed conflict--would make the femininity of the sacred House centre appear far less totalizing than Schroter maintains. Schroter (2005:328) then writes:
   Saka pu'u and saka lobo houses are distinguished by miniatures on
   the roof. Pu'u is symbolically represented by another little house,
   comparable to the bhaga though much smaller, which is set on the
   roof. It is further distinguished by stylized hair pins fixed on
   each end of the roof, which are called ana lie, a term that refers
   both to female genitalia and to a firstborn child.

A first if minor problem is that this reads as if the 'stylized hair pins' are called ana lie, but this must be a slip of the pen; what Schroter probably intended to say is that the miniature house is called ana lie. At least that would tally with what she writes a few lines further: 'Interestingly, the feminine is represented by repeated references to the house which manifests itself as the bhaga and the ana lie'.

But more vital than this probable slip is Schroter's identification of the symbolism of the miniature house ('female genitalia') and on this we disagree. In my own work, I have noted the term ana ie for the miniature house (Smedal 2002:496). According to Arndt, (24) this is the correct term and Arndt provides no further semantic breakdown of the expression. However, there is a separate entry for cice (now conventionally spelled ie) (Arndt 1961:211) which he translates as 'friend', and another entry for lice (now conventionally spelled lie) (Arndt 1961:296) which he also translates as 'friend'--the two words being dialect versions of the same basic term. A reasonable conclusion so far is that the compound ana ie signifies exactly the same thing as does ana lie (ana by itself having a range of meanings, of which 'child', 'human', and 'member' are among the most common). Thus an approximate guess at an apposite English gloss would be 'friendly people'--not an unreasonable appellation for a symbol (itself a house) placed atop the ritually most significant House of an entire woe. What, then, might be the basis for Schroter's claim that ana lie refers to female genitalia? Perhaps the answer lies in her misreading Arndt. According to him, cie means 'clitoris' (Arndt 1961:211) and so does lie (Arndt 1961:295). But note the very absence of the 'c' prior to the ultimate 'e' in these two words. In Arndt's orthography such absence of a 'c' between two vowels signifies a glottal stop--a distinct phoneme in the Ngadha language--between the vowels in question. This absence means that these two words, using present conventions (where the apostrophe signifies the glottal stop), would be spelt i'e and li'e, respectively. (25) Schroter's assertion, therefore, is baseless. There is no such thing atop the Ngadha sa'o saka pu'u (trunk House) roofs as a miniature house symbolizing either the clitoris or any other part of the female genitalia. What is on top of these roofs is instead a miniature house symbolizing an affable collectivity, a collectivity which consists of men as well as women. (26)

To impute gender-specific meaning to symbolic representations where there is none is ethnographically irresponsible, just as it is to omit mention of dual-gendered symbolic representations. I agree with Schroter that in Ngadhaland, named Houses are conceptually more 'feminine' than 'masculine,' but the point is that they are always both. In this aspect, Ngadha Houses have much in common with Houses elsewhere in Flores and eastern Indonesia. Whether such objects are best referred to as 'androgynous' (Howell 2002) or 'dual-gendered' (Hoskins 1998:187-90) is certainly debatable, but a prior task is to render faithfully the relevant ethnographic evidence itself, which I think Schroter fails to do.

As already noted, I have taken inspiration from Levi-Strauss in my own work on Ngadha social organization. Yet, as I have written elsewhere (Smedal 2002:517-8; see also Howell 1995:150-1), my view is that, for comparative purposes, the analytical value of Levi-Strauss's 'house societies' concept is limited. There are two reasons for this. The first is that there is such a plurality of principles, dispositions, and proclivities that 'the house' (as Levi-Strauss specifies it) may contain--in so many combinations--that the range of difference would render comparative exercises all but vacuous. 'The house' approaches, in fact, a polythetic category (Needham 1975). The second is that to the extent that Levi-Strauss writes about 'the house' in specific societies, he discusses it in most detail as it is found in social formations that can best be described as adhering to a unilineal principle of descent (Carsten and Hugh-Jones 1995:18-9; Gibson 1995:131). The problem with this is that in such social formations 'the house' is, as it were, redundant; 'the house' is demonstrably there, but so is an incontrovertible unilineal principle (a 'constitutive rule') which is more elemental. There is a small irony in this. Levi-Strauss (1987) first developed the notion of 'house-based societies' precisely in an effort to understand how groups that are most fundamentally organized according to a cognatic principle can remain fairly stable over time, as is evident from his comments on various Borneo ethnographies. A key result of Levi-Strauss's considerations (1987:127) in this regard is his suggestion that what takes the place of 'rules of descent' in these cognatic social formations are 'rules of residence', rights in land assuming priority (over descent as such):
   In effect, [cognatic] systems bring in a new dimension, since they
   define, perpetuate and transform the mode of social cohesion
   through a relation that is no longer to a fixed rule of descent,
   but to a system of rights in land.

Since the organizing principle in cognatic systems has been so difficult to pin down one is obliged, when searching for the basis of the social order, says Levi-Strauss (1987:154),
   to move successively from relations of kinship and descent to
   property relations, then, in the absence of a rule of perpetual
   succession to a hereditary office or domain, to relations of
   residence, the greater part of them contingent.

What I hope to have demonstrated is that unlike most ethnic groups in eastern Indonesia, the Ngadha are not, pace Schroter, unilineal. When the ethnographic phenomena are explored in their specificity 'before declaring the relevance or not of already accepted understanding' (Kapferer 1997:11), it becomes clear that there is no incontrovertible descent dogma, no principle, no ideology, no doctrine which dictates that a Ngadha House is perpetuated by women only, or by men only. (27) Instead, by fetishizing marriage and the procreative principle, the 'cognatic problem'--how can cognatic kinship give rise to stable groups? (Strathern 1992)--is solved: married couples live in Houses (Bloch 1993). This way, the contingent relations of residence (see Levi-Strauss just quoted), marriage, procreation, identity, and belonging are dealt with simultaneously. The means for fixing residence is--as Arndt, Robert H. Barnes, Molnar, and I have stressed--bridewealth. In short, the absence of a unilineal dogma in Ngadhaland is the ground on which the figure of the House must be understood. Thus the presence of the House would justify placing the Ngadha under the rubric Levi-Strauss (1983:1222) labelled 'pseudomorph': social formations which appear to be either patrilineal or matrilineal but which on closer inspection turn out to be neither.

A function, perhaps, of the absence of a unilineal dogma (at least concurrent with it) is that marriage may take two forms: either the husband joins the wife's House, or the wife joins the husband's House--although strictly speaking (because Houses are not exogamous units), they may both be born to the same House. Marriage, as already noted, is permitted between (third) cousins and beyond on both sides; House endogamy is thus not uncommon. In either of the two marriage types, children will be affiliated to the House of the spouse who remains put. Routinely, this involves residence near the House itself and the right to cultivate House land as well, although House affiliation remains unaltered even when people move to other islands.

Two particulars about these separate forms of marriage are indisputable. The first is that both involve bridewealth prestations, but in unequal proportion. When prestations are modest the woman remains in the House, and when they are substantial she moves out. The second is that each form is practised to the near-exclusion of the other in separate geographical areas. People have common knowledge of both forms, yet in both areas they stick to their practices and say that what is done 'over there' is unworkable for them. What unites them is that the system, understood in its totality, is exchange driven--not descent driven.

To sum up, in the greater part of Ngadhaland (except in the Were region, that is), the outcome of persistent uxorilocality and matrilateral affiliation over time is that it seems as if House affiliation is dictated by a principle of matrilineality. Sisters and daughters remain, brothers and sons marry out generation after generation. Moreover, unlike the case in the New Guinea Highlands, Ngadha genealogies are not shallow. They are well kept, and may show how daughters have regularly replaced daughters for eight or ten generations. It is even sometimes said that a House without daughters is a House that dies out. As noted earlier in this article, there are emergency procedures (such as pagu raru dhutu nu and dhepo haki) that the Ngadha can resort to, but the procedures depend on certain micro-demographic particulars which may or may not be present.

Rules for uxorilocal residence and matrifiliation produce a very solid accumulated pattern of 'Houses of women'. But these rules--that daughters remain, and sons marry out--are regulative, not constitutive rules. The pasa option--that a resident son can bring a woman to the House--is well known and sometimes chosen. Ngadha House rules can be bent (and those who do the bending are sometimes admired). A constitutive rule of descent is something of an entirely different order.


Schroter (2005) observed that in the greater part of Ngadhaland, sa'o (and hence woe) recruitment regularly follows matrifiliation, and she concludes from this fact the presence of an ideology of unilineal, that is, matrilineal, descent. I made the same observation, but I also observed that in the Were region sa'o (and hence woe) recruitment follows patrifiliation. The fact that marriage in Were is of the pasa type, and that elsewhere in Ngadhaland it is also permissible--however rarely it may occur--to contract pasa marriages, led me to the conclusion that there exists no incontrovertible principle, no constitutive rule, which dictates that sa'o and woe recruitment takes place according to a unilineal principle. Instead, I have already shown, and I think conclusively (Smedal 2002), that whether or not recruitment into Ngadha social units follows matrifiliation or patrifiliation varies with the size of bridewealth. To put it in the simplest of terms, in accordance with the arguments advanced by Fortes and Scheffler already referred to, while it is true that in much of Ngadhaland a child is affiliated to the sa'o (and woe) of its mother, it is not necessarily the case that the mother herself was affiliated to the same sa'o at birth. The possibility that the mother was affiliated to another sa'o at birth is what the institution of pasa provides for. Thus Schroter's assertion, that the Ngadha are 'unilineal' and 'matrilineal', is untenable.


A period of 25 months of field research in Flores (1990-1991, 1993; 1997-2010 intermittently) in three 'administrative villages' (desa), Dariwali, Nenowea, and Watumanu, in what is now the subdistrict (kecamatan) of Jerebuu (formerly part of subdistrict Aimere), supplemented by intermittent investigations in Golewa subdistrict and among transmigrant Ngadha in Kalimantan was facilitated by fellowships and travel stipends from the (then) Norwegian Research Council for Science and the Humanities; the Institute for Comparative Research in Human Culture, Oslo; the Department of Anthropology, University of Oslo; the Faculty of Social Science and the Meltzer Fund, University of Bergen; and by a Research Council of Norway grant to the MEISA project (Migrants and Entrepreneurs in Insular Southeast Asia). Research was conducted under the auspices of the Lembaga Ilmu Pengetahuan Indonesia (LIPI, Indonesian Institute of Sciences) and sponsored in various ways by Nusa Cendana University in Kupang, Timor, and by Palang Karaya University, Central Kalimantan. I am grateful to the bodies mentioned for their assistance and to the numerous Ngadha women and men for having endured my intermittent presence in the course of two decades. I also wish to thank my colleagues Eldar Braten and Edvard Hviding, as well as Bijdragen's two anonymous reviewers, for critical and constructive comments to earlier versions of this article. The usual caveats apply.


Allerton, Catherine 2004 'The path of marriage: Journeys and transformation in Manggarai, eastern Indonesia', Bijdragen tot de Taal-, Land- en Volkenkunde 160:339-62.

Arka, I Wayan, Jeladu Kosmas and I Nyoman Suparsa 2007 Bahasa Rongga: Tatabahasa acuan ringkas. Jakarta: Penerbit Universitas Atma Jaya (PUAJ).

Arndt, Paul 1936 'Deva, das Hochste Wesen der Ngadha' [Part I], Anthropos 31:894-909.

1954 Gesellschaftliche Verhaltnisse der Ngadha. Wien-Modling: Verlag der Missionsdruckerei St. Gabriel. [Studia Instituti Anthropos 8.]

1960 'Mythen der Ngadha', Annali Lateranensi 24:9-137.

1961 Worterbuch der Ngadhasprache. Fribourg, Suisse: Pertjetakan Arnoldus, Endeh/Posieux. [Studia Instituti Anthropos 15.]

Bader, Hermann 1953 Die Reifefeiern bei den Ngada (Mittelflores, Indonesien). Wien-Modling: St. Gabriel. [St.-Gabrieler Studien 14.]

Barnes, John A. 1962 'African models in the New Guinea Highlands', Man 62:5-9.

Barnes, Robert H. 1972 'Ngada', in: Frank M. LeBar (ed.), Ethnic groups of insular Southeast Asia: Vol. 1: Indonesia, Andaman Islands, and Madagascar, pp. 83-6. New Haven, CT: Human Relations Area Files.

1980 'Marriage, exchange and the meaning of corporations in eastern Indonesia', in: J.L. Comaroff (ed.), The meaning of marriage payments, pp. 93-124. London: Academic Press. [Studies in Anthropology 7.]

Blackwood, Evelyn 1995 'Senior women, model mothers and dutiful wives: Managing gender contradictions in a Minangkabau village', in: Aihwa Ong and Michael G. Peletz (eds), Bewitching women, pious men: Gender and body politics in Southeast Asia, pp. 124-58. Berkeley, CA: University of California Press.

1999 'Big houses and small houses: Doing matriliny in West Sumatra', Ethnos 64-1:32-56.

Bloch, Maurice 1993 'Zafimaniry birth and kinship theory', Social Anthropology 1 (1B):119-32.

Boas, Franz 1897 The social organization and the secret societies of the Kwakiutl Indians. Report of the US National Museum for 1895. Washington, DC.

1920 'The social organization of the Kwakiutl', American Anthropologist (N.S.) 22:111-26.

Carrier, Achsah H. and James G. Carrier 1991 Structure and process in a Melanesian society: Ponam's progress in the twentieth century. Chur: Harwood Academic Publishers. [Studies in Anthropology and History 1.]

Carsten, Janet and Stephen Hugh-Jones (eds) 1995 About the house: Levi-Strauss and beyond. Cambridge: Cambridge University Press.

Codere, Helen 1990 'Kwakiutl: Traditional culture', in: William C. Sturtevant (general editor) and Wayne Suttles (volume editor), Handbook of North American Indians: Northwest Coast, pp. 359-87. Washington, DC: Smithsonian Institution. [Handbook of North American Indians 7.]

Cole, Stroma 2007 Tourism, culture and development: Hopes, dreams and realities in east Indonesia. Bristol: Channel View.

Evans-Pritchard, Edward Evan 1940 The Nuer: A description of the modes of livelihood and political institutions of a Nilotic people. Oxford: Oxford University Press.

Fischer, H.Th. 1964 'The cognates in the Minangkabau kinship structure', Oceania 34:96-110.

Fortes, Meyer 1945 The dynamics of clanship among the Tallensi. Oxford: Oxford University Press.

1949 The web of kinship among the Tallensi. Oxford: Oxford University Press.

1959 'Descent, filiation and affinity: A rejoinder to Dr. Leach', Man 59, Nos. 309, 331:193-7, 206-12.

1969 Kinship and the social order: The legacy of Lewis Henry Morgan. London: Routledge and Kegan Paul.

Forth, Gregory 2007 'Can animals break taboos? Applications of 'taboo' among the Nage of Eastern Indonesia', Oceania 77-2:215-31.

Fox, James J. (ed.) 1980 The flow of life: Essays on eastern Indonesia. Cambridge, Mass.: Harvard University Press. [Harvard Studies in Cultural Anthropology 2.]

Gibson, Thomas 1995 'Having your house and eating it: Houses and siblings in Ara, South Sulawesi', in: Janet Carsten and Stephen Hugh-Jones (eds), About the house: Levi-Strauss and beyond, pp. 129-48. Cambridge: Cambridge University Press.

Goody, Jack R. 1983 The development of the family and marriage in Europe. Cambridge: Cambridge University Press.

Helms, Mary W. 1998 Access to origins: Affines, ancestors, and aristocrats. Austin: University of Texas Press.

Hoskins, Janet 1998 Biographical objects: How things tell the stories of people's lives. London/ New York: Routledge.

Howell, Signe 1995 'The Lio House: Building, category, idea, value', in: Janet Carsten and Stephen Hugh-Jones (eds), About the house: Levi-Strauss and beyond, pp. 149-69. Cambridge: Cambridge University Press.

2002 'Nesting, eclipsing and hierarchy: Processes of gendered values among Lio', Social Anthropology 10-2:159-72.

Jonaitis, Aldina 1991 Chiefly feasts: The enduring Kwakiutl potlatch. Seattle/London: University of Washington Press.

Kapferer, Bruce 1997 The feast of the sorcerer: Practices of consciousness and power. Chicago: University of Chicago Press.

Kuper, Adam 1988 The invention of primitive society: Transformations of an illusion. London/ New York: Routledge.

Lepervanche, M. de 1967 'Descent, residence and leadership in the New Guinea Highlands (part I)', Oceania 37:134-58.

Levi-Strauss, Claude 1982 The way of the masks. Translated by Sylvia Modelski. Seattle: University of Washington Press. [Originally published as La voie des masques. Geneve: Skira, 1975.]

1983 'Histoire et ethnologie', Annales 38:1217-31.

1987 Anthropology and myth: Lectures 1951-1982. Translated by Roy Willis. Oxford: Basil Blackwell. [Originally published as Paroles donnees. Paris: Plon, 1984.]

McKinnon, Susan 1991 From a shattered sun: Hierarchy, gender, and alliance in the Tanimbar Islands. Madison, WI: University of Wisconsin Press.

Molnar, Andrea Katalin 2000 Grandchildren of the Ga'e ancestors: Social organization and cosmology among the Hoga Sara of Flores. Leiden: KITLV Press. [Verhandelingen 185.]

Needham, Rodney 1974 Remarks and inventions: Skeptical essays about kinship. London: Tavistock.

1975 'Polythetic classification: Convergence and consequences', Man (N.S.) 10:349-69.

Peletz, Michael G. 1995a 'Neither reasonable nor responsible: Contrasting representations of masculinity in a Malay society', in: Aihwa Ong and Michael G. Peletz (eds), Bewitching women, pious men: Gender and body politics in Southeast Asia, pp. 76-123. Berkeley, CA: University of California Press.

1995b Reason and passion: Representations of gender in a Malay society. Berkeley, CA: University of California Press.

Reenen, Joke van 1996 Central pillars of the house: Sisters, wives, and mothers in a rural community in Minangkabau, West Sumatra. Leiden: Research School CNWS. [CNWS Publications 45.]

Rivers, William Halse Rivers 1924 Social organization. Edited by W.J. Perry. London: Kegan Paul, Trench Trubner. [History of Civilization.]

Scheffler, Harold W. 1985 'Filiation and affiliation', Man (N.S.) 20:1-21.

1986 'The descent of rights and the descent of persons', American Anthropologist 20:339-50.

2001 Filiation and affiliation. Boulder, CO: Westview Press.

Schlegel, Alice 1972 Male dominance and female autonomy: Domestic authority in matrilineal societies. New Haven, CT: Human Relations Area Files.

Schroter, Susanne 2005 'Red cocks and black hens: Gendered symbolism, kinship and social practice in the Ngada highlands', Bijdragen tot de Taal-, Land- en Volkenkunde 161:318-49.

Smedal, Olaf H. 1996 'Conquest and comfort: The Ngadha "bad death" ritual', in: Signe Howell (ed.), For the sake of our future: Sacrificing in eastern Indonesia, pp. 43-72. Leiden: Research School CNWS. [CNWS Publications 42.]

2000 'Sociality on display: The aesthetics of Ngadha Houses', RES: Anthropology and Aesthetics 37:106-26.

2002 'Ngadha relationship terms in context: Description, analysis, and implications', Asian Journal of Social Science 30-3:493-524.

2009a 'On the value of the beast, or the limit of money: Notes on the meaning of marriage prestations among the Ngadha, central Flores (Indonesia)', in: Knut Rio and Olaf H. Smedal (eds), Hierarchy: Persistence and transformation in social formations, pp. 269-97. Oxford/New York: Berghahn.

2009b 'Hierarchy, precedence, and values: Scopes for social action in Ngadhaland, central Flores', in Michael P. Vischer (ed.): Precedence: Processes of differentiation in the Austronesian world, pp. 209-27. Canberra: ANU E Press.

Strathern, Marilyn 1992 'Parts and wholes: Refiguring relationships in a post-plural world', in: Adam Kuper (ed.), Conceptualizing society, pp. 90-116. London: Routledge.

Tule, Philipus 2006 'We are children of the land: A Keo perspective', in: Thomas Reuter (ed.), Sharing the earth, dividing the land: Land and territory in the Austronesian world, pp. 211-36. Canberra: ANU E Press.

(1) It is perhaps necessary to state that what prompted me to employ Levi-Straussian terminology is not that Ngadha social organization in my view is based on cognatic principles (which may or may not precipitate 'house societies'). It is because of the crucial role of the House among the Ngadha, a view which Schroter and I share.

(2) There are presumably very few empirical examples of group composition that match this model. Scheffler (1986:344-5, 2001:107-19) notes that land-owning groups of agnates in northern Pakistan, the so-called khel, have features that seem to match the model well. The key question is whether khel membership hinges on patrifiliation per se or on inheritance--two issues that Rivers taught us to keep separate. Khel membership appears to accompany individually acquired rights to land, and these rights can only be transmitted from father to son; never from grandfather to grandson. A further albeit not decisive concern is the eventual destinations of potential members who, so to speak, fail to qualify. Obviously, they cannot join another khel, for we know already that patrifiliation is a necessary condition for membership in any one of them--and sons can only be patrifiliated to one man.

(3) 'It would be interesting to work out how, say, the Nuer might have been described if the only analytical models available had been those developed to describe, say, Cimbu and Mbowamb.' (J.A. Barnes 1962:9.)

(4) Schroter uses the spelling 'Ngada'. While in the past some authors have spelt the ethnonym that way (and in other ways as well), it is clear that the name of the kin group ('clan') from which the name of today's administrative district (kabupaten or regency) was taken, is 'Ngadha', the 'dh' signifying a phoneme (an implosive or preglottal 'd') distinct from 'd' (the two phonemes coexist in the Ngadha language)--a distinction not recognized in official Indonesian (see also Molnar 2000:1, note 1). It is true, as Schroter (2005:319) notes that in his first writings Paul Arndt, a missionary-priest-cum-ethnographer who spent some 40 years in Ngadhaland, was not consistent in rendering this phoneme. But since 1936 he settled on 'Ngadha' as the preferred spelling of the name of the group--the district Ngada being the home to several other ethnic groups, some of whose languages are rather different from the Ngadha language.

(5) By 'greater Ngadhaland' I refer to a sociocultural (not administrative) unit in the southwestern part of the administrative district (kabupaten) Ngada in central Flores, which is 1. a linguistically homogenous area (linguistic variation is chiefly phonological) within which 2. people celebrate the annual village festival known as reba and where 3. kin groups (woe) comprising a set of Houses are represented by two conceptually gendered ritual foci in the 'traditional village' (nua) plaza: the male ngadhu (or madhu) and the female bhaga. The north-eastern corner of this area is occupied by the village of Naru, a few kilometres north of the district capital Bajawa on the main road to So'a. The remote mountain village of Lopijo, close to the border of the district of Manggarai, is located in the north-western corner and the coastal subdistrict (kecamatan) capital Aimere is found in the south-west corner. The coastal village Boba occupies the south-eastern corner of Ngadhaland. The population groups beyond this 'greater Ngadhaland' area are the Manggarai and the Rongga to the west, the Nage and the Keo to the east, and the people in So'a to the immediate north; none of them celebrate reba, none of them display ngadhu and bhaga, and all of them speak languages other than the ones spoken by the Ngadha. See Molnar 2000:24, for a more detailed set of Ngadha diacritica.

(6) That said, certain woe play a decisive, structural role in that marriage between their affiliates is prohibited.

(7) Like Schroter, I was naturally intent on determining whether 'cili-bhou' is still in use (and if so, for what kind of social unit), arriving, as she did, at a negative conclusion.

(8) In my experience, the uncompounded word woe applies exclusively to ideologically stable and discrete social networks claiming the right to separate ngadhu and bhaga. By definition these networks have a name and consist of minimally two named Houses or sa'o. The names of both types of unit can be changed. By definition, also, woe are autonomous polities at the socio-ritual level. I have never heard woe used for lower-level units, in some parlances, as 'sib' or '(extended) family'--and indeed none of the numerous Ngadha men and women I ever consulted about the matter recognized such usage either--nor have I recorded the existence of a unit above that of woe.

(9) 'Voce' is Arndt's spelling of what is now conventionally spelt 'woe'.

(10) Cole 2007:76; Molnar 2000:89-114; Tule 2006:219; see also Forth 2007:219 on the neighbouring Nage to the east; and Arka et al. 2007:50 on the neighbouring Rongga to the west; the languages here have much in common with the Ngadha language.

(11) Smedal 2002:502-5. See also Molnar 2000:121-4 for the very similar way bridewealth payments operate among the neighbouring Hoga Sara.

(12) See also Arndt 1954:42. In Were, where postmarital virilocal residence is the norm, full bridewealth consists of 21 named items totalling 19 horses and 4 buffaloes. In order to amass wealth of this scale and organize its transfer, grooms (together with their close patrilateral kin) draw on the fact that they have previously assisted other men in their efforts at producing bridewealth and can now call on their assistance to support the groom in question. In Were, therefore, 'a relatively large economic value characteristically floats about, chiefly in the form of claims--settling temporarily among groups of kin (Houses) as marriages are contracted. In the other region (Jerebu'u) the "floating value" is very much smaller' (Smedal 2002:503-4). In Were, a prospective bride's father and brothers will seek to retrieve the bridewealth expenses they incurred in connection with their own marriages by insisting on comparable bridewealth payments for her. A crucial point is that unlike the Ngadha's neighbours (for example, the Manggarai to the west, the people in So'a to the north, and the Nage and the Keo to the east), Ngadha bridewealth prestations are not reciprocated by specific items (valuable cloth, for example); Ngadha counter-prestations consist in large-scale feasting--that is to say, in the (joint) consumption of perishables--whenever a bridewealth item is presented. The Ngadha say that in economic terms, the resources expended by bridewealth recipients during such feasting is usually equal to and may even exceed the value of the bridewealth items received. These particulars--that donors are not reciprocated with valuables and that the bride's group spends much upon receiving--help explain why in Ngadhaland bridewealth is paid in instalments over many years. In many cases the greater part of the bridewealth is never paid, in fact never even asked for. The rationale for this, according to the Ngadha, is that the greater the unpaid part of the bridewealth, the stronger the bond between the two groups.

(13) The following statement does not in my view inspire confidence in Bader's powers (1953:98, note 20) of observation:
   The single family, both in the restricted sense (sa tuka, mogo sa
   tuka; who come from the same body) and in the extended sense (all
   closer kin who live in the same house: isi sa'o, tuka sa'o: the
   entire House) plays only a minor role in Ngadha and hardly emerges
   in public life.

Depending partly on how one construes the meaning of the last part of the sentence ('minor', 'hardly'), the general drift appears to be that in Bader's opinion the House--as a kin group--is of negligible sociological import. I think Schroter would agree with me that Bader's grasp of Ngadha sociology was less than firm.

(14) Robert H. Barnes (1980:111, citing Arndt 1954:60-70) summarizes Arndt's remarks on the matter by saying that 'divorce is rare when bridewealth has been paid, more frequent otherwise'. But it must be borne in mind that the divorces Arndt could have referred to would concern men and women born 20, 30, even 40 years earlier (prior to his arrival in Flores in 1924, see Arndt 1954:vii), when Christian conversion--and with it, church solemnization--would have been nonexistent or at best erratic.

(15) See Arndt 1954:44-6 who is absolutely clear about this.

(16) The institution of Ngadha bridewealth is a central topic in Smedal (2002, with further analytical treatment in Smedal 2009a) and the buffalo features prominently.

(17) This ritual takes place relatively infrequently and in my field locations the ritual never took place during my doctoral field research. Having now witnessed two occurrences (in October 2007 and June/July 2009) I hope to present an analysis of it in the future.

(18) Later, Schroter (2005:338-9) contradicts herself on this matter, writing (this time correctly) that, 'if a high-ranking girl or woman was accused of having sexual relations with a low-ranking boy or man, both were ritually killed'.

(19) See Smedal 2002 where this comparative point is discussed.

(20) The term more narrowly means 'one kind'; it signifies a 'kinship group' and, significantly, this 'basic social group' ... 'owned one or more plank houses in a winter village and several seasonal sites where it had the right to harvest certain resources' (Jonaitis 1991:86 quoted in Helms 1998:17); for a succinct exposition of the Kwakiutl numaym, see Codere 1990:366-8.

(21) Schroter 2005:321, note 6. It is worth recalling that two decades ago McKinnon (1991:30-1), reviewing some of the 'house society'-inspired literature published then, and wary of the confusion potentially arising from the attempts at reformulating ethnographica in the new idiom, sensed that however promising the concept of the 'house' and 'house societies' may seem, the old categories of 'descent' and 'alliance' persist in writers' vocabularies and have yet to be disentangled from them.

(22) Schroter 2005:328. That I find the term 'matrilineal descent' inapplicable in the Ngadha context has already been made plain.

(23) It is perhaps worth noting, too, that charred leaves of the kobho plant are often smeared on the stump of newborn babies' umbilical cord.

(24) Arndt 1961:212, see cana cice; Arndt's Ngadha orthography is somewhat idiosyncratic.

(25) The primary meanings of li'e are 'seed' and 'kernel', the extended meaning of which is any small solid object (its function as a numeric classifier is analogous to Bahasa Indonesia and Standard Malay biji). Li'e is also a common locution for gold pieces and is, curiously enough, sometimes used to signify, in addition to clitoris, testicles.

(26) Another possible interpretation is that ie also signifies the neighing of a horse (Arndt 1961:212). Thus the ritual parallel speech couplet
ie moe jara masi     neigh like a horse at the top
kako moe manu jago   crow like a rooster

signifies triumph, or pride; the connotation being that the ana ie can only be placed atop the ridge purlin of the trunk House when it is ritually inaugurated at great expense (including buffalo sacrifice at the ngadhu) during the ka sa'o ritual, when hundreds of guests are usually present. A variation of the significance of this triumphant 'horse at the top' is that Jara Masi is also the personal name of a mythical figure--the husband of Inerie, the highest mountain in Nghadha land (Arndt 1960:134-7).

(27) Molnar (2000:151-3), in her study of the neighbouring Hoga Sara whose sociocultural life has much in common with that of the Ngadha, is equally clear: 'The concept of lineality does not help much in explaining house membership [...] I propose that we dispense with this concept of lineality in the sense of descent, as it does not help us understand the various aspects of group membership and of social organization [...]. The use of the concept of "house society" is more applicable to the Hoga Sara. [...] the concept of the house cross-cuts the traditional categories of descent and affinity.'

OLAF H. SMEDAL is Professor of Social Anthropology at the University of Bergen. He holds a PhD from the University of Oslo. Some four years of fieldwork in Indonesia have resulted in Order and difference: An ethnographic study of Orang Lom of Bangka, West Indonesia, Oslo: University of Oslo, 1989, and a series of articles on social organization, kinship, language, ritual, value hierarchies and aesthetic expressions of two ethnic groups in Indonesia (the Lom, Bangka and the Ngadha, Flores). He edited, with Knut M. Rio, Hierarchy: Persistence and transformation in social formations, New York: Berghahn, 2009. His most recent research is on migration processes in Indonesia and local effects of political and economic changes in the country following the end of President Soeharto's rule of power. Professor Smedal can be reached at
COPYRIGHT 2011 Brill Academic Publishers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2011 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Smedal, Olaf H.
Publication:Journal of the Humanities and Social Sciences of Southeast Asia and Oceania
Article Type:Essay
Geographic Code:9INDO
Date:Apr 1, 2011
Previous Article:Dakwah, competition for authority, and development.
Next Article:Performance and political change in Indonesia.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |