Abstract
This paper takes a critical look at the discursive construction of the identity of "the Mentawaians" who are purported to inhabit the Mentawai islands in Western Indonesia. Out of a range of possible indigenous representations of identity "Mentawaian" is the one that travellers and scholars universally use. I trace the formation of this representation in the very early literature on the islands authored by traders, travellers, scholars and missionaries. I then go on to examine the (hegemonic) effects of this on the way in which anthropologists have subsequently come to discursively construct the "culture" of the "Mentawaians". As an alternative, I propose that anthropological scholarship needs to take greater heed of the ways in which the local inhabitants construct their own identities. I briefly illustrate this through the example of my own work which describes a shifting and contextual construction of local identity in a particular locale on Siberut, the largest of the Mentawai islands.
 

The last decade and a half has been a time of reassessment and soul-searching for anthropology, with an extensive examination of the relationship between anthropology and its ‘object’, the Other. Or, in other words, anthropologists long aware that "dominant conceptions of the world" circulating amongst the people they study tend to exert "not only the constraining power of dominance over other modes of thought but also [exhibit] the inertial authority of habit and instinct ... the horizon of the taken for granted, what the world is and how it works ..." (Hall 1988:44), are increasingly mindful that this refers equally to their own circumstances. Hence the need within scholarship to engage in eternally reflexive and vigilant critical practices in order to bring this horizon continually into relief. Allied to this is the awareness that social reality is complex in the late 20th century and requires, accordingly, complex analytic strategies to deal with it. This has resulted in the appearance of a variety of theoretical innovations within anthropology(1[see endnotes]). A prominent theme is that the once taken-for-granted notion of a world constituted by discrete bounded peoples/societies defined by various distinctive characteristics is increasingly viewed as having limited utility (Wolf 1982, 1988). Closely related to this is a move away from grand narrative explanations towards particularist analytic strategies which deal with complexity on its own terms (cf Fardon 1990; Lomnitz-Adler 1991; Alonso 1994; Alvarez 1995; Moore 1994; Kearney 1995; Mouzelis 1992). All this implies in the first instance the need to revisit the ethnographic record, in the interests of subjecting the portraits of the past to critical scrutiny. It also implies the need to guard against the taken-for-granted constructs that shape ethnographic projects in the present, an ideal increasingly realised in much contemporary ethnography.

These were the methodological sensibilities that accompanied me to the field in early 1992 with the general aim to conduct research on the kinship system of the ‘Mentawai’ people or ‘Mentawaians’ on the island of Siberut, the northernmost of four making up the ‘Mentawai’ islands located approximately 150km off the west coast of Sumatra, West Indonesia. And indeed following a brief period in the field ‘Mentawaian’ language, culture, and society proved to be a little more subtle and complex than prevailing characterisations in the literature allowed. This paper, then, attempts to sketch the development and structure of a discourse—what I term the "discourse of the Mentawaians"—as it manifests itself in much of the literature about the islands, an "horizon of the taken-for-granted" that, I contend, acts to limit ethnographic recognition of the range of possible socialities in this corner of the world. Adapting Hall's (1996) concept of "identification" I demonstrate that the category ‘Mentawai’ and its concomitant identity categories ‘Mentawaian’ or ‘the Mentawaians’ do not do justice to the complex and contradictory nature of identities as these are constructed in the life-world of those who can be conceived to fall within the boundaries of these generic categories. But more fundamentally the paper attempts to critically assess an approach that has formed the foundation for a good proportion of the ethnography of Indonesian societies, and which has been instrumental in producing the fully developed phase of the "discourse of the Mentawaians", the "field of study" perspective. I argue that in this particular ethnographic context, it tends to obscure more than it illuminates.
 

The invention of ‘Mentawai’.

I have placed the usages ‘Mentawai’/‘Mentawaian’ within inverted commas since it is my contention that we need to treat these other than as an unproblematic and self-evident categories or entities, despite the fact that, somewhat paradoxically, ‘Mentawai’, from the point of view of administrators and scholars remains, a self-evident entity in need of no qualification (although as an official administrative category it is of less relevance). Indeed the usage ‘Mentawai’ does enjoy a quasi-official status indicating a general geographical location, although there is very much more than this encapsulated within it when treated as a complex metaphor. It is accorded differential valency in popular, administrative, and scholarly discourse.

The islands constituting the group Siberut, Sipora, and North and South Pagai, are divided up into four Kecamatan (districts) amongst several constituting the Kabupaten (regency) of Padang/Pariaman. On the Sumatran mainland, in particular the provinces directly adjacent to the islands, West Sumatra and Riau, the names Sipora, Pagai, or Siberut mean little to people in popular discourse. What meaning they do have is in terms of the entity known as ‘Mentawai’. When mainlanders learned that I was conducting research in their area they would invariably ask where my location was exactly. In the early stages of the research I would reply that I was working on Siberut. When many expressed puzzlement as to where this was, I would rephrase the answer stating that I was located on ‘Mentawai’ which they would immediately be able to locate cognitively. People who knew of Siberut, Pagai and Sipora would invariably qualify my answer to their question: "Oh, so you’re working in Mentawai". Having gained some understanding of this I came to answer with ‘Mentawai’ first and ‘Siberut’ second.

The term ‘Mentawai’ has a long history. It was established as customary usage early in the 19th century and became firmly entrenched in the 20th where the category ‘Mentawai’ has become rigid, self-evident and unquestioned. Articles written about the southern (Pagai) islands in the early 19th century generally took the ‘Pagai’ islanders as their subject matter rather than ‘the Mentawaians’. However in the 20th century, work concerning the Pagai islands(2) and Siberut(3) is specifically about ‘the Mentawaians’ and not the ‘Pagai islanders’ or the ‘inhabitants of Siberut’ as they are often described in the 19th century literature. So how, then, did this come to be the case? In what follows I will concentrate upon several key texts that I argue have given birth to the contemporary anthropological and administrative identity category, 'Mentawai'.

John Crisp's (1799) brief account marks the entry of the islands into western discourse. Crisp took as his subject the "inhabitants" of the "poggy", or "nassau" islands which were to be found off the west coast of Sumatra (Crisp 1799:77). He reports that the name "nassau" was given to the islands by the Dutch. However, he also reports that the "inhabitants" refer to themselves as "Poggy", whereas the "people of Sumatra" call them "Orang Mantawee" (Crisp 1799:81). He considers this usage to originate in the "natives" own language from the word "Mantaoo signifying a man" an explanation for the origin of the usage ‘Mentawai’ cited by subsequent analysts (Loeb & Heine-Geldern 1935:158; Schefold 1991:32). He goes on to relate that "Porah or Fortune island" (Sipora) is "inhabited by the same race of people" (Crisp 1799:83). Several pages later we learn that the "different tribes of Orang Mantawee, who inhabit the Poggy islands" do not "war" with each other although this is not the case in relation to the "inhabitants ... of some island to the north-ward, whom they called Sybee", ie. Siberut (Crisp 1799:85). It is clear that the usage "Orang Mantawee" applied by the "people of Sumatra" is the one he favours, a designation, however, restricted to the Pagai islands.

Summarising Crisp’s account, as well as drawing on the reports by Marriot, Saul, and Forest published in Dalrymple (1777), Marsden (1811:468) makes the following observation:

Consistent with Crisp’s approach but portraying signs of the uncritical acceptance of the category "Mentawei", and producing a tension between the general and the particular, Von Rosenburg (1853) some forty years on, writes about the "inhabitants" of the "Mentawei" islands. Having briefly situated the islands geographically in relation to Sumatra the article opens with the observation that the "Mentawai-archipel is inhabited by people of all manner of customs and habits specific to the area in which they are found", a strident image replete with diversity. The islands themselves, he relates, consist of firstly "Siberoet", "Sibero" being the word in the "Mentaweijers language" (von Rosenburg 1853:403). To the south of this is "Pora, also South-pora, and Goedfortuin, (Sikobo for the Mentaweijers)". "Pora" is separated by the Nassau strait from Poggij which is variously known as the Nassau islands, or North and South Poggij, or Pake by the Malays, Sigalagan by the Mentaweijers..." (von Rosenburg 1853:404). One of the lesser islands nearby is known as "Poelo Mego" (Mego island) or "Biriloga" in the "Mentaweijers language", and is the most southerly of the group (von Rosenburg 1853:404). As with Marsden, the islands are defined as the abode of the "Mentaweijers", even though they use the term "Tschagallelegat" to "refer to themselves in their own language" and despite the existence of many other indigenous terms ("Sikobo", "Sigalagan", "Mego") indicating various ‘parts’ of the "Mentawei-islands" (von Rosenburg 1853:409). Here the account turns to similarities and differences between the "Tschagallelegat" as a unit rather than the "Mentaweijers"—although the latter usage as the title indicates is dominant—and Australia, Hawaii, and the Marquesas. The rest of the article, however, deals with the various ethnographic vagaries pertaining to the "Mentaweijers", not the "Tschagallelegat, with one reference to the "Seijbie [Saibi] district on Siberut" (von Rosenburg 1853:411) and the occasional reference to what generally happens on Siberut (eg. p.413, p.416), or "some districts in the Nassau islands" (von Rosenburg 1853:422). In effect the account may well have been entitled the "Tschagallelegat" or "Sigalagan", although this would have made it difficult to talk about other areas which one is able to do through a blanket term such as the "Mentaweijers".

Logan’s (1855) "The Chagalelegat, or Mantawe Islanders" is similar in style and scope to Marsden’s except that, in addition to the reports in Dalrymple (1777) and Crisp’s account, there were several more detailed articles published in the intervening 50 years or so that he could draw upon, the most significant of these being von Rosenburg’s. Logan’s summarising, like Marsden’s, is pivotal in this process. Included in Logan’s account is a map of the islands, authored by von Rosenburg, entitled "Chart of the Mantawe Islands". The Pagai islands ("Poggy"), Sipora ("Pora"), and Siberut ("Siberut") are all clearly marked under the umbrella of the "Mantawe islands".

Logan’s treatment of Crisp is also revealing upon whom he draws heavily, a dependence involving a process of selection and refining where the usage "Mentawei/Mentawai" is favoured over and above others also available in these early texts. Crisp, he relates, "visited the Mantawe Islands in 1792" (Logan 1855:274). Of course Crisp himself wrote no such thing, merely mentioning that he had visited the "Nassau or Poggy islands" (Crisp 1799:77). He went on to mention the "Orang Mantawee" as merely one particular (exogenous) appellation for the inhabitants of those islands. Crisp also mentioned that in the Sikakap ("See Cockup") area "are a few houses inhabited by some Malays from Fort Marlbrough" one of whom "had acquired a competent knowledge of the language of the natives" (Crisp 1799:80). However this becomes, in Logan’s words, as follows: "One of these Malays ... had acquired the Mantawe language" during his residence there (Logan 1855:274).

The article is, intriguingly, entitled "The Chagalelegat, or Mantawe Islanders", the former usage coming to Logan from von Rosenburg who himself applied both terms, the "Chagalelegat" to be understood as a subset of the "Mantawe Islanders" in the last instance. An insight into this process is contained within Logan’s appraisal of von Rosenburg’s (1853) account, a poignant testament to the essentialising gaze underwriting ethnographic endeavours at this time as well as displaying a characteristic ironic unwitting self-critique. He praises the detail of the account, in contrast to Crisp’s, as "one of the best accounts that have been given of any of the lesser Indonesian tribes" (note the "Mantawe islanders" as ‘tribe’) although he berates von Rosenburg for not going further than he did (Logan 1855:276). Indeed

Despite this caveat, Logan launches into his account explaining the relation between the "Chagalelegat" and the "Mantawe islands" indicating that the former are the "sole occupants" of the latter, also referred to as the "Mantawe group", so making explicit what was latent in von Rosenburg’s paper and extending it by making the Chagalelegat ubiquitous across the archipelago (Logan 1855:278). The discourse is carried on in a fashion that much discourse about ‘Mentawai’ has—the "archipelago", "group", or "islands" on one hand and "its" inhabitants on the other are collapsed into each other creating a unitary "thing", a "people" who possess or express a "culture" or a particular way of doing things. In Logan’s case in the substantive part of his portrait the "Mantawe islands" are the abode of the "Chagalelegat". However he immediately casts doubt on the homogeneity of the situation by focusing on the differences and complexities of the content of the unit he has constructed. He firstly geographically contextualises the islands in relation to the Sumatran mainland, Enggano to the south, Nias to the North, on to the Andaman islands further to the north in the Bay of Bengal (Logan 1855:279). Then he situates the islands in relation to each other, in effect reproducing von Rosenburg’s mosaic very much in opposition to the essentialising "Mantawe" image:

As with Marsden, Logan seems to desire straightening out the record, to put order into the confusion and complexity. He relates how Veth’s (1849) "account of the Mantaweans ... criticises the errors and confusion in the names that have been applied to them in books and maps" (Logan 1855:275-276), incognizant it would seem of his own role in ironing out, not so much "errors and confusions", but what may perhaps more accurately be described as ensuring that the varying interpretations by outsiders of the inhabitants prevail over the varying interpretations the inhabitants hold of themselves. In Logan’s authoritative hands all four islands became further entrenched as the "Mantawe islands" inhabited by "the Mantaweans" who speak the "Mantawe language", despite his own initial ambivalence as to a suitable appellation. At base we find the coalescing of a unifying, totalising discourse which came to be applied in greater or lesser degrees, a discourse which still operates in the present day.

In Marsden’s hands Crisp’s ideas underwent an initial consolidation with the result that an essentialising gaze was brought to bear on the Poggies, Porah, Nassau and Sybee, in which the basis was laid whereby all four islands could be seen to be the abode of the "Orang Mantawee".(4) With a few strokes of the pen Pagai, Sipora, and Siberut ("Sybee") were stigmatised as being occupied by the one "race of people ... having the appellation of orang mantawei", an ultimately exogenous epithet which appears in accounts over the next 150 odd years as "Mentawai" or "Mentawei", a place inhabited by "the Mentawaians" or "the Mentaweiers". Following Marsden’s synthesising efforts the application of the term "Mentawai" to cover the inhabitants living on the Mentawai islands has become ensconced. Articles about "Mentawai" and "the Mentawaians" are ostensibly about Pagai and Sipora, or recently, Siberut, whilst conversely publications about the Pagai islands concern their inhabitants, "the Mentawaians" (eg. Hansen 1915), although Hinlopen and Severijn (1855) are exceptions, referring to the inhabitants of the "Poggi" islands as "Poggijers" (Hinlopen & Severijn 1855:329;327) or "Poggi-inhabitants" (pp. 324;329), and to the people on Sipora as the "inhabitants of Pora" (p. 330).

Yet, as we have seen, in much of this discourse about "the Mentawaians" there is present an undercurrent of difference that various writers take on board in differing degrees. But since the (implicit) goal is an understanding of "the Mentawaians" as a comparative project, there is no real contextualisation of difference outside the totalising dominant discourse which, whilst not absolute in the 19th century has congealed into a hard inflexible a priori in the 20th.
 
 

Against Essentialism

I wish to pursue the question of homogeneity and variation further in order to investigate in more detail the machinations behind this difference-suppressing discourse that shows up time and again in so many colonial and post-colonial texts, texts which reproduce the essentialist discourse of "the Mentawaians" whilst skirting over great variation in practices and local constructions of identity.

Kruyt’s (1923) "The Mentawaians" is a masterful piece of ethnography when compared with much of what went before it and even some which came after, masterful for its attention to detail and ipso facto the consideration of difference, even though the author relies on usages such as "on Mentawai", "on the Mentawai islands" as well as "the Mentawaians" throughout the article. Kruyt carried out his research in four separate locations: Silaoinan/Taikako, Silabu ("Silaboe"), and Saumangania ("Saoemangania") on north Pagai, and Katurei in southeast Siberut. Interestingly, the major variations are between these three areas on Pagai itself which are all relatively close to each other. In his text Kruyt often remarks on differences between Katurei on Siberut and the villages on north Pagai, differences he puts down to different "adat" (Kruyt 1923:10).

Kruyt reports, for example, that on "Siberut" (ie. Katurei) the rusuk institution is absent(5). Here, instead, a girl’s suitor might make a discreet visit to the field-hut where she is staying, and if she agrees to a liaison then he approaches her parents in order to obtain formal permission to proceed (Kruyt 1923:8). No formal rusuk type arrangement is entered into. In Silabu, in contrast to all his other research sites, Father’s Father’s Brother’s Son’s Child is available for marriage so as to prevent the division of "family wealth" (Kruyt 1923:7), an observation to which Kruyt devotes but a tiny paragraph. On the contrary, from my own research in Madobag, similar to the situation Kruyt describes for Silaoinan, Saumangania, and Katurei, sociological distance is the most important criterion in determining whether or not a marriage is permissible. No one who can in any way be considered to be descended from the same ancestor may marry. Nooy-Palm similarly reports that children of same-sex or opposite sex siblings are not allowed to marry on Pagai or Sipora, or more specifically "the ancestors disapproved of marriages between too close relations" (Nooy-Palm 1968:208). Nooy-Palm, interestingly enough, mentions Kruyt’s research villages including Silabu in drawing on his information concerning funeral practices, but only in passing, and only as support for the general description of Pagai-Sipora "culture" (Nooy-Palm 1968:171;224).(6)

Publishing several years later, Loeb noted that "there are certain linguistic and cultural differences between the three groups of the Mentawai islands; Siberut, Si Pora, and the Pagehs", differences which "become accentuated in North Siberut" due to influences from Nias (Loeb 1928:409). He then goes on to observe that "Little is known ... of the Mentawai islands with the exceptions of the Pagehs" (Loeb 1928:409), even after more than a century of contact. Loeb is uneasy about the usage "Mentawei" in all of his papers. Whom Loeb refers to as the "natives of the Mentawei islands" or the "Mentawei islanders" (Loeb 1928:408) might just as easily been called "the Mentaweians". A rare occasion where he does refer to "the Mentaweians" is in paraphrasing Karny’s (1925) origin theory. Thus in Shaman and Seer he begins: "In my introductory paper on Mentawei (Pageh) culture ..." (1929a:60). Henceforth in this paper he gets around the problem of "the Mentaweians" through recourse to the phrase "In Mentawei" (p.66) or "the people of Mentawei" (p.70). He then alternately refers to the "Pageh islanders" or "Mentawei". In "Mentawei Social Organisation" Loeb articulates his unease thus: "I shall make use of the word ‘‘Mentawei’’ rather than ‘‘Pageh’’ in this article, with the understanding that data mentioned were collected in the Pagehs, rather than throughout the islands as a whole" (Loeb 1928:409)—a troubled victim of the weight of history and scholarly convention. Nevertheless the Mentawei "people" or "islanders" collectively are the focus within his discourse, a literary practice which has become orthodox in the present.

Nooy-Palm (1968) commences her extended portrait exhibiting a tension between the general and the particular. The paper is titled the "culture of the Pagai-islands and Sipora, Mentawei", the information coming from "a short field-study ... on North Pagai, while on Sipora a few additional data were collected" (Nooy-Palm 1968:154), through interviews conducted in Indonesian. It begins with some facts about the "Mentawei-islands" and shortly moves into substantive issues concerning "Mentawei culture" (Nooy-Palm 1968:153). Nooy-Palm makes a pertinent comment about the state of knowledge of "Mentawei" in offering the opinion that "one can hardly refer to the local speech as the Mentawei language, because on these islands no linguistic research has been carried out yet" (Nooy-Palm 1968:159). One wonders why one is more entitled to refer to "Mentawei culture" so unproblematically, since not a great deal more intensive ethnographic ‘cultural’ research, has been carried out on all four islands by the 1990s? With regard to language, Nooy-Palm notes that the "languages spoken on the various islands show differences; moreover, it seems as if several languages (or widely divergent dialects) are spoken on Siberut" (Nooy-Palm 1968:159).

On a slightly different note, in her contribution to Lebar’s (1972) Ethnic Groups of Insular Southeast-Asia, Nooy-Palm divides "The Mentaweians"—who can be variously referred to as "Orang Mantawei", "Mentaweier", or "Poggy-Islanders" covering the major usages introduced by Crisp, Marsden, and von Rosenburg—into

These terms sakalagan, sakalelegat, and sikobou occur variously in the papers of Marsden, von Rosenburg and Logan and would appear to continue to have relevance in the present day. Schefold (1991:18) relates that the "people of Siberut know that located to their south are the islands [Sipora and the Pagai islands] where the Sakalagan live". The word, he speculates, is derived from "laggai", meaning "place of settlement", although "there are those who say this word derives from eilagat, a species of tree". The inhabitants of Sipora "refer to themselves as Sakalelegat, which derives from the word lelegat meaning "place", or Sakobou—from the word kobou ‘source of salt-water’". Along with this the "inhabitants of both the Pagai islands ... refer to themselves as the Sakalagan people [whereas] all three of the southern islands refer to Siberut’s inhabitants as the Siberut people ... named after a river in the island’s southeast" (Schefold 1991:18). Coronese, somewhat confusedly and without making clear his sources relates that "The Mentawaians themselves refer to the Siberut people by the name Sakalelegat and Pagai by the name Sakalagan (Coronese 1986:10).

But it is in Schefold’s (1972-73,1982,1986,1989,1991,1994) work that, in my view the inchoate nature of the discourse operating to reproduce the "Mentawaians" comes into relief. Schefold carried out fieldwork with the Sakuddei, a group located in the south western corner of Siberut who, it is clear, are "Mentawaians" first and "the Sakuddei" second, the latter being a subset of the former. In "Culinary Code", for example, we learn that "the Mentawaians" live "on Mentawai" or the "Mentawai islands" and are by dint of this fact an example of the "culture on the Mentawai islands" (Schefold 1982:68). This is the characteristic structure underlying all Schefold’s works despite the existence, once again, of a very muted but nonetheless quite evident undercurrent of difference. For example similar to Kruyt, as well as highlighting this disparity between the Pagai islands/Sipora and Siberut, Schefold notes the major difference between the set of marriageable kin on Siberut and Sipora where

Similar to Loeb, there is unease—although it is much more implicit—regarding the assumed generality implied by the category ‘Mentawai’ evident in "The unequal brothers-in-law..." (1986). As Schefold remarks, the "data to be presented in the following paragraphs all derive from Siberut ... and differ in some respects from the situation on ... North and South Pagai" (Schefold 1986:73). Indeed on Pagai, as opposed to Siberut, women inherit taro fields due, Nooy-Palm (1968:213f) has argued, to a double-descent ideology. This Schefold rejects, maintaining that in "reality" women’s inheritance of taro fields can be read as a "specific instance of the general bilateral extension of the right of succession in Pagai, which is connected with the division of labour and has no structural significance" (Schefold 1986:78). This is because, when a woman’s brother marries, she gives a "suitable proportion" of her taro fields to her BW. This assimilating, homogenising approach finds its rationale in the paper’s general objective namely to connect "Mentawai" to the Indonesian "structural core".

In reconsidering J.P.B. de Josselin de Jong’s (1977[1935]) structural criteria upon which is based the unity of "Indonesia" as a "field of ethnological study", that is asymmetric circulating connubium, double (bilineal) descent, and socio-cosmic dualism, Schefold argues that the circumstance of the "floating alliances of the Mentawaians"—referring to a new husband’s subordinate relationship to his wife’s relatives, that is his group’s wife-givers—can be considered to be a "transformation of the established connubium in the asymmetric cross-cousin marriage in less archaic Indonesian cultures" since

This theme can also be seen as the basis for Schefold’s "involution" argument where, adapting Geertz’s (1963) Agricultural Involution thesis from the agricultural to the social realm, Schefold argues that the (Mentawaian) "taboo system" becomes more and more elaborate, ever more differentiated to the point where, in a novel move, "devolution" occurs, that is, a simplification of an area of complexity that has become too rigid. There is never "evolution", defined as the "fundamental change of religious ideas" (Schefold 1972-73:72), simply due to the "force of tradition" (Schefold 1972-73:80). From this point of view, then, the fact that "Within the Mentawaian archipelago there exist from place to place significant differences in the details of religious observances" is not a problem as these are differences in degree and not kind, as there has never been, nor is there likely to be, fundamental, evolutionary change: "The basic concepts ... are ubiquitous; concrete manifestations, however, differ from each other persistently" (Schefold 1972-73:79). ‘Mentawai’ as a structural whole internally involutes and devolutes ad infinitum, a result of the "force of tradition"—"communal belief in the validity of inherited ideas restricts all changes to within prescribed causes" (Schefold 1972-73:76). Involutionary and devolutionary moments

Detectable within this theory’s machinations is the metonymic power it draws as a part within the total metaphor qua structure that is ‘Mentawai’. In a sense the argument is that the "culture" on the Mentawai islands also has a structural core, specifically, through the equation of Siberut and Pagai as transformations of each other, and collectively, as a subordinate transformation of the dominant "Indonesian" core. But it is through this that it finds its meaning—‘Mentawai’ metonymically evokes the structural whole. Yet on its own, the power of the concept, ‘Mentawai, performs this function quite well whereby all differences over the islands are fundamentally ‘Mentawaian’, although nowhere has anyone actually explicitly articulated what the essence of ‘Mentawai’ might consist in. It remains an unarticulated assumption embedded in, and thereby shaping, the discourse. The argument could be made, quite apart from the structural argument, that differences are indeed only superficial and therefore inconsequential, making a critique of ‘Mentawai’ irrelevant. But in the absence of a wide range of intensive, specific studies from a wide, representative variety of areas, supplying an up-to-date ethnographic ‘snapshot’, this would be premature. More urgently required are studies valorising indigenous conceptions of identity, explanations of who particular groups consider themselves to be—who are the Selves and who are the Others? In present and past discourse, meaningful difference is silenced, enduring paradoxically only as ‘noise’, as a sub(verting)-text in which the core discourse’s integrity is retained. Beguiled by its own magic, the discourse has, understandably, not seen any need to question its self-defining self-evident epistemological and ontological status.

A closer look at de Josselin de Jong’s ‘ethnologisch studieveld’ approach enables us to catch a glimpse at some of the forces at work shaping the discourse. De Josselin de Jong defines his ‘studieveld’ as follows: "By this term we mean certain areas of the earth's surface with a population whose culture appears to be sufficiently homogenous and unique to form a separate object of ethnological study and at the same time apparently reveals sufficient local shades of differences to make internal comparative research worthwhile" (167-168), a practice securely grounded in "objective evaluation according to established clearly defined rules" (167). The "studieveld" that the author has in mind here is, of course, "Indonesia" or the "Netherlands East Indies". Despite the paradoxical acknowledgment that this area "is by no means homogenous; that the only thing knitting together the many parts of this archipelago, which is as heterogenous as can be with regard to race, language and culture, is the authority of the mother country..."(168), our attention is nevertheless drawn to "a few phenomena which shed light on the significance of the Malay archipelago as an ethnological field of study [in order to] reveal at least something of the unity which makes the diversity all the more instructive and interesting"(168). In the blink of an eye these "phenomena" congeal from shadowy heuristic suggestions into what can be considered an "integrated whole or system". And the further we look into this system "the clearer it shows itself to be the structural core of numerous ancient Indonesian cultures in many parts of the Archipelago". That is, as Schefold indicated "clan connubium" (168), "socio-cosmic dualism" and the "double descent system" (171). Of note here is the dialectic of authority where the "field" as structure is self referential—it draws authority from that which it purports to discover as an instance of an "objective evaluation according to established clearly defined rules". These in turn recursively support the initial premises with their authority. The "few phenomena" are the de facto dominant discourse which subordinates diversity within the unity that is the (dominant) "whole or system". Differences are in fact to be seen as "different local variants of the system" (174). Similar to Orientalist methodology "It shares with magic and with mythology the self-containing, self-reinforcing character of a closed system, in which objects are what they are because they are what they are, for once, for all time, for ontological reasons that no empirical matter can either dislodge or alter" (Said 1978:72). It is also not surprising that de Josselin de Jong, whose paper provides Schefold with his inspiration, in seeking to persuade his readers of the worthiness of his "ethnologisch studieveld", points out that "conclusions" arrived through a study of the "phenomena" outlined in his paper would be useful for "the administration, for the missions, and for all those who occupy a leading position in Netherlands East Indies society" (de Josselin de Jong 1977[1935]:181). The scientific gaze and the colonial gaze reveal their common grounds in a power which subjugates: the former in its Platonic quest to reveal the primary structures of "reality" beneath the messiness of that "reality", the latter in its quest to similarly grasp its essence in order to, simply, be better able to craft it, an exercise of power in a movement from understanding to management.
 
 

Where does ‘Mentawai’ begin and where does ‘it’ end? What essential invariable characteristics does ‘it’ display, or is ‘it’ a "fuzzy set" akin to the New Guinea Highlands (Hays 1993)? ‘Mentawai’ is surely finite, unambiguously so when taken as a purely geographic category—‘it’ ends somewhere. But this is a matter of intensive empirical investigation, not a matter of unreflexive a priori givens. In the preceding examples of variation is a snippet of information providing a glimpse of what may well be a whole body of meaning in itself. Then again it may not. The answer is probably somewhere in between. But certainly, in the comparative perspective these glimpses are given short shrift, subordinated and silenced in the pursuit of ‘Mentawai’ as a quasi ‘culture-area’. Moreover the crucially important issue of indigenous representations of identity is given no space at all.

The general approach hitherto has been to conduct research, in greater or lesser degrees of intensity or duration in (a)generally discreet location(s), research about the ‘Mentawaians’ which legitimates the inclusion of selective examples from separate locations sometimes some distance removed from the research site in order to highlight that they do things differently there. But these are differences in degree not of kind, arguably similar to the situation in Fiji where the colonial construction of a "Fiji customary order ... glossed over considerable internal diversity in kinship, ritual, social structures, and language ..." (Thomas 1992:69).

The essentialisation reproducing ‘the Mentawaians’ is sustained through this broad, vague comparative focus which is sustained itself by implicit recourse to the category. The essence of this gaze is succinctly articulated by Wagner where he writes

People of the Rereiket—indigenous conceptions of identity

Consequently my research has involved not the ‘Mentawaians’ per se but rather a select series of events occurring within the community, Madobag, which is a part of one of numerous river valleys located across Siberut known as the Rereiket. From the point of view of the anthropological gaze(8) the inhabitants of the Rereiket can be understood as having similar sorts of practices that are directed towards similar sorts of ends, people who in certain contexts see themselves as Rereiket people rather than Sagalubbe people or Silaoinan(9) people, Madobag people rather than Ugai(10) people in other contexts and, significantly, people of a certain suku (‘descent group’) as distinct from all other suku, in many other contexts. I would contend that this is the point at which a consideration of the complexities of identity in this particular ethnographic context should begin.

Hence this approach stems from the concerns of the people of Madobag and the multiple facets of their complex identity that are evident as a manifestation of those concerns. It seeks to move towards embracing the situated and contextual, constructed nature of identity that takes place in the flux of lived reality in this part of the world. It is perhaps best described, however, not in terms of 'identity' which, as Steedly (1996) cogently demonstrates in the case of Karo identity in North Sumatra, is an artefact of the collision of colonialism, political-economy, and historical consciousness, a codified, delimited range of possibilities distilled from the "bewildering mutability of forms and substances" that make up the object of inquiry. The more relevant concept is Hall's "idenfication" in which identities are conceived as "never singular, but multiply constructed across different, often intersecting and antagonistic, discourses, practices, and positions" (Hall 1996:4).(11) In this mode of thinking, every encounter, whilst firmly situated in the contingencies and immediacies of the time-space context in which it is taking place, is nonetheless also firmly interconnected with non-localised material realities involving the state and the world-system. This has a bearing, then, upon how the anthropologist—an exemplary product of and manifestation of the world-system—constructs the practices to which he or she is privy. As Steedly remarks the "categories we recognise and the content of those categories are constantly subject to processes of internal and external revision" (1996:464). Given this the anthropological task is to trace the outlines of the processes of identification that constitute the historical moment in which the anthropologist is embedded.

On a general level, then, the people of Madobag construct themselves as "Indonesians" since they have a definite relation to the entity "Indonesia", this relation varying amongst people according to the length and quality of contact with "Indonesians". They have, to varying degrees, an awareness of Indonesia as an entity among other identities in a geo-political context, although for those who have not had secondary education this is in terms of one pulaggajat(12) amongst other pulaggajat. Hence, younger people and those educated in the schools in Muara Siberut, the administrative and trading centre for South Siberut, in certain contexts refer to themselves unequivocally as "Indonesians" using the national language (Bahasa Indonesia) to articulate this identity: we are/I am "orang Indonesia". They construct themselves as such when distinguishing themselves as members of the entity "Indonesia" as distinct from the various other entities, Jerman (Germany) and Inggiri (England) for example.

For those who do not obviously originate from Siberut and use the national language to communicate with them in the first instance, the identification invoked is sareu which has to be considered in relation to the opposition between simatawe and (sa)sareu, a distinction placed at the centre of several accounts of indigenous conceptions of identity (Naim 1977; Coronese 1986). This dichotomous conceptualisation represents an advance on the way in which identity on the islands has been dealt within the context of the discourse of the ‘Mentawaians’, although these categories are still conceptualised as discrete generalised identities rather than being viewed as contingencies within the play of identification.

The existence of the word simatawe (cf. ‘Mentawai’) arguably represents a case of what Thomas (1992) has described in relation to Fiji as substantivisation, where colonial discourse led to the selection, reification, and institutionalisation of a specific cultural order consisting of certain practices selected from a possible variety of practices. Or in terms of identity formation the unity encapsulated within the concept 'Mentawai' has been constructed within the machinations of power, of inclusion, and importantly, exclusion: it is a result " not of a natural and inevitable or primordial totality but of the naturalised, overdetermined process of 'closure' "(Hall 1996:5). The machinations of closure in this case have resulted in the acceptance of an exogenous reification which is reproduced as a living reification in the present by the very people who were its original subject in the term simatawe. The evidence suggests that "matawe" is traceable to the usage taken by Logan from Crisp, a word that has come to be applied in colonial discourse as ‘Mentawai/Mentawei’ and has eventually come to be used by the people themselves to refer to their collective identity in certain contexts. How that might have actually occurred is unclear, although it would be short of the mark to suggest that this is purely a product of literary convention. It is more likely that having been forged within the functioning of the political economy and the colonial gaze in the early 19th century, it became ensconced with the various discourses that constituted the growing entanglement of the agents of colonialism with indigenous populations.(13)

In the Rereiket, then, simatawe refers to anyone who can be defined as originating from one of the Mentawai islands who can speak the "Mentawai language" (nganga (si)matawe). It is part of a strategy to identify broad categories of "people" (sirimanua): simatawe, Minangkabau, Nias(ians), Batak people, Jerman (Germans), Inggiri (English) and so forth. However, the word simatawe is hardly ever used in this positive sense, except when actively elicited from informants. When the word is used in this way it is by non-locals particularly foreigners, Indonesian travellers and state officials. For the indigenous population it is, rather, a case of whether an individual is sareu or non-sareu, not whether they are simatawe or not.

Sareu ("those from afar") embraces other Indonesians or anyone simakotkot tubu ("dark/black-bodied") who cannot be identified as originating from river valley areas on Siberut or the other islands, Sipora or North and South Pagai, whose first language is not nganga matawe, which includes (white) foreigners. These foreigners are referred to as orag turi ("tourists"), since they almost universally appear with backpacks, cameras and guides, these constituting the defining elements of an orag (Indonesian: orang) turi. Other Indonesians are never usually referred to as ‘orag Indonesia’ but rather simply as sareu. It should also be noted that this facet of identity is only relevant in contexts where sareu may or may not be involved, the language criteria being of primary importance here. It is otherwise of no consequence. And I surmise that prior to colonial entanglement it would have been a case of ‘our suku’(suku mai), that is, our descent group against the rest, sirimanua, a subset of whom would have been sareu, the designation simatawe not being a part of the field of possibilities.

An important point in all this is that seemingly discrete identity categories particularly the simatawe-sareu distinction are, in the play of identification , poignantly indiscrete—they often run counter to the binary logic the observer brings to them. Seeming contradictions from the objective, discursive, ethnographic standpoint that deals with identities rather than identification, are not contradictory when contextualised within the play of identification. For example one older informant highlighted the complexities at work here within the processes of identification , in declaring himself to be Indonesian yet not be sareu which encapsulates things Indonesian:

On a day to day, practical level, then, in opposition to this formal, discursive distinction between sareu and simatawe, the most important differentiation is that constructed between each suku (‘decent’ group) and all the rest, one practically reproduced, or rather lived, in a multitude of contexts and is at the heart of quotidian life.(15) Thus there is "us" (kai, or more usually, suku mai, "our suku") and "them" (sirimanua) the broad undifferentiated mass of humanity (within which is included sareu and orag turi). Sirimanua can be also differentiated according to river region in the relevant context. Important places in the landscape of Siberut are identified with reference to the river which passes through them. Each of the major river systems gives its name to the valleys through which it and its tributaries flow (cf.Schefold 1991:121). The people of Madobag along with those in the nearby settlements of Rokdok, Ugai, and Matotonan define themselves as "from the Rereiket", the name given to the river passing through the valley in which they are located. The "rereiket" exists in opposition to other nearby regions such as Silaoinan, Sagalubbe, as well as those more distant, such as Saibi, Simatalu, Samukob, Terekan to name some of the regions more frequently referred to in local discourse. Sirimanua can be conceived also to consist of all the other suku whether they be Sabetiliakek (a suku in Ugai), the suku Satoinong of Silaoinan, the pulaggajat of Padang, Minangkabau, Jerman (Germany), or Inggiri (Britain) and so on. It all depends upon the context—this determines which identities will be invoked, lived, and reproduced in the ongoing play of identification.

In summary the somewhat simplistic formal dichotomy between simatawe and sareu, relied on by several of those working within the discourse of the ‘Mentawaians’ (eg. Naim 1977; Coronese 1986), does violence to the lived, practical reality of identity as it unfolds within the processes of identification in the Madobag context. These represent merely two identifications that can be invoked from a range of possibilities. As for ‘Mentawai’ it suffices to say that the category has a relevance limited to certain contexts where it is used to evoke the collectivity of islands directly adjacent to the province of West Sumatra, and thereby metonymically encapsulate the inhabitants of these islands.

In this paper I have sketched the outlines of a hegemonic discursive construction of the inhabitants of the Mentawai islands, encapsulated and exemplified in the "field of study" approach which has formed the foundation and justification for the ethnographic modus operandi in this region. As a move away from this I have cast critical light upon this discourse through which the complex construction of sociality on these islands is seen but though a glass darkly. As an example of a possible direction in which the ethnographic study of the area could proceed I have, furthermore, briefly considered the broad contours of a more complex identity than the discourse of the ‘Mentawaians’ allows. My aim in this is to open a space within which the complex nature of indigenous identity can begin to be explored as an important phenomenon in itself in which greater analytical emphasis is placed on the specifics of particular constructions of identity at particular times in particular places. Minimally I would suggest that a sustained consideration of the specificities of identity formation be a sine-qua-non element of any ethnographic endeavour wherever it might be carried out in the area in the future. Thus this a beginning not an end; it represents a point of departure in which the nature of the ethnographic enterprise can be more fully explored and a greater degree of reflexivity achieved in relation to the art of ethnographic research as it has been practised in the region than has hitherto been the case.
 

Acknowledgements

This paper has been previously published (in a slightly modified form) in  The Australian Jounal of Anthropology Vol. , No.1 (1999). It appears here with permission of the Australian Anthropological Society.

Endnotes

(1) A sample of these might include Barth  (1989); Kottak & Colson (1994); Moore, S.F. (1994); Vayda (1990); Hannerz (1992); Douglas (1996); Marcus & Fischer (1995).

(2) For example Loeb (1928; 1929a; 1929b; 1929), Nooy-Palm (1968; 1972)

(3) For example Schefold (1972-73; 1982; 1986; 1989;1991; 1994).

(4) Kipp (1993:24-65) points out that a similar process was implicated in the construction of ‘Batak’ as a distinct entity and, based on this, the construction of the Karo Batak as a subset of the larger ethnic group.

(5) On Pagai a man and a woman could commence living together and have children without marriage taking place. The institution and the special dwelling erected for them were referred to as rusuk.

(6) There exist equally significant variations in relation to death practices. On "Siberut" a corpse is cleaned in the house, but on Pagai this occurs at the cemetery (Kruyt 1923:167). This is especially significant in light of the fundamental cosmological import of "death" and the role of the House in distancing the deceased’s spirit (sanitu), which has now become the very essence of death, from the rest of the House’s members. At the least it suggests a different cosmology. Another issue is the conveying of the corpse to a canoe at the river which takes it to the cemetery. On Pagai a completely new path to the river is cut, whereas on Siberut the existing path is used. In addition the man and the woman who take the canoe are, on Siberut, husband and wife. But on Pagai and Sipora they are not yet married (Kruyt 1923:169). What differential valency, then, is being attached to these activities? Perhaps on Pagai the association of death is too powerful and must immediately be separated from the living, hence the rapid removal to the graveyard via a new track which is not a common conduit for the living, to be finally escorted to its final resting place by people who may not have children (although they may if they are living in rusuk together), that is who are not yet nurturers of life. This all points back to the role of the House and its differential relation to ‘death’ in both circumstances. Furthermore, in Taikako a corpse is placed with the head facing in any direction. In the other areas the head may only face the east (Kruyt 1923:174; Mess 1870:359). Head and direction are very significant in some other societies within the Austronesian area and beyond. The lack of importance of this for one group, when all the groups around it place great importance on it, is intriguing.

(7) Nooy-Palm had mentioned these in the (1968) article, noting that the language in use on Sipora "ie. the Sakobou or Sakalelegan (Sakalelegat), ... only in respect of a few words differs from the dialect or language spoken by the Sakalagan. The Sakalagan (=inhabitants of the village) live in North- and South-Pagai ..." (Nooy-Palm 1968:159).

(8) That is, this particular anthropological gaze.

(9) River valleys located to the immediate west and north respectively.

(10) A settlement located 2 km to the north.

(11) The discourse within which Hall is working attributes this process to contemporary cultural, political and economic processes constituting "globalization". However this approach perhaps unwittingly situates itself within an essentialist position by downplaying the dynamism that has arguably characterised the non-European world (the "people without history" as it were) that did not require European stimulation to animate it.

(12) An area of land (laggai) belonging to a particular descent group. It is used by the locals to articulate the fundamental sociological categories which make up their social landscape.

(13) And as Keane points out (drawing upon Voloshinov, 1973) "the ideological trappings of language are powerful precisely because of the ‘‘social ubiquity’’ of speech, which is at once public and intimate" (Keane 1997:38).

(14) A large body of literature has emerged in the past two decades highlighting the importance of the space of the ethnographic encounter constructed between Self and Other in the research process. For example Bachnik (1986); Bourdieu (1977); Clifford (1983); Dumont (1986); Dwyer (1977;1979;1982); Geertz (1973); Taussig (1989); Tedlock (1979); Tyler (1987a;1987b); Wagner (1981). The problem is identified thus by Bourdieu: "The anthropologist's particular relation to the object of his [sic] study contains the makings of a theoretical distortion inasmuch as an observer, excluded from the real play of social activities by the fact that he has no place (except by choice or by way of a game) in the system observed and has no need to make a place for himself there, inclines him to a hermeneutic representation of practices, leading him to reduce all social relations to communicative relations, and more precisely, to decoding operations".

(15) In a forthcoming paper "Places of Life; Places of Death: Towards a Phenomenology of the Austronesian House" I examine the way in which particular identities are actively reporduced through the reporduction of the House as  a "strategy of intervention" in the cosmos, the aim being to reproduce a haven for "life" in opposition to the forces of death that threaten to overwhelm the living. The production of a particular "place", the House, is effected through the production of identities within the processes of identification.