Vous êtes sur la page 1sur 427


Monograph Series
Vern Carroll, Editor
ASAO Monograph No. 1
[?d HONOLULU 1970
. .
Edited by Vern Carroll
Editor's Preface vii
I. Introduction: What Does "Adoption" Mean? 3
Vern Carroll
2. Traditional and Modern Adoption Patterns in Hawaii 21
Alan Howard, Robert H. Heighton, Jr., Cathie E. Jordan,
Ronald G. Gallimore
3. Adoption in the Society Islands 52
Antony Hooper
4. Tahitian Adoption as a Psychological Message 71
Robert I. Levy
5. Adoption on Rangiroa Atoll, Tuamotu Archipelago 88
Paul Ottino
6. Adoption on Nukuoro 121
Vern Ca"oll
7. Adoption on Kapingamarangi 158
Michael D. Lieber
8. Banahan Adoption 209
Martin Gary Silverman
9. Some Legal Aspects of Gilbertese Adoption 236
Henry P. Lundsgaarde
l 0. Adoption, Guardianship, and Social Stratification in
the Northern Gilbert Islands 261
Bernd Lambert
I I. Adoption on Ponape 292
J. L. Fischer
12. Adoption on Romonum, Truk 314
Ruth Gallagher Goodenough
13. Adoption on Rotuma 343
Alan Howard
14. Kinship and Adoption in the Northern New Hebrides 369
H. W. Scheffler
15. Epilogue: Transactions in Parenthood 391
Ward H. Goodenough
Contributors 423
A modest literature for Eastern Oceania helped to define a whole
new field of anthropological effort. The pioneering works of
Bronislaw Malinowski, Raymond Firth, and Reo Fortune soon
became classics in the nascent discipline of social anthropology.
But the early significance of Eastern Oceania (Polynesia, Microne-
sia, and island Melanesia) did not endure. A world depression, a
global war, and a shift of interest to areas of greater political con-
cern-each in its turn prevented more than token advances beyond
the early work. While Eastern Oceania languished, social anthro-
pology in other areas (such as Africa) proceeded apace. As late as
the mid-l 950s, scarcely more than a handful of social anthropolo-
gists could be found working, at any one time, in this vast region.
Although social anthropology, by definition, is comparative, a
paucity of primary data has retarded the development of a social
anthropology of Eastern Oceania. Such data are now becoming
available. Increased availability of research funds and improved
transportation and communication, among other factors, have
lured ever-increasing numbers of students into this geographical
area. Their enterprise has been well rewarded. Eastern Oceania is a
fascinating laboratory, well stocked with small societies isolated
from each other by miles of open ocean.
A score of junior scholars have served an extended apprentice-
ship in this laboratory and are now prepared to confront problems
of common concern. The contributors to this volume have re-
cently formed an organization, the Association for Social Anthro-
pology in Oceania, whose primary function is to sponsor symposia
on topics of comparative interest and to publish the reports of
these symposia. The present volume is the first of what we hope
will be a long series.
As with most initial efforts, the present work has been in the
making a long time. In response to a proposal circulated by the
editor in the summer of 1964, John L. Fischer, Antony Hooper,
Paul Kay, Bernd Lambert, Henry P. Lundsgaarde, and Harold W.
Scheffler prepared papers which were presented at a symposium at
the annual meetings of the American Anthropological Association,
in November, 1964. The symposium, on Adoption in Eastern
Oceania, was chaired by David M. Schneider; Ward H. Good-
enough served as principal discussant. During the ensuing year and
a half, the other contributors to this volume-many of whom were
in the field-prepared preliminary drafts of papers. These papers
were read and discussed at a conference, chaired by David M.
Schneider, held on March 17 and 18, 1967 at Stevenson College,
University of California, Santa Cruz. There was general agreement
that the conference had been of great value in clarifying the issues.
The participants in that conference would like to acknowledge
their gratitude to Dean E. McHenry, Chancellor of the University
of California, Santa Cruz, and to Charles H. Page, Provost of
Stevenson College, for their financial sponsorship of. the confer-
ence and their warm hospitality. We are also indebted to Roger M.
Keesing, who ably attended to the local arrangements. Following
the conference, each of the participants prepared a thoroughly
revised paper. Further revision followed the circulation of these
papers among the symposium participants. A spirit of cooperation
has characterized this project since its inception; the same spirit
has inestimably lightened the labor of the editor, making that
often tiresome task a rewarding experience.
It is a pleasure to acknowledge our gratitude to the Wenner-
Gren Foundation for Anthropological Research for a grant which
aided in the preparation of the manuscript for publication.
Pacific and Asian Linguistics Institute
University of Hawaii
May, 1968
Vern Carroll
Single quotes are used consistently throughout this work for
glosses of native words, concepts, and conventional phrases.
A dashed line indicates an adoptive relationship; the arrow points
to the adoptee:
.... , -a-do-p-te-r 1 ~ adoptee I
Vern Carrol/
An important part of the recent development of anthropological
theory in English-speaking countries has been the continuing ef-
fort to use ordinary English words in ethnographic description
without allowing the cultural connotations of these words to dis-
tort the analysis which makes use of them. However heroic the
effort, any anthropological use of a term like "marriage" or
"family" or "kinship" seems destined to prejudge the central
issues by introducing unspecified assumptions derived from the
anthropologist's culture. No sooner does an anthropologist provide
a working definition of a crucial term than other anthropologists
discover an ethnocentric bias.
On the other hand, anthropologists who have worked in Eastern
Oceania have all confidently identified something to which the
term "adoption" seems applicable. Although the contributors to
this symposium did not formally agree upon a working definition
of adoption, it seems to me that we had in mind something like
the following:
Adoption: any customary and optional procedure for taking as one's own a
child of other parents.
There are several difficulties, as we shall see, in refining such a
definition to apply fully to the cases which have been selected as
generally conformable to it. The first of these is the matter of con-
tent. Is adoption in Eastern Oceania like adoption elsewhere-in
the United States, for example-in certain specifiable respects, and
unlike adoption elsewhere in specifiable respects? Is it judicious, in
other words, for anthropologists to use this term to label customs
that conform only in part to adoption practices in Europe and the
United States? This question can be usefully discussed by com-
paring the forms of adoption in the United States and in Eastern
For those of us brought up in the United States
-and here I am
relying exclusively upon myself as an informant-"adoption" calls
to mind the picture of a couple, who have tried unsuccessfully for
many years to have children of their own, who finally, with con-
siderable misgivings, have secured a child of unknown parentage
from an institutional intermediary, called an "adoption agency."
The child is normally presumed to have been unwanted by his
natural parents; he is probably the child of an unwed mother.
Usually great efforts are made to keep the natural parents and the
adoptive parents from knowing each other-lest the former try at
some future time to claim their child from the latter. The whole
business is attended by legal proceedings which are so complicated
and so time-consuming that they often serve to inhibit potential
adopters from realizing their aim.
Adopters are normally required to be young, healthy, happily
married couples of substantial social position and income,
though one hears that there has been some liberalization of the
rules recently, so that mature single persons are occasionally per-
mitted to adopt children. Prospective adoptive parents appear to
have quite definite notions about what sort of child they want.
Children with chronic illness or deformity, personality problems,
or minority ethnic status are notoriously difficult to "place." Our
culture presumes the altruism of an adoptive parent, especially
since the adoptive relationship is generally known to be frought
with difficulties. Parents who voluntarily give up their children in
adoption are not regarded highly, although it is recognized that
the parents (especially the mother) may have had little real choice,
given the difficulty of their circumstances.
The above view of American adoption is deficient in several re-
spects. There are many cases of adoption by persons who are not
strangers to the adoptee. About half of the adoptions in America
are by relatives-the adoption of illegitimate children by their
grandparents, the adoption of stepchildren, and the adoption of
children of deceased or divorced relatives.
Even when allowance
is made, however, for such cases, there are striking differences be-
tween adoption in the United States and in Eastern Oceania.
Whereas American adoption is often a transaction involving
total strangers, adoption in Oceania is generally a transaction be-
tween close relatives. In the typical case, the adopter is related to
one of the natural parents of his or her adopted child as a full or
classificatory 'sibling' or 'parent'. Whereas American children are
generally put up for adoption by their natural parents, Oceanic
parents rarely "put up" their children for adoption. Unlike Ameri-
can adoption, the traditional Oceanic variety is not characterized
by formal legal procedures. American children are generally
adopted by couples, whereas Oceanic children are often adopted
by a single individual (albeit, more frequently than not, one who is
married). American adopters are most often childless,
adopters in Oceania frequently have children of their own. In
some Oceanic societies adoption is much more frequent than in
the United States, and Oceanic adopters seem much less choosy
about the physical attributes of the child they get. For example, in
Oceania a child may be adopted even before it is born.
The social
attributes of the child's natural parents, however, are generally of
great importance to Oceanic adopters.
Prospective adopters in Oceania are rarely denied the privilege
of adoption because of economic circumstance, ill-health, or per-
sonality characteristics. They normally get no kudos for adopting
another's child, while natural parents who give up their children in
adoption normally are not stigmatized; on the contrary, they are
usually considered generous. The natural parents of American chil-
dren who have been adopted are unwilling or unable to assume
parental responsibility. In Oceania the natural parents of adopted
children are usually ready, willing, and able to keep their children.
In the United States, the defining feature of adoption appears
to be the legal act. (Note the following dictionary definition:
"Adopt: ... to take as one's .own child, specifically by a formal
legal act .. . "[American College Dictionary 1964] .)
There appear to be several reasons for an increasing emphasis in
the United States on court-approved adoptions.
One important
factor in the growing involvement of the state in adoptions is con-
cern for the child's welfare:
The child is imbued with inherent rights entitling him to what is essential for
his well-being. He should be provided the conditions, experiences and oppor-
tunities favorable to healthy growth, and to the development, use and enjoy-
ment of his individual capacities (Child Welfare League of America 1958:4).
Another factor of relevance is that "A changing concept of the
family as a unit based on ties of affection and actual performance
of the parental role, not only on biological relationship, makes it
possible for children and adoptive parents to enjoy the satisfac-
tions of family life" (Child Welfare League of America 1958:2).
The legal proceedings in adoption are designed to clarify the
adopted child's status. The adopted child must first be decisively
removed from the authority of his natural parents, in order to pre-
vent the complications which would arise if the natural parents
were later to claim their rights in the child (Child Welfare League
of America 1958:61). Second, the responsibility for the child's
welfare (care, feeding, protection, etc.) must be assured (Child
Welfare League of America 1958:64). Adoption is considered a
drastic remedy: "It is a generally accepted premise that a child is
best reared in the home of his natural parents ... " (American
Academy of Pediatrics 1959: l ). Only under certain conditions is
adoption considered advantageous to the child: "Adoption should
not be considered for ... children with close family ties, whose
own relatives can, or may be helped to meet their needs" (Child
Welfare League of America 1958:6).
In Oceania, the interests of the parents, rather than those of the
child, are considered paramount. There appears to be little effort
to sever legal and social relations between a child and his natural
parents. Adoption in Eastern Oceania tends toward an informal
arrangement between the parties most immediately concerned.
Oceanic adopters usually do not trouble to register adoptions with
local courts operating according to European procedures, even
when such courts are easily accessible to them, except when adop-
tive status must be determined for land cases (see Lundsgaarde,
chapter 9).
The definition of American adoption in legalistic terms results
in the grouping together of a number of distinct cases which have
little in common apart from their manner of legal validation. Each
of these categories of American adoption would be considered
somewhat differently by Oceanic peoples.
In most of the societies examined in this volume, the natives
clearly distinguish, terminologically and conceptually, between
"fosterage" (temporarily taking care of others' children as an
obligation of kinship) and "adoption" (permanently assuming the
major responsibilities of natural parents).
The sort of American
adoption in which a couple takes an orphaned young relative
would in Eastern Oceania usually be considered .. fosterage." If the
foster parents did not ask for the child from his natural parents,
then the child-from the natives' point of view-could not possibly
have been "adopted."
In Oceania there is little, if any, adoption of total strangers.
Moreover it does not appear to be important to Oceanic step-
parents to "adopt" their stepchildren. Thus there is little substan-
tive comparison which is possible between the adoption complex
in Eastern Oceania and that in the United States, when we con-
sider the two adoption institutions in toto. Even when we com-
pare only specific parts of these institutions we find few points of
similarity. The characteristics of adopters and adoptees, the moti-
vations of adopters and the natural parents from whom they
adopt, the social form of the process of adoption, and the cultural
meaning of the act-all differ significantly between Eastern Oceania
and the United States.
Although it might have been expected that the research re-
ported upon in this volume would have led to a definition of the
term "adoption" which provides a clear indication of what is and
what is not to be classified under this heading, it will come as no
surprise to those who have followed the recent terminological con-
troversies in the anthropological literature that no definition of
this sort has emerged. It would appear in fact that substantive defi-
nitions are not possible at this time.
The most careful efforts to provide "ethnographic" definitions
of customary practices in terms of their "essential features" or in
terms of a "lowest common denominator" of attributes have suc-
ceeded only in showing that it is a hopeless task (see, for example,
Leach 1955 and Gough 1959 on the definition of "marriage"). As
Levi-Strauss has demonstrated ( 1963 ), any such effort must con-
tend with a wide range of intergrading phenomena, and will lead
inescapably to the conclusion that the custom, in fact, does not
exist! What is always required in efforts of this kind is an analyti-
cal reference point, a coherent motivation for assembling data rele-
vant to a particular topic. For adoption, such a reference point can
be provided only by a general theory of kinship.
Although it is not my intention to provide such a theory here, it
might be useful to provide an illustration of the sort of thing that
is needed. Let us propose, for example, the theorem that any cul-
ture must reinforce certain values concerning the care of one's
own children (lest children be regularly abandoned) and may well
reinforce these values to such a degree that adults without children
feel deprived. One obvious way of insuring a desire for children is
by equating the liberated status of "adulthood" with "parent-
hood." (Note that a married pair in the United States is merely a
"couple," whereas a couple with children is elevated to the status
of a "family.") If adults without "own children" feel the lack of
them, then it is not remarkable that some customary means be
provided for making "own children" available to such persons. It
was with this point of reference in mind that I constructed the
definition on page 3 . A definition of this sort may have some
heuristic value in delineating a body of data which may repay fur-
ther analysis. For example, this definition should suggest to the
ethnographer that almost any data relevant to the transfer of
rights in children, or desire for children, or filiation-in short, any-
thing having to do with arbitrarily establishing some portion of a
parent-child relationship-may prove illuminating in a study of
adoption. The definition also suggests that the volitional aspect of
fostering is of some importance.
The complexity of a construct like "adoption" imposes certain
limitations on the degree of refinement which is possible when it is
used as an analytic tool. Goodenough takes up one of these limita-
tions in chapter 15, in which he argues that adoption, as defined
by any group of natives (including ourselves), might be usefully
analyzed by considering that "parenthood" is a complex of rights
which may be shared, transferred, acceded to, delegated, surren-
dered, or otherwise circulated among social persons according to
culturally specific rules.
It is clear that a notion like "parents" must be looked at from
two points of view.
There is first of all a cultural construct (the
differentiating symbolism of which we see as biological); secondly,
there are cultural notions about the role functions of parents. Just
what parts of the latter are supposed to follow from just what
parts of the former is an empirical question. We cannot assume
just because we feel that child care (for example) is a responsi-
bility of 'mothers' and 'fathers' (defined biologically) that the
natives have the same view of the matter.
The danger in making
assumptions of this kind lies in the possibility of unconscious dis-
tortion of the data. If non-parents in another society assume some
functions which we suppose are integral to the cultural definition
of natural parents, then we are forced to construe adopters as be-
ing like natural parents in ways for which the natives see no need.
To put the matter another way: What we think of as "parent-
hood" is not necessarily a unitary entity which can be thought of
as residing all together in one place or another.
The definition on page 3 thus fails in two places to character-
ize Oceanic adoption: by stipulating that it is a "child of other
parents" who is taken in adoption we have glossed over the many
different ways in which it is possible to be a "child" or a "parent";
by stipulating that an adopted child is "taken as one's own" we
have committed the same error.
But even if the definition were amended to take account of
such difficulties, it still could not be used as part of a cross-
cultural typology in which acts or beliefs or practices which con-
stitute "adoption" are clearly distinguished from those which fall
under the heading of "fosterage," or "ritual kinship," or the like.
Consider, for example, the problem of deciding who is and who is
not "adopted" in Eastern Oceania.
A child who is taken in temporarily (i.e., "fostered") may re-
main with his foster parents until adulthood, treating them as
natural parents and being treated by them as their natural child.
That is, the foster parents may come to regard such a child as
"their own" (especially if they were not entirely unwilling to take
the child in the first place). It may even transpire that the child
receives from the foster parents some land or other evidence of
social filiation. On the other hand, a child who was clearly
"adopted" may soon leave his adoptive parents and return to his
natural parents, with all parties to the adoption henceforth acting
as though it never took place. Under circumstances such as these
the natives might easily be confused as to whether or not such
children were "really adopted." In the absence of a clear legalistic
basis for deciding such matters, the natives will employ a multi-
variable "ideal-type" model of what constitutes "proper adop-
tion," to which few actual cases will conform and to which many
marginal cases will seem at least vaguely germane. The number of
variables which are relevant to any such ideal-type is so great that
any classification which attempts to take all of them into account
is doomed to failure.
A similar problem of classification occurs in relation to customs
which are considered by the natives as akin to child-adoption, but
which involve establishing sibling or parental relations. In at least
some parts of the culture area under consideration, the natives
appear to feel that the 'adoption' of parents or siblings is similar in
some way to the adoption of children (Silverman, chapter 8). Any
general theory of Oceanic kinship will have to account for such
beliefs. Child-adoption appears to be only one of many possible
ways of establishing as if kin relationships. The presence and the
absence of any way of establishing kin-like relationships are
equally deserving of explanation.
At one remove from the adoption of 'children', 'parents', and
'siblings', (and of 'grandchildren' and 'grandparents') is the adop-
tion of 'aunts', 'uncles', 'cousins', 'nieces', and 'nephews'. In
Polynesian societies, these categories are not terminologically dis-
tinct from primary kin, although from an American point of view,
it is individuals in the last two of these genealogical positions who
constitute the bulk of the adoptees. If adoption establishes an as if
parent-child (or sibling or grandparent-grandchild) relationship,
then it may not be reasonable (as I have argued in chapter 6) to
think of Oceanic 'adoption' as adoption, since the adopter and the
adoptee are often already in such a relationship to each other
before the adoption-that is, there is no change as a result of the
'adoption' in many important aspects of their relationship.
To expand the problem just a bit further: Adoption, of what-
ever sort and however defined, establishes long-term, solidary
relationships of the kind which are usually ascribed culturally to
close biological kin. Closest to this in intent are marital relation-
ships and the relationships established between friends. A marriage
or the initiation of a formal friendship, like an adoption, often
appears to initiate relationships which are tantamount to those of
consanguinity in at least some respects.1
How is one to clearly
distinguish such processes from "adoption?"
It will be clear from the preceding discussion that there is great
hazard in using the term "adoption" in descriptive ethnography
without indicating carefully what it is (if anything) that is being
translated by the term. Without divesting the term of its many
connotations in English, the ethnographer (and his readers) will
tend unconsciously to make inferences based on their own cultural
experience-inferences which are generally not warranted. It will
be noted that the authors of the following chapters have taken
considerable care to specify what the term "adoption" means in
each chapter. This is done usually by associating the term with a
native term and with the concepts to which the native term
alludes. By starting the analysis with a culturally specific and lexi-
cally labelled category, we insure our consideration of the full
range of information relevant to what the natives themselves con-
sider 'adoption'. However refractory to analysis such an array of
data may be, the fact remains that the data must be analyzed all
together, in all of their complexity, if we hope to understand the
native point of view. As an analytic tool the term "adoption" may
prove to be quite useless; as a descriptive term it appears to be
Of considerably greater interest than the matter of definition
are the innumerable research leads suggested by the ethnography
in the following chapters. Some of these have already been
touched upon; another is the matter of motivation.
The extremely high incidence of adoption in many parts of
Oceania requires an explanation. What is the motivation for prac-
ticing it on this scale? Levy (chapter 4) reviews some of the theo-
retical statements in the earlier literature, and much additional
information on the matter of motivation is available in each of the
following chapters. For reasons which Fischer (chapter 11) dis-
cusses at some length, childlessness (due to infertility) appears to
be a common problem in Oceania. R. Goodenough (chapter 12)
and others demonstrate a correlation between the rate of adoption
l l
and the rate of infecundity. But infertility is not the entire expla-
nation for high adoption rates, since many people adopt others'
children after their own have grown up, or to "complete" a family
with young children of a single sex, or to substitute for deceased
children. "Childlessness" is thus a frame of mind caused by many
factors, of which physiological incapacity for parenthood is only
one. The etiology of this frame of mind is discussed by Howard et
al. (chapter 2) and R. Goodenough (chapter 12), who adduce
psychologically-oriented theories to account fer it. Their theories
deserve to be tested in other Oceanic groups which do not practice
adoption to such a great extent.
Of special interest in this connection is the finding that high
adoption rates may persist in the face of radical social and cultural
change (see Howard et al., chapter 2) and in different economic
contexts (Kay l 963b). Hooper's study of adoption in relation to
economic-ecological variables (chapter 3) indicates the value of
controlled comparions in formulating general principles to account
for dependent variables.
All of the field studies in this volume support the empirical gen-
eralization that there is a universal desire for children throughout
Eastern Oceania. Whether this desire is somehow "stronger" in this
area than elsewhere is a matter for further investigation. In fact,
the whole matter of cultural differences in the value attached to
children and to child-raising is an important area of kinship re-
search that is suggested by the present volume. There appear to be
great differences, for example, in the status of unwed mothers in
Eastern Oceania and in the United States; differences based, at
least in part, on the value attached to having young children.
Another problem is the discrepancy between the motivations
advanced by different informants, and also the discrepancy be-
tween what the same informant will say on one occasion and what
he or she will say on another day or in a different context. As
Lieber (chapter 7) points out, the motivations for maintaining an
adoptive relationship may change dramatically over the lifetime of
the parties to the adoption. The motives for adoption which are
adduced by adopters tend to differ from the motives which are
ascribed to the adopters by the natural parents of the adopted
child, or by disinterested third parties. This appears to be part of
the larger question of the relationship between social acts (such as
adoption) and the stated or conscious motivations for engaging in
them. Is any individual a reliable informant about the forces which
impel him to engage in a particular act? It is certain that the stated
motivations for adoption are interesting data towards an analysis
of those public attitudes which are culturally valued, but do they
suffice to explain social acts?
The motivations for adopting are only one side of the issue;
there is also the question of the motivation of parents who give up
their children in adoption. If children are especially cherished (as
appears generally to be the case in this area) and if childbearing
and child-rearing are valued (as also appears to be the case) how is
it that any sizable number of parents can be prevailed upon to re-
linquish their offspring? Do parents not find it difficult to part
with the children they give up in adoption? Is not the Oceanic
tendency to define an "appropriate" adoption as one wherein the
child is asked for before birth and taken from his parents as soon
as he is able to survive (i.e., at weaning) an indication that parents
cannot reasonably be expected to part with their children once
they have become attached to them? If this is so, what explains
the occasional willingness of parents to allow the adoption of
older children?
Why is it difficult for parents to refuse requests for the adop-
tion of a child? What are the sanctions which militate against the
rnfusal of adoption requests? Are there no limits as to those who
have a right to request another's child? If parental responsibilities
and obligations are so minimally stressed that adoption becomes a
routine matter, why are parents unwilling to kill their children or
abandon them totally? These matters are taken up in chapter 6
and in other chapters. It suffices to say here that all of the avail-
able evidence suggests that specific obligations to give up children
in adoption are merely part of more general obligations between
certain categories of kin; this seems evident in all of the societies
which practice adoption on a large scale.
In apparent intent and overt form, adoption establishes the
same sort of link between a child and an adult as exists, by cultural
definition, between the child and his natural parent. As Silverman
(chapter 8) points out, "culture" must cope with insufficiencies of
"nature" (i.e., the ideal world of things-as-they-should-be). Cer-
tuinly this is the sense of American adoption. But why is there no
decisive legal act in Oceania which formalizes the severance of
kinship bonds between an adopted child and his natural parents?
Is this merely an instance of the general rule that procedures for
changing status (e.g., marriage) tend toward informality in this
part of the world?
Do Americans really believe that ties of consanguinity can be
effaced and then be recreated by a legal act,
or does the form of
American adoption indicate a special emphasis on only some of
the numerous components of the parent-child relationship, such as
responsibility and authority? Is there reason to suppose that
Oceanic peoples do not recognize the possibility of supplanting a
relationship founded on natural parenthood? (For that matter, do
people anywhere believe that this is possible?) Should Oceanic
adoption not be thought of more as creating an additional parental
relationship for the adopted child rather than as replacing the
bond between parents and their natural children? Is there some
special significance to the fact that stepchildren are not fre-
quently "adopted" in Oceanic societies, and why do Oceanic
peoples generally not adopt 'strangers'?
1 5
The answer to questions about the nature of kinship can only
be determined on the basis of investigations into the precise extent
to which adoptive relations are construed as tantamount to "bio-
logical" relationships. To what degree does adoption really re-
arrange ties of consanguinity? Is it possible that, appearances to
the contrary, adoption is perceived as not changing the "natural"
relationship of a child to his "biological" parents (as I have argued
in chapter 6). Or is it perhaps the case, as Levy has suggested
(chapter 4), that biological parenthood may, in some societies, be
usefully construed as "contingent" in the same way that an adop-
tive relationship is contingent?
If an adoptive relationship is precisely equivalent to the rela-
tionship between natural parents and their children, why are
adopted children forbidden to marry their natal kin? Why is it that
natural parents in Oceania tend to resume a parental role when the
adoptive parents die? What accounts for the tendency of adopted
children to leave their adoptive parents and return "home," espe-
cially in illness or adversity? Is an adopted child treated like a nat-
ural child? Is he treated with more or less consideration than a
natural child, and if so, why? Is he under some legal disability
(perhaps with regard to inheritance of land or rank)? Is greater
care taken in adoptive relationships to fulfill parent-child role ex-
pectations? Is natal clan membership maintained by a child who
has been adopted out? Is land generally inherited by a child from
his natural parents or from the kin groups of his natural parents?
To precisely what degree are ties maintained between an adopted
child and his natal kin group, and to precisely what degree is
the adopted child incorporated into the group of his adopter?
For example, are the members of the adopting group obliged to
accept the adoptive child as "one of their own?" Are they con-
sulted by the prospective adopters? Do adoptive ties remain as
relevant as consanguineal ties in future generations? Particularly
full coverage of these points is provided by Ottino (chapter 5).
The answers to questions such as these are the answers to the
question of the nature of kinship. We shall understand what con-
sanguinity means when we understand the ways in which ties that
are not consanguineal can never become consanguineal ties, as
Scheffler points out (chapter 14 ). In short, it may be easier to
understand consanguinity in terms of what it is not, than in terms
of what it is. Similarly, our understanding of why some people
adopt will be greatly enhanced by an understanding of why others
do not.
I am indebted to Raymonde Carroll, Alan Howard, Robert I. Levy, Henry
P. Lundsgaarde, Harold W. Scheffler, David M. Schneider, Albert J. Schutz,
and Martin G. Silverman for detailed comments on earlier drafts of this
2 The value of a direct comparison between American and Oceanic adoption
institutions was suggested to me by Robert I. Levy.
3 The ideal qualifications for adoptive parents in the United States are set
out in Child Welfare Leagu.e of America Standards for Adoption Service.
In addition to the criteria mentioned, it would appear that religion is also
important: "It is presumed that a child who is relinquished for adoption
ordinarily will be placed in a home where the religion of the adoptive
parents is the same as that of his parents, unless the parents or mother
have specified that the child should or may be placed with a family of
another religion" (Child Welfare League of America 1958: 12).
4 Of crucial relevance to American adoption is the position of the unwed
mother: "Although there has been considerable modification of society's
harsh attitude toward her, it still remains extremely difficult for the un-
married mother to raise her child successfully in our culture without
damage to the child and to herself" (Child Welfare League of America
5 It has been estimated that adoptees who are unrelated to their adopters
constitute 1 percent of the population of the U.S.A., while another 1 per-
cent of the population consists of adoptees who are related by marriage or
blood to their adopters (U.S. Children's Bureau 1964). Children adopted
by relatives (mostly stepparents) in 1955 were usually living in their adopt-
ers' homes. Half of these who were adopted by non-relatives "were placed
in their adoptive homes independently, about half of them by their
parents or relatives" (U.S. Children's Bureau 1958:2).
6 Childlessness is in fact a desirable characteristic of American adopters, at
least from the point of view of the social service agencies. "In selecting a
family for a particular child, childless applicants may be given prefer-
ence ... " (Child Welfare League of America 1958:37).
7 I am informed by Robert I. Levy that many American couples who
want to adopt children, but who are reluctant to accept children who have
been in institutions and possibly traumatized, attempt to arrange adop-
tions with physicians on a "when available" basis. To this extent, at least
some American adopters appear willing to commit themselves to an adop-
tion before the child is born.
8 The history of adoption in the common law will be found in Pollock and
Maitland 1898; a brief history of adoption laws in the United States is
contained in chapter 1 of Witmer et al. 1963; a comparative study of
adoption legislation in European countries is contained in United Nations,
Department of Social Affairs 1956. The conception of adoption in
Western Europe and the United States has changed from that of providing
a child for a family to that of securing a proper home life for a deprived
child (International Child Welfare Review 1949, 1950, cited in Pringle
1967: 175). The radical change in European adoption law during the past
two centuries is obvious when we remember that: "The Code Napoleon,
1894, articles 343 ff, provided only for adoption by a person who was at
least fifty years old; who, at the time of the adoption had neither children
nor legitimate descendents; who was at least fifteen years older than the
person to be adopted; and who had (with certain exceptions) furnished
assistance to, and taken care of, the person to be adopted during his mi-
nority and for at least six years. The person proposed to be adopted had to
have reached his majority and he retained all his rights in his natural
family ... "(Witmer et al. 1963:20).
9 These working definitions are my own; they conform generally to the
usage in the chapters following. In the ethnographic literature there is no
noticeable consistency in the use of these two terms.
10 I am indebted to David M. Schneider for reminding me of this point (see
Schneider 1965, 1968).
11 The fact that responsibilities for child care are shared among a wide circle
of kin is stressed in many of the ethnographic accounts in this volume.
12 Compare the following definition of adoption: "the method provided by
law of establishing the legal relationship of parent and child between
persons not so related by birth with the same mutual rights and obligations
that exist between children and their natural parents" (Report of the De-
partmental Committee on the Adoption of Children [England], cited in
Goodacre 1966:15; also Child Welfare League of America 1958:1, italics
13 Ethnographic investigation which explores the similarities in the institu-
tions of adoption, marriage, and formal friendship will also serve the useful
purpose of indicating the limitations of the distinction between "ascribed"
and "achieved" statuses, of which so much has been made in sociological
theory. The logical connection of adoption with other institutions is im-
plicit in many dictionary definitions of the word "adopt." For example:
"adopt ... to accept, receive or choose as one's own ... when not natur-
ally so. The receiving into a clan or a tribe of one from the outside, and
treating him as one of the same blood" (American Academy of Pediatrics
1959:38, italics added).
14 The importance in America of natural parenthood as a symbol which is
closely associated with symbols of role performance is amply attested to
by the literature on adoption. Adoptees are said to experience "genealog-
ical bewilderment" (Sants 1964); and the behavior of adoptive parents is
often modified by the knowledge that they are not the "real parents" and
that their adoptive child is not their "real child" (Krugman 1964; Kirk
1959). Thus, there are numerous and difficult dilemmas in which adoptive
parents and their adopted children find themselves (Paton 1954; Kirk
1964, chapter 3), dilemmas which all hinge on the performance of roles
conventionally assigned to the closest sort of consanguines by those who
are not in fact so related.
15 The frequent adoption of castaways in aboriginal times perhaps indicates
that a 'stranger' is two things: someone who is not a relative and who also
has other kin to whom he has access. The castaway is cut off from his
other relatives and thereby becomes eligible for adoption because he is
something less than a 'stranger'.
Alan Howard
Robert H. Heighton, Jr.
Cathie E. Jordan
Ronald G. Gallimore
Adoption was an important aspect of traditional Hawaiian culture,
and research among modern Hawaiians demonstrates that it is still
very important, despite the fact so many other features of the
traditional culture have been lost or distorted beyond recognition.
In this chapter we attempt to accomplish three tasks: to review
the literature describing traditional Hawaiian adoption patterns, to
present material from our current researcn detailing modern prac-
tices, and to provide an explanation for the continued significance
of adoption among modern Hawaiians. We define "adoption" as
the establishment of relationship rules appropriate to a specific set
of kinsmen between persons not occupying those genealogical
positions. Adoption may be formal (ceremonially or legally vali-
dated) or informal, and it may or may not involve a change of
residence. In discussing the traditional forms of adoption in this
chapter, we use the Hawaiian terminology, but this proves unsatis-
foclory when it comes to modern practices. For dealing with the
luller we distinguish three forms: legal adoption; hanai, a Hawaiian
lcrm designating 'an agreement to transfer primary parental rights
over a child'; and fosterage, which involves the taking care of a
child without a transfer of primary parental rights. These are
operationally defined, based on a questionnaire sent out to over
one thousand Hawaiian households. Throughout the paper the
unqualified term "adoption" is used as a general term which in-
cludes all of the more specific categories mentioned above.
The aboriginal Hawaiians distinguished three forms of adoption:
ho 'okama, ho 'okane - ho 'owahine, and hanai.
Ho'okama may be literally translated 'to make a child' (ho'o
being a causative prefix; kama 'child' or 'person'). The implication
is that the adopting parents took as their own either someone
else's child or an adult for whom they had a special regard. Con-
cerning the ho 'okama adoption of children Handy and Pukui
The adopting parent becomes to the child makua ho'okama (literally 'parent
making child his own'), while the child is known as kaikamahine ho'okama if
it is a girl, or kaiki ho'okama if a boy. The relationship comes about as a
result of mutual affection and agreement, at first tacit, then unobtrusively
discussed, between the child and the older person; the part of the child's true
parents, if living, is normally negative; although if there is a strong dislike for
the would-be adopting parent the true parent is capable of interfering. This is
a relationship involving love, respect and courtesy, but not necessarily respon-
sibility of any sort, and rarely a change of residence ( 1958: 71 ).
Ho 'okama referred also to the adoption by older persons of
younger adults. Between adopted adults sibling terms were often
used: kaikua'ana ho'okama (kaikua'ana 'older sibling of the same
sex as speaker'), kaikaina ho'okama (kaikaina 'younger sibling of
the same sex as speaker'), kaikuahine ho 'okama ( kaikuahine 'sister'
when male speaking) or kaikunane ho 'okama ( kaikunane 'brother'
when female speaking) (Pukui and Elbert 195 7: 115 ).
The ho 'okane - ho 'owahine relationship is described by Handy
and Pukui as "an adoptive platonic marital relationship between
persons of opposite sex."
Such a relationship might be initiated
by a boy who took a fancy to a girl and wanted her as his wahine
ho 'owahine. He proposed this either to the girl or her parents, and
if the offer was accepted, his relatives made a feast and roasted a
pig to seal the relationship. Alternatively the girl might make the
suggestion that the man become her kane ho 'okane. "Sometimes
the wahine may be but a child of six or seven while the kane is an
adult farmer or fisherman; or the kane may be just a little lad
while his wahine is a mature housewife" (1958:54-55). The fol-
lowing two incidents cited by Handy and Pukui illustrate this form
of adoption:
Hoeawa of Puna became the kane ho'okane of a prominent Hilo-pali-ku
woman named Hela. Both were married. Hoeawa and Hela were as good to
each other as brother and sister. Hoeawa's niece used to go to Hela's with her
cousin, Hoeawa's daughter, and both were treated like own nieces. Hela died
many years before Hoeawa. She used to give him gifts to take home, and his
wife used to make fine mats for him to take to his wahine ho'owahine. Such
wahine ho'owahine and kane ho'okane never made love to each other
I've always liked fishing and dancing and have learned to plait mats, sew, keep
house and to plant. I did not know that someone was watching me until a
man came to my father's house with his daughter. As soon as I got home with
some fish that I had caught that morning, I was called in by my father who
told me that I was wanted to be the girl's husband. I looked at her. I had not
seen her before. She was well built, homely, but had a sweet disposition. Still
1 did not want to marry her and leave my father alone. I said, 'I am poor and
unable to buy shoes for myself. How could I support a lovely wife like this
maiden? Let her be my wahine ho'owahine instead.' The father was satisfied
and I think the girl was glad, too. For years we were the best of friends, this
girl and I. I caught fish for her and carried her gifts whenever I could and she
always treated me like her own brother until she died twenty years ago. Her
husband was a good man (1958:55-56).
Beaglehole (1939:62) reports a somewhat different use of the
terms, stating that they referred to parents adopting into their
household the favorite playmate of their child.
The term hanai, as a verb, means 'to feed'. As a noun it refers to
the provider or to a person for whom one provides food. In tradi-
tional times it was used to refer to the chiefs who were provided
with food by their subjects. "In early Hawaiian traditions, the
Alii-nui or leige-lord was referred to by the people as their hanai
and they, in tum, were his ohua [retainers). The ohua were desig-
nated either hoaaina, tenants placed upon the land by agreement,
or kupa, hereditary tenants. The word hanai, to the Hawaiians,
meant more than just 'the fostering relationship'. It implied 'a
sympathetic embrace toward one, whose very existence depended
upon that embrace' "(Kenn 1939:46-47).
When used in reference to children the concept of hanai implied
that the child had been taken into the household of his makua
hanai 'feeding parents' and reared as their offspring. They assumed
complete social rights and obligations in raising their kama hanai
'feeding child' (Beaglehole 1939: 162-163). The adopting parents
were also known as the child's kahu hanai. Kahu is translated
'guardian' or 'nurse', and has a secondary meaning referring to
cooking, specifically, 'to tend or cook at an oven' (Pukui and
Elbert 1957:105).
The evidence available to us from elderly informants and docu-
mentary sources suggests that four principles were of particular
importance in the traditional patterning of hanai relations. These
were kinship and seniority between the natural parents and the
adopting parents, and the age and sex of the child.
Apparently hanai children were almost always taken from
within one's own 'ohana 'kindred grouping', although Handy and
Pukui assert that "Sometimes a child was asked for by a friend
... before it was born, with the idea of thus cementing the friend-
ship through the care of the child, but not if it was the eldest"
(1958:72). The significance of kinship for adoptive relations is
apparent in the comments of Charles Kenn: "The [adopted] child
became a part of the new household (ohana) if the [adopters]
were also blood relatives; otherwise, it remained a part of the
ohua, or those that were attached to the household unit but not
related in any way blood [sic] to the akana, or family proper. The
Hawaiians were very careful as to the parentage of a keiki-hanai or
[adopted] child and did not [adopt] 'indiscriminately' as is often
believed" n 939:47).
Within the kindred grouping seniority was relevant, for if a
senior relative asked for a junior relative's child it was almost im-
possible for the latter to refuse the request. The rights of grand-
parents were particularly strong. Some informants said that grand-
parents' claims on their grandchildren took precedence over those
of the natural parents, and that the parents had to have the grand-
parents' consent in order to keep their own children. The older
siblings of parents could also exercise seniority and demand a
child, but younger ones could not. The latter might ask for a child,
but they had to rely upon implicit supernatural sanctions, rather
than authority, to back up their request. If a child's parents were
to die, his siblings would be expected to hanai him. Within the
child's own generation the first-born was responsible for his sib-
lings in the case of the parents' death or absence, particularly if
there were no senior relatives to assume the responsibility. If the
first-born was old enough to care for his siblings, he was expected
to treat them as hanai until they were old enough to take care of
themselves. If the oldest died or was absent the next oldest took
the responsibility, and so on.
Sex was significant in that the first-born male child was con-
sidered to belong to the father's side 'ao 'ao kane and the first
female child to the mother's side 'ao 'ao wahine (Handy and Pukui
1958:72; Kamakau 1964:26). An example of this is recorded in
the tale of a Kauai chief named Holoholoku:
When Holoholoku, the [adoptive] parent of Maihinali'i, came to Oahu from
Kauai (as he was instructed to do in a dream) to seek the wife chosen for his
ward, he found her grandmother at Makapu'u. The girl, Malei, was sent for
and before she left for Kauai the grandmother expressed her thoughts to
Holoholoku. If a daughter should be born of this union of the Oahu chiefess
and the Kauai chief, then she (the grandmother) would want to rear her here
on Oahu-but if the child was a boy (like his father) then he was to be reared
by those of his father's side of the family. Should there be no daughter born,
then after death Malei must be brought back to her old home for burial
(Handy and Pukui 1958:43).
For subsequent children the picture is somewhat obscure and
it is likely that a number of considerations, including sex, were
taken into account when one of them was adopted.
Age was a factor in that there was a strong preference for adopt-
ing children at birth or shortly thereafter. Kenn states that a child
was often promised before birth, especially by one sister to an-
other if the second had no children of her own, and if the child
was asked for ( 1939:47); but according to Handy and Pukui some
Hawaiians regarded it as a bad thing to ask for an unborn child
while it was ma kahi haiki 'in an uncertain place', and that it was
better to wait until after it was born, or ma kahi akea 'in a wide
space' (1958:72).
At the time that the child was given away the natural parents
might utter a ceremonial phrase such as Ke haawi aku nei maua i
ke keiki ia olua, kukae-a-na 'au, 'We give the child to you, excre-
ment, intestines and all'. The contract was binding and there were
severe supernatural sanctions supporting it. If a disagreement arose
between the two sets of parents, and the 'natural parents' (makua
pono i) tried to recover their child, it was believed that since the
faith had been broken the child would either become very sick or
would die. Such a disagreement was called hukihuki 'pulling back
and forth'. If, however, the child returned of his own accord, then
he was referred to by his adopters as kukae ka ke kahu hanai, 'an
ungrateful and unappreciative child, after all the care and atten-
tion had been given him' (Handy and Pukui 1958:72; Kenn
Indications are that adoption was quite frequent among Hawai-
ians during the precontact and early postcontact periods, but we
have no reliable estimates. The reasons for adoption are somewhat
clearer, or at least they are more explicit, in the literature. Ho
'okama, for example, figured importantly in the rank system:
This form of relationship existed in order to retain the power in a ruling
house, and, most important of all, to keep the blood undefiled and so to per-
petuate this mana or psychic force in the clan. If a chief had no direct heir, he
adopted one, but in doing so he had to choose from the closest of kin, chil-
dren of his brother or sister only. In making a child hookama, he passed on to
it all the prerogatives, rights, and privileges of his own high position, in order
that it might succeed him to leadership. In the case of the hanai relationship,
even in the same family, the rights of the [adopters] are not necessarily trans-
mitted to the keiki-hanai ['adopted child'] . Hookama ... is in essence, an
elevating instrument. This form of adoption was also used in another way.
For example, two brothers of royal birth might choose mates. The elder,
designated as haku and therefore possessing the right of leadership in his own
generation, let us say, married a woman of low caste. His son, if born before
that of his brother, became the haku within the new generation. If the
younger brother married a woman of high caste and his son were born after
his brother's son, this child, being younger, paid respect to his cousin. But, his
grandparents, in order to give this second boy a higher place in his generation,
might adopt him and thus bestow upon him all the rights and privileges
enjoyed by themselves. This act immediately placed the boy on the same
social level as his father, becoming, as it were, his father's brother. Although
the son of the older brother was still the haku, the son of the younger brother
automatically became the leader of his generation. This is done to retain the
blood purity (Kenn 1939:47-48).
A chief might also adopt a child if it wet or soiled him, for
under such circumstances the child had either to be killed or
adopted (Handy and Pukui 1958:48-49). Presumably this could
happen only with the children of close relatives, since no others
were allowed to get close enough to a chief for the problem to
The desire of older people to hanai children was frequently
motivated by a wish to be provided for in their old age. This was
voiced in the expression, Hana a ka mea kama 'ole he/e kuewa i ke
a/anui 'What a childless person will eventually do is to wander
about uncared for on the highway', while another expression ad-
vises that one should Hanai kanaka, hiki ke ho'ounauna, 'Feed
human beings for they can be sent on errands' (Handy and.Pukui
1958:173, 168). Other economic motives were the desire to pass
on wealth to the adopted relative and the desire to relieve close
relatives of hardship if they had more children than they could
adequately provide for.
The most frequently cited motive for adoption, however, was
simply a fondness for children and a desire always to have some in
the household. People without children seem to have felt that
something very vital was missing in their lives, so they tried to
remedy the situation. In contradiction to this presumed motiva-
tion, the alleged prevalence of infanticide in early Hawaii might be
cited. Thus Ellis claimed that two-thirds of all children were de-
stroyed (1828:325), and Dibble, citing the same frequency, stated
that parents had no desire for children and that if children could
not be given away they were killed ( 1909: l 08). Nevertheless,
there are very few documented cases of infanticide among Hawai-
ians. Ellis mentioned only one, and the weight of evidence clearly
suggests a desire for children.
The motives for giving children in adoption seem to have in-
cluded the following:
l. The desire to create a bond between one's family and that of
the adopting parents. This was especially important for chiefs who
were thus assured alliances with other a/i'i (Hom 1948:24; Goo
1958: 17).
2. The belief that twins must be reared apart lest one or both
of them die (Hormann n.d.).
3. The belief in 'uha kapu 'taboo lap'. According to this belief
some women were so kapu that they could not raise their own
1.:hildren. Uha kapu was believed to result from the fondness of an
aumakua 'a personal or family god' for a woman. The aumakua
did not want the woman who was dear to it to be soiled by the
urine and feces of an infant; if such a woman attempted to raise
her own children, it was thought that the children would die or
become crippled (Handy and Pukui 1958:48-49). The children of
such a woman had to be raised by others.
4. Ali'i siblings of opposite sex might be raised apart in order
that they might marry later in life without regarding each other as
brother and sister. Such unions would produce offspring who were
ali'i niau pio 'chiefs who were higher in rank than either of their
parents' (Green and Beckwith 1924:246).
5. The belief that to refuse a request to hanai a child was to
risk the death or sickness of the child from the sorcery of the jeal-
ous would-be adopter (Yamamura 1941:137-138; Forster 1960:
6. A desire to have a child learn skills not possessed by his
parents. In such a case, the child might be apprenticed to an ex-
pert, into whose home he was taken, becoming for all practical
purposes an adoptive member of the family (Handy and Pukui
The role of a hanai child in his adoptive family depended to
some extent on the conditions under which he was adopted:
The 'feeding child' may be a mere waif taken in out of kindness, who in the
course of time automatically assumes a tacitly accepted role of servant in rela-
tion to the family and to the true children of its 'feeding parents'. This is said
to be hanai 'ai i kanaka, or reared to serve the true children of the family. It
may be, on the other hand, an orphan or the child of a relative or dear friend,
formally adopted and for whom the 'feeding parent' comes to have affection
that may be as great as that for the biological offspring. Under such circum-
stances the 'feeding child' comes to feel more active affection for the family
that raises it and in whose home it spends its childhood than for its true
parents; consequently, in later years, when the 'feeding child' is grown and
the 'feeding parents' are ageing, the deepest devotion is ofttimes felt and
shown on both sides in this relationship (Handy and Pukui 1958:71).
In some cases the hanai child was held to be so precious that he
was not allowed to work, or even to feed himself:
Sometimes boys thus taken as ... (hanai) were not allowed to do any work,
or to carry anything in the hand, or to plant, or to carry anything on the
shoulders, or to fish. This was in accordance with a vow taken by the [adopt-
crs] never to see the child perform any kind of labor as long as they were
alive. It was the same with some girls; the grandparents or [adopters] made
great pets of them, and they were not allowed to carry anything in their
hands, nor were they taught to beat or to print tapa because that was work
that soiled the hands. Such children would be seated on piles of mats or tapa,
or on the chest or lap of their attendant (kahu), to be fed poi by dropping it
into their mouths (e kau ai ka ai), and fish by mouth (ka i'a a o ka pu-'a), lest
they choke on lumps in the poi or on fishbones; and they would be given
water from the mouth (mumu i ka wai i ka waha) lest they choke and the
precious ones come to harm (Kamakau 1964:26-27).
There was also a temporary form of fosterage, usually involving
older children rather than infants. When parents became ill or had
to be away from their home for a period of time, they sometimes
asked one of their relatives to care for their children. These chil-
dren were referred to as their adopters' luhi 'burden' and it was
understood that they would be returned to their natural parents
upon request.
Finally, relations were sometimes established between families
which, though not formalized, were modeled on kinship:
l lousehold guests not related to the family proper, were referred to as ohana
makamaka. They were allowed to share with the family whatever it had to
offer, and were different from the ohua in that they were not compelled to
Jo any work. They became the aialo, privileged to eat at the same eating
place as the ohana. This was a high honor bestowed upon the guest in ancient
I luwaii. The outgrowth of this practice has come to be called "calabash" rela-
lionship, in which one family claims relationship to another because in the
past, their common ancestors ate together out of the same calabash of poi
(Kenn 1939:47).
Adoption still has a high incidence among the Hawaiian popula-
tion. 2 For a people whose customs have otherwise changed so
drastically, the persistence of adoption is dramatic, not only be-
cuusc of its frequency, but because it is so often a focus of con-
Cl'rn and powerful expressions of emotion. Indeed, on the basis of
our current field research, we would assert that the importance of
udoption is one of the characteristics which distinguish contem-
pornry Hawaiians from other American ethnic groups.
Through the auspices of the Hawaiian Homes Commission, we
distributed household questionnaires to four Hawaiian Homestead
communities on Oahu. In order to lease homestead land a person
has to be able to demonstrate that he is at least half Hawaiian by
genealogical ancestry. Included in the questionnaire were items
pertaining to legally adopted and hanai children in the household.
Forms were received from 681 households out of a total of 1022
to whom they were sent, representing a 66.6 percent return.
Either a legally adopted or hanai child in the household was re-
ported on 132 or 19.4 percent of the forms. An additional 56
households contained children under eighteen years of age who
were living in a condition of fosterage, i.e., neither of their legal
parents were present in the household. If we consider all these
children as "adopted," the total number of households containing
adopted children was 188, or 27.6 percent of the total number
responding. All together, 334 adopted children were included in
these 188 households, for a mean of 1.8 per household. A break-
down of the figures by type of adoption in the four communities
appears in Table 1.
The figures obtained for the homestead communities are consis-
tent with those obtained by Forster in a study of two rural com-
munities on Maui. He found that 17 out of a total of 73 Hawaiian
households, or 23.2 percent, contained adopted members. The 17
households contained a total of 39 children who were considered
hanai but were not legally adopted and three who were legally
adopted, for a mean of 2.5 per household ( 1960:97).
Although adoption remains prevalent, the form that it takes,
the terminology used to describe it, its functions, and the motives
behind it have all been modified. This is particularly true of the
ho 'okama and ho 'okane-ho 'owahine types of adoption. These
terms are understood only by the older folks, many of whom are
vague as to the exact meanings. Today ho 'okama usually refers to
legal adoption. The only modern usage of ho 'okane and ho 'o-
wahine that has come to our attention is in reference to the taking
of a sweetheart or spouse. The term luhi is also used infrequently
nowadays. Instead, Hawaiians simply say that they are taking care
of a child for a time. As a matter of fact, Hawaiian terms are rarely
used when talking about adoption. When speaking about the sub-
ject, our informants usually say something like "I took him when
he was two weeks old," or "She was taken by her grandmother."
TABLE I Frequency of Adoption by Types in Four Hawaiian
Homestead Communities
I louseholds Reporting
and No. of Children Nanakuli Waimanalo Kewalo Papakolea Total
I louseholds reporting
legal adoption 30
No. of children 40
I louseholds reporting hanai 44
No. of children 60
I louseholds reporting
foster age 34
No. of children 58
Total households reporting
all types of adoption 108
Total no. of children 158
Total households responding 265
Total households with adopt-
ed children (all types)'1 86
% households with adopted
children 32.5
31.6 25.5
Not equal to the total of households with legal adoption, hanai, or fosterage, since some
hm111cholds contained children of more than one category: 3 contained legally adopted
111d hanai children, 8 legally adopted and foster children, 9 hanai and foster children,
111d 3 legally adopted, hanai, and foster children.
ll'l'hll probable reason that the Waimanalo rate is significantly lower than the other three
homllsteads is because it is a newer community and has a lower proportion of women
uvor thirty-five years old (see Table 6).
The term hanai is still used on occasion, but its definition varies.
Nome people use it to refer to any permanent or temporary foster-
lll&C of a child or adult that involves the assumption of economic
ro11ponsibility for more than a few weeks. Others restrict it to rela-
tively permanent arrangements involving the full assumption of
parental rights and obligations. When pressed, most of our infor-
mants distinguished between adoption and hanai, with the former
referring to a legal and the latter to a non-legal assumption of
parental rights and obligations.
Despite the assertion by Kenn that today "there is often indis-
criminate adoption without knowing the background of the child"
(1939:47), our evidence suggests a strong preference among
modern Hawaiians for adopting the children of relatives or close
friends. Thus Forster found that 37 of the 39 hanai children in his
communities had come from the family of a relative. The two ex-
ceptions were cases involving close friends (1960: 97). In our
homestead sample of 334 adopted children, 269 (80.5 percent)
were known to have come from relatives, for 17 (5.1 percent)
there was no information, and 48 (14.4 percent) came from non-
relatives. Of the latter, 24 were said to have been obtained from
friends, and for 13 no additional information was available; only
11 had come from institutions (see Table 2). Significantly, the
Liliuokalani Trust, whose child welfare section is concerned with
the placement of orphaned Hawaiian children, reports that in
1965 only 4 of the 17 children they placed in foster homes were
taken by Hawaiian families. Only as a last resort do Hawaiians rely
upon adoption agencies.
Several considerations seem to be involved in the desire to
adopt only from relatives or close friends. One is rooted in the
widespread belief that a child's character is largely inherited; not
knowing the parents of a foster child means that "you don't know
what you're getting." Secondly, it is easier to deal with people
with whom one already has an established "account" of reciprocal
rights and obligations. Hawaiians feel that conflict is less likely
with such people and that any problems which do arise can be
more easily mediated. If the natural parents are neither relatives
nor friends, there is a lingering fear that the contract will be
broken either by them or the child. This fear is implicit in the re-
marks of a young mother: "I would rather take care of someone in
the family. You feel closer to them. I feel if children are adopted
from outside the family, they will ask about their parents when
they grow up and will go to find out who their parents are; I
wouldn't keep it from them. If they're from within the family,
they know who their parents are."
In contrast to the ideal, middle class, mainland pattern, which
stresses the severance of ties between an adopted child and his
natural parents, Hawaiian ideology stresses the reverse. It is re-
garded as not only desirable for the child to know his natural par-
ents but for him to maintain intimate contact with them as well,
as the following case illustrates.
The M's have a hanai child, Danny, who has lived with them since he was a
few days old. His natural father is a friend of Mr. M's. He and his wife live a
few miles away and occasionally visit the M family. On these occasions Danny
is made to kiss his natural parents and is instructed to call them "mommy"
and "daddy." Periodically, at his natural parents' request, he is sent to visit
them. All these activities are imposed on Danny against his will. The M's do
this partly because they wish to remain in the favor of Danny's natural par-
ents in order to forestall any attempt on their part to take the child back, and
also because the M's feel that it is right for Danny to know who his parents
are and for him to maintain contact with them.
Finally, and perhaps most importantly, it appears that Hawai-
ians adopt from relatives and friends because they are accustomed
to fulfilling their needs through personalized transactions rather
than through formal Western institutions.
The rule that Hawaiians adopt only from relatives or close
friends is sometimes broken when no children are available or in
situations in which a child has been abandoned or is otherwise in
desperate circumstances. Such children will often be taken into a
household, even if it is already very large and poverty-stricken:
A family ... with eleven children took in an unmarried girl who was about to
have a child. They were having a difficult time supporting the large family,
but did not refuse aid. The father worked for the WPA earning about forty
dollars a month. This amount was supplemented by fishing and by taro
planted in the mountains. The young lady was allowed to stay in the home as
long as she wanted to, and she left the child in the care of the family when
she departed. The child is treated as one of the family and no discriminations
urc made (Yamamura 1941:43).
However, despite a great show of compassion for children who
need a home, objections are sometimes made by relatives who dis-
upprove of adopting unrelated children, as can be seen in the ac-
count of an elderly Hawaiian informant who was recounting the
:adoption of her foster daughter: "In 1920 the Spanish influenza
TABLE 2 Relationship of Adopters to Adoptees
Relationship Nanakuli Waimanalo Kewalo Papakolea Total No. Total%
d d d d

.... .... .... ....
.... u
Q. Oil Q. Oil Q. Q. Oil Q. Oil
0 -

- "C

- "C
"C -
... ....
--o u u u
c; "'

"' "'
o;S "'
"' "'
..... .... ....

:f l!
:f l!
c.8 c.8 ..c c.8 c.8 c.8
SoCh 2 10 9 0 1 3 0 0 2 1 0 0 3 11 14 28
DaCh 4 19 31 2 13 8 4 6 13 1 0 0 11 38 52 101
ChCh* 0 0 0 1 1 2 5 0 23 1 4 4 7 5 29 41
Total 6 29 40 3 15 13 9 6 38 3 4 4 21 54 95 170 50.9
SiCh 6 3 1 0 0 1 1 3 3 l 2 0 8 8 5 21
BrCh 0 5 0 2 0 0 2 0 0 0 0 0 4 5 0 9
MoSiCh 0 2 0 0 0 0 0 0 0 0 0 0 0 2 0 2
MoBrCh l 0 0 0 0 0 0 0 0 0 0 0 l 0 0
FaSiCh I 0 0 0 0 0 0 0 0 0 0 0 1 0 0
ego's generation 0 0 0 I 0 7 0 0 0
0 0 0 1 0 7 8
- I generation 0 3 3 0 0 5 2 0 I 0 0 I 2 3 10 IS
- 2 generation 0 0 1 0 0 3 0 0 3 0 1 0 0 1 7 8
Total 8 13 5 3 0 16 5 3 7
I 3 1 17 19 29 65 19.5
PATltlL..\TER..U. ~ O C1D1.D
SiCh 1 0 2 l 0 0 l 0 0
0 0 0 3 0 2 5
BrCh 2 0 0 0 0 0 3 0 0
1 1 0 6 1 0 7
FaSiCh 0 1 0 0 0 0 0 0 0
0 0 0 0 1 0
-1 generation 0 2 0 0 0 0 0 1 0 1 0 0 1 3 0 4
-2 generation 0 l 0 0 1 0 0 0 0 0 0 0 0 2 0 2
Total 3 4 2 I I 0 4 I 0 2 I 0 IO 7 2 19 5.7
unknown) 3 1 4 0 0 0 4 2 1 0 0 0 7 3 5 15 4.5
friend 8 7 4 0 1 0 1 1 2 0 0 0 9 9 6 24
institution 1 1 3 0 0 4 2 0 0 0 0 0 3 1 7 11
unknown 3 4 0 1 1 0 1 1 0 1 1 0 6 7 0 13
Total 12 12 7 I 2 4 4 2 2
I I 0 18 17 13 48 14.4
8 1 0 3 0 0 4 1 0 0 0 0 15 2 0 17 5.1
TOTAL 40 60 58 11 18 33 30 15 48 7 9 5 88 102 144 334 100.0
*Exact linkage unknown.
came to Hawaii in the wake of World War I. Many all over the
islands died of it. Some families were wiped out entirely. At this
time, a Japanese woman on Kauai died in giving birth to a baby
girl. A month later, the infant's father also died, and she was
brought here with the hope that a Japanese family would adopt
her; but that hope did not materialize. I wanted her. My husband
objected strenuously, and my relatives voiced their shocked com-
plaints-auwe, an unrelated child!"
The emphasis on kinship seniority has also remained significant
among modem Hawaiians. Grandparents in particular assert pres-
sure to hanai their children's children. In our homestead sample,
grandparents were the adopters of 170 of the 334 children (50.9
percent). If only those children adopted by relatives of known
generation are considered, 70.9 percent were taken from the
grandchild generation, 24.0 percent from the child generation, and
5.1 percent from the adopter's generation (see Table 2). While
demands from grandparents are potent, they are not always ac-
ceded to by the younger generation, however. Many young parents
refuse to give their children to anyone as long as they can ade-
quately care for them, especially since acculturation has lessened
the fear of supernatural sanctions for denying a request. This has
resulted in overt conflict in many instances:
Mrs. K, who is herself a hanai, is now living with her husband and three chil
dren in the same house with her hanai parents. The latter want to adopt her
youngest son, but she, dissatisfied with her own experience as a foster child,
refuses to let them. She says that they can "take care" of the boy if they
want to (since they are living in the same house), but they cannot hanai him.
Mr. A had an older son who had been a hanai This boy was now married and
the father of a son. Mr. A asked for the child to raise as his own, and his hanai
son refused. Mr. A says that he dearly loves his small grandson and that, ac-
cording to the old Hawaiian way, the child belongs to him. He says that his
son now believes in the haole ('Caucasian') way. Mr. A thinks that this is
wrong, but can do nothing. He insists that the grandchild should be his to
raise and love as he grows old.
Seniority also continues to operate within the parental genera-
tion, as can be seen in the case of Mr. B:
Mr. B's first son is a hanai from his eldest brother. It came about this way:
Mr. B and his wife had been married for two years but did not yet have a
child. His eldest brother, whose wife was then pregnant, hinted that Mr. B
would be welcome to hanai the coming child; finally Mr.Basked. His brother
promised the child if Mr. B would come and get it before it was a week old.
Mr. B said that any sibling could hanai a child from any other, but, where-
as elder siblings could ask for a younger sibling's child outright, it was usual
for younger siblings to wait for hints from older ones before requesting one
of the latter's children. He also said than an older sibling might force a
younger one to take a child, and that his second oldest brother asked him to
hanai an illegitimate son. Mr. B tried to refuse, but finally had to take the
Within the children's generation the concept of responsibility of
older siblings for younger ones is likewise retained to a marked
degree among modern Hawaiians. The H family provides an ex-
In the H family there are thirteen children who range in age from two to
thirty years old. The family is very cohesive and each of the older siblings
takes considerable responsibility in caring for the younger ones. The older
males, for example, contribute a large portion of their income to the expense
of raising the youngsters. They also take pride in playing a quasi-paternal role.
After work they come home and play with the smaller children-"sing, play
cards, but mostly talk story with them." They listen to a younger sibling's
problems, advise him, and act as disciplinarians when it is called for. Mrs. H
said that she wanted her oldest boy to inherit the property so that he could
stay and care for the small ones.
The older females are expected to assume the responsibility of "bringing
up" one or more of their younger siblings. Mrs. H said, "It's good for the
girls. They learn to be mothers before they have their own." The older girls
have almost complete jurisdiction over the younger children who are assigned
us their wards, with parental interference only under unusual circumstances.
The strength of this responsibility is manifest in the comment of one of the
older girls who said, "If I get married and have to leave home, Suzie will come
to live with me because she is my hanai." Kathy, a child of five, became
emotionally upset when her older sister Joan, who had raised her, got married
und went to the mainland. Mrs. H said that she couldn't do a thing with
Kuthy. "She keeps asking where Joan is and when she's coming home. All she
docs is cry when I tell her Joan is on the mainland. I guess we'll have to send
her to the mainland for a while." This responsibility extends to situations
outside the household as well. The older siblings are expected to look after
their juniors at all times, and negligence is met with severe disapproval by Mr.
und Mrs. H and the rest of the older siblings.
An examination of Table 2 reveals that adoption transactions
urc usually carried out between matrilateral kin. In the grandchild
category, for example, daughters' children outnumber sons' chil-
dren by l 0 l to 28, while in the other categories ma trilateral kin
outnumber patrilateral kin by 65 to 19. If we assume that the
degree of formality in relationships is inversely correlated to the
strength of kinship ties, the weakness of patrilateral links is re-
vealed by the types of adoption represented in the matrilateral and
patrilateral categories. Among patrilateral kin, 10 out of 19 chil-
dren (52.6 percent) were legally adopted, while among matrilateral
kin 17 out of 65 (26.2 percent) fell into this category. In the
grandchild category, where one would expect relationship ties to
be strongest, only 21 out of 170 adoptions ( 11.8 percent) were
legalized. The data also show a tendency for individuals, when
adopting in the child generation, to take children from a sibling of
the same sex. Thus female adopters took 21 children from their
sisters and only 9 from their brothers, while male adopters took 7
children from their brothers and 5 from their sisters.
The preference for adopting children at birth or shortly there-
after is still pronounced among our contemporary informants. The
comments of Mrs. Pare typical: "A friend of mine separated from
her husband and wanted to give me her first boy. I didn't want to
because he was already five years old, and at five he understood
the ways of his mother. If she had given him to me at one year it
would have been all right, but at five he would cry and get sick
thinking of the mother."
Although this sentiment is widely shared, there are cases in
which older children are taken as hanai without hesitation, parti-
cularly if the child is willing. An example is provided by Mrs. S:
Valerie is Mrs. S's brother's daughter. She is now six years old. Her
mother and father were divorced, and Mrs. S's mother came and
took all the children to Hilo. Then about two years ago she came
to Honolulu and brought Valerie with her. "She was a shy child
... she didn't talk too much, but you became attached to her
after a while." Valerie did not want to go back to Hilo because she
was afraid of airplanes. "We asked her if she wanted to stay and
she said yes, so we kept her."
Mrs. S is also taking care of her sister's son, whom she dis-
covered sleeping on the beach after he had run away from home.
Nevertheless, it is clear from the data in Table 3 that the large
TABLE 3 Ages of Children Taken into Household
Nanakuli Waimanalo Kewalo Papakolea Total %
Legal adoption
birth 13 3 5 0 21 25.6
l wk. to 6 mos. 19 3 15 0 37 45.1
6 mos. to 1 yr. 2 0 5 8 9.8
more than 1 yr. 6 l 7 2 16 19.5
Total 40 8 27 7 82 100
birth 20 8 3 0 31 32.3
l wk. to 6 mos. 19 7 7 4 37 38.5
6 mos. to 1 yr. 9 0 0 2 11 11.5
more than 1 yr. 7 3 5 2 17 17.7
Total 55 18 15 8 96 100
birth 33 11 8 0 52 29.2
1 wk. to 6 mos. 38 IO 22 4 74 41.6
6 mos. to 1 yr. 11 0 7 19 10.7
more than 1 yr. 13 4 12 4 33 18.5
TOTAL 95 26 42 15 178 100
1 nformation was not available as to when foster children entered these households.
majority of hanai cases involve infants. In our homestead sample
29.2 percent of the children who had been legally adopted or
taken in hanai were taken at birth, and an additional 41.6 percent
were taken before the child was six months old. Only 18.5 percent
were taken after the age of one year.
The principle that the first-born male child belongs to the
fother's side and the first female child to the mother's side is less
in evidence among modern Hawaiians. Perhaps this is because the
rule of alternate sexes was highly specific. It may well have been
lost, therefore, in the breakdown of Hawaiian custom, that is, the
failure to transmit specific decision-making principles from gener-
ulion to generation. The remarks of Mr. H are indicative: Mr. H
TABLE4 Sex of Adopted Children
Legal Adoption Hanai Fosterage Total
Community Male Female Male Female Male Female Male Female
Nanakuli 26 14 33 27 27 31 86 72
Waimanalo 7 3 7 11 18 15 32 29
Kewalo 14 16 7 8 23 25 44 49
Papakolea 4 3 5 4 2 3 11 IO
TOTAL 51 36 52 50 70 74 173 160
% 58.6 41.4 51.0 49.0 48.6 51.4 52.0 48.0
said that a grandchild he wanted belonged to him, no matter what
sex it was. The wife's parents had no rights to the first child. He
said that he did not know if it was old tradition, but he thought
the males of the family were supposed to determine who had the
rights to adoption.
Our data in Table 4 on the sex of adopted children indicate
only a slight preponderance of males (52.0 percent). It should be
noted, however, that in the category of legal adoptions males are
more strongly represented (58.6 percent). This may be a reflection
of the importance attached by most Hawaiians to "passing on the
One very noticeable trend stands out in our investigations into
modern Hawaiian adoption practices, namely, an increasing con-
cern for legalizing adoptions. Of the 334 adopted children in our
homestead sample 88 (26.3 percent) were legally adopted. While
this tendency is most pronounced when the foster child is unre-
lated (37 .5 percent), even close relatives are apt to ask, or in some
cases demand, permission to legally adopt hanai children. This is
sometimes motivated by a desire for legal protection of the child
with regard to inheritance, but by far the most frequent reason
given is a fear that the natural parents will take back the child. Our
informants have provided us with numerous instances in which
such conflicts have arisen, and they invariably reported them with
great emotion. Here are some examples:
Mrs. N said that her hanai son is part Negro. His father used to work with
Mrs. N. When the boy was born his parents came to her and asked her to be
his godmother. Shortly after the child was baptized his parents left for Japan,
and they left him in Mrs. N's care. She took care of him for four years, with-
out receiving any support from the baby's parents. At the end of four years
the parents returned to Hawaii and asked for the child back. She told them
that she loved him and wanted to keep him; that, after all, they had not even
provided support for their own baby. The parents did not agree and brought
the case to court. Although it cost her five hundred dollars for an attorney,
Mrs. N fought the case and won.
Mrs. A and Mrs. B were neighbors. Both had inherited their homesteads from
their parents, who were close friends. When Mrs. A was married, at about age
twenty-one, Mrs. B was already middle-aged. Her children were all grown, and
she wanted more in her home. Mrs. A and her husband were just starting their
family and were experiencing financial difficulty. When Mrs. A had a child
she gave it to Mrs. B to care for, at Mrs. B's request, but when the next one
came along Mrs. A took back the first child and gave the baby to Mrs. B. This
was repeated with the next child, but following that Mrs. B refused to do it
again. "Three times my heart was broken." She is now considering taking a
hanai from another family.
Mr. S's sister has taken two of his children as hanai. First she took at birth
one of his twins, now age sixteen. She had one child of her own at the time,
but it was Mr. S's seventh child. She asked S for the baby when his wife got
pregnant, and he told her that it was all right with him but that she had to ask
his wife. She did, and his wife agreed. When the day of birth came, there were
twins and his sister wanted both, but S refused. He said, "God must have
meant one for each of us." His sister went to a lawyer without S knowing it
and started procedures for legal adoption, but when she brought back the
papers he refused to sign. She said that S might try to take back the child
after he had grown up, but S told her he would not-she did not have to
worry about that. Then he asked her what would happen if she and her hus-
band died or were incapacitated-"I've got insurance for them, but you don't.
What if you get sick and can't support them? I love my children and want to
be sure they will always be taken care of," he told her.
In 1946 Mrs. C took two children as hanai, a two-year-old girl who was her
namesake and the girl's brother who was still an infant. The children's mother
died shortly after the boy was born. According to Mrs. C, the mother had
Nuspected that she was going to die and wanted the children to go to Mrs. C,
who had not seen the woman for some time and was told about her death by
the girl's godfather. He asked Mrs. C to take the girl. Shortly after that the
children's father came to Mrs. C and asked her to take the boy also, and she
uareed. She said that for the first year their father contributed twenty-five
dollars per month for their support, after that ten dollars per month for an-
other three or four years, and then he refused to contribute anything. As a
result, Mrs. C filed for legal adoption and the father fought it, so they
brought the case to court, where Mrs. C won after a four-hour hearing.
These cases illustrate the great attachment to hanai children
that is characteristic of Hawaiian parents. Our data indicate that
under most circumstances hanai children are treated with some-
what more indulgence than natural children. When there are both
hanai and natural children in the household, however, a pro-
nounced egalitarianism is usually espoused. All those informants
who had assumed full parental rights insisted that they loved all
their children equally and that hanai children were entitled to
their share of the inheritance. Thus far we have insufficient evi-
dence to either support or repudiate this claim to equality of in-
heritance, but "playing favorites" is considered a very bad thing
for parents to do, and, although emotional commitments to dif-
ferent children within a household may vary, most parents make a
strong effort to see that no one child gets substantially more or
less material benefits than any other.
The strength of adoptive ties, even under rather grave economic
hardship, is illustrated by two cases reported by Beaglehole
One case recorded is that of a crippled husband, seventy-two years old, and
his wife, seventy-one. With the old couple live two hanai children, female sib-
lings aged six and five respectively, daughters of the husband's niece. This
niece has a family of 11 children. She is a school teacher. She refuses to sup-
port her uncle because of a quarrel. The uncle, because of his desire for chil-
dren, refuses to allow the two girls to return to their blood parents. The
family is living in great poverty and the two girls have health problems. The
old woman owns a small property of doubtful economic value. She could
obtain a government pension if she would allow the territory to take out a
lien on this property. She refuses to do this because of her desire to save the
property and pass it on to her two foster children when she dies.
Another case is that of a very old pure-Hawaiian woman who has adopted as
a hanai her great grandchild. a boy of mixed Hawaiian, Chinese and Cauca-
sian blood, 14 years old. The old woman receives a pension of fifteen dollars
per month of which five dollars are paid out in rent. The boy is clever, but
beyond the control of the woman. He shines shoes, sells newspapers, but is
rapidly becoming a social misfit, having already appeared several times in the
juvenile court for being a truant. The old woman feels that she must have this
boy or other children with her to be happy even though she has no means to
support children.
The terminology used by adopted children to refer to their fos-
ter parents varies with the circumstances. If the adoption took
place at birth or shortly after, the common English parental terms
-'ma', 'mom', 'mama', 'dad', 'daddy', etc.-are applied, although
if the adopters are grandparents they are still usually called 'grand-
ma' and 'grandpa'. When adults speak about relatives by adoption
they most often use the same terms as they use for consanguineal
relatives, although they may qualify relations at times by saying
such things as, "my hanai mother" or "my uncle on my hanai
father's side." Such qualifications are likely to be used only if the
person speaking has maintained relations with his natural parents
or their relatives. The degree to which this occurs generally de-
pends upon residential proximity more than any other factor.
When the natural parents live near the adopters, the foster children
may have nearly equal contact with both family groups. Both sets
of parents sometimes exercise parental rights, feeding the child
and disciplining him. This is known to have become a source of
conflict between such families, as one might expect. When both
sets of parents are socially present the child often uses two sets of
contrasting terms to refer to them, perhaps calling one set some-
thing like 'mom' and 'dad' and the other 'mama' and 'papa'. When
the natural parents live far away or otherwise do not maintain
their ties with the adopting family, they are likely to become no
more than genealogical shadows for whom the child has little re-
gard. What is important under most circumstances is the social and
emotional ties between individuals, rather than the specific nature
of kinship links (natural or adoptive) between them. In fact con-
siderable confusion about the nature of their kinship ties to one
another often arisl(s in the minds of individuals involved in adop-
tive relationships. This can be seen in the attempt of a woman to
describe her relationship to her hanai daughter:
Kealoha is my brother's child. Of course my brother isn't really my brother as
both he and I are hanai children of my father. I guess my father isn't really
my father, is he? I know who my real mother is but I didn't like her and I
never see her. My hanai brother is half-Hawaiian and I am pure Hawaiian. We
aren't really any blood relation I guess, but I always think of him as my
brother and I always think of my father as my father. I think maybe Papa is
my grandfather's brother; I am not sure as we never asked such things. So I
don't know what relationship Kealoha really is though I call her my child
(Hormann 1960: 13-14).
We have thus far discovered few consistencies in the attitudes
toward their natural and foster parents of adults who grew up as
hanai children. The whole range is represented, from people who
as adults are very close to their foster parents and completely alien
to their natural parents to the reverse. Much depends upon the
vicissitudes of residential proximity, economic circumstances, and
interpersonal conflicts and crises. A tentative generalization, how-
ever, is that whereas Hawaiians do not seem to feel compelled to
honor biological relationships for their own sake they usually do
feel a sense of obligation toward their foster parents for having
raised them. An illustration of this is the case of Jane, who, in
spite of a deep-felt desire for a home of her own, continues to live,
along with her husband and four children, with her hanai parents
and to provide a large part of their support. She remains there, she
says, because she has a binding sense of obligation to take care of
them, since they took care of her when she was young.
Several motives seem to affect the decisions of present-day par-
ents to allow their children to be adopted by others. These include
some of the traditional reasons plus a few new ones. The fear of
harm to the child through sorcery if a request to adopt is refused
is not as strong or as prevalent a motive as it once was, but it is
still present. Thus some of our informants have indicated that
would-be adopters, if refused, might bring harm to the child
through sorcery as a result of their envy or jealousy. Beaglehole
noted such a case:
The old man P is indigent. He has quarrelled with his daughter who has re-
fused him economic support and has also refused to allow the old man to
take his eldest grandchild as a hanai. The granddaughter became sickly and
the daughter firmly believed that her father was a kahuna 'sorcerer' who had
bewitched her small child and was making it impossible for her husband to
obtain work. Some months later the sickly child died. The daughter became
unbalanced mentally through fear of her father's power and consequent
worry, and her children had to be removed from the household for their own
safety (I 939:67-68).
Related to this concept is the idea that harm may come to a
child over whom there is ill feeling, not as a result of intentional
sorcery, but from the presence of the hostile feeling itself. The
same sanctions also operate against attempts to take back a child
who has been taken in hanai. Evidently it is not the regaining of
the child that threatens disaster, but the act of taking him from
the foster parents. Once the foster parents die, or the child has
been taken from the foster parents by a third party, the natural
parents may take their child back without fear of harm:
An old indigent grandmother was physically neglecting her two young
adopted granddaughters. The mother of the two children refused to take
them back and give them proper care for fear of the old lady's magic powers.
The mother related that the old lady took as an adopted child one of her girl
children. Later she asked for a boy when it was born. The mother refused the
old lady telling her that she neglected the first child. The old lady argued that
the mother had both boys and girls while she had only a girl. She also wanted
a boy. The boy became ill when it was six months old. The mother knew that
there was nothing physically wrong with the infant but that its sickness was
caused by outside influence. She went to see the old lady, explained that she
should have given over the boy and now she would give away her next boy if
the old lady would see to it that the ailing infant got well. The infant re-
covered immediately and the old lady received the next boy as her child.
Since then no one has been sick in the mother's family. The mother wished
she could take back her children and give them proper care. Under the cir-
cumstances, however, she did not dare to do this. She felt that if a social
worker removed the children from the old lady and then later turned them
over to her, their own mother, this would give the old lady no excuse to be-
witch the mother or her family (Beaglehole 1939:58).
A second form of sanction, more secular in nature, is the ex-
pressed feeling that anyone with many children, especially many
young children, has an obligation to "share the wealth" by giving
1.:hildren to those relatives without children who want to hanai
them. In talking about attempts to adopt, remarks like "She has
six children," or "This is her eighth child," are frequently made
with reference to the natural mother, carrying the implication that
unyone so blessed can certainly afford to give away one child. A
family with many children which refused to give one to a close
relative without any would be accused of being stingy. Related to
this is the belief that the adoption of a child by a childless couple
promotes fertility and thereby permits them to produce children
of their own:
Mrs. N said that she gave her six-month-old daughter to one of her sisters. Her
Nislcr had no children at the time. At first it was hanai but the sister's hus-
band wanted to adopt the child "in black and white," so they let him. Then
the sister started to have her own children, and now she has four boys. In
response to a question about the relationship between the adoption and the
sister's fertility, Mrs. N broke in and said, "That's what did it. It's an old
legend, called ho 'opili" (literally, 'To bring together, stick; to attach oneself
to a person'[Pukui and Elbert 1957:303)).
Perhaps the most interesting question that emerges from the data
is why, among a people whose culture seems to have been other-
wise completely shattered, has the prevalence of adoption so
dramatically survived? If there were political reasons they are gone
now, and it does not seem to make much sense to search for social
systemic explanations when the majority of Hawaiians now live in
American-type communities. Economic explanations are likewise
unconvincing. The value of children for old-age insurance is as
questionable today for Hawaiians as it is for other Americans. On
the other hand the economic burden of raising children has been
vastly increased in a wage economy. Furthermore, the willingness
to pay adoption costs, often by people who could be classified as
indigent, suggests that economics is not a primary consideration in
most instances. This is not to deny that economics plays a role in
particular cases. Thus a person may offer to take a relative's child
in order to relieve the weight of poverty rather than from a strong
desire for the child, or a child of at least half-Hawaiian ancestry
might be adopted by someone wishing to pass on his homestead
land to someone in the family;
but our evidence strongly suggests
that most adoptions occur in spite of economic considerations
rather than because of them. What then is the explanation?
The answer appears to lie in the continued significance of nur-
turance as a motivational force among Hawaiians. Within the tradi-
tional culture nurturance was an institutionalized value around
which a substantial body of custom was built. People were ex-
pected to take care of those in need, and the strong were expected
to assist the weak. Prestige accrued to the generous. Today the
value is still expressed in the guise of the commercially exploited
"Aloha Spirit." Our thesis is that, despite the elimination of sup-
portive custom, nurturant behavior has been perpetuated among
lhe Hawaiian population by virtue of a widely shared child-rearing
pattern. The key features of this pattern and its motivational con-
sequences are presented below.
During the first two or three years of life, Hawaiian children are
greatly indulged. They are continually fondled, their needs are
instantaneously responded to, and they are rarely disciplined. As
children become more mobile and more verbal, however, or when
u subsequent infant enters the household, their demands for atten-
tion are increasingly rebuked and their intrusions are rejected.
Parental pressure toward independent behavior begins early;
11kills are not purposely taught, and there is little positive rein-
forcement for correct behavior. Instead, an older child's activities
ure virtually ignored unless they irritate adults, in which case pun-
ishment-spontaneous and unpredictable-is likely to be the only
response. The most salient feature of this training sequence is the
~ r t y withdrawal of nurturance coupled with punishment for de-
pendency. This early indulgence and subsequent rejection is very
much like the patterns described by Ritchie for New Zealand
M:1ori (1963: 127-141) and Levy for the Tahitians (Levy 1968). It
results in powerful dependency needs which are inhibited as a con-
Neq uence of the punitive training. The maturing child discovers
that dependency overtures frequently provoke punitive responses,
1uu.l he becomes increasingly anxious in situations where the grati-
fication of his needs is controlled by others. He learns to avoid
Interpersonal involvements in which dependency strategies are re-
As boys grow older they spend an increasing amount of time
11way from the household and in the company of peers. They may
"ul and sleep at home but otherwise have only marginal contact
with the adult members of their household. Their relationship
with peers is, as one would expect, intense and unstable. That is,
they tend to form very tight friendships that are easily broken
when one of the parties fails to meet the other's dependency de-
111u11ds. Girls, on the other hand, spend a great deal of time at
home assisting with housework. They are socialized into the
housewife role at a very early age, and it is the only role in which
they are consistently encouraged. Their responsibilities include
curing for younger siblings, a role in which they control the pre-
cious resources of nurturance. Reinforcement of nurturant be-
havior comes not only from the parents, but from the infants as
well. It is the responsiveness of their wards at a time when they get
so little from adults that provides the major source of inter-
personal gratification for young Hawaiian girls. Babies-dependent
and responsive-therefore become identified as the only human
objects that provide an unthreatening opportunity for intimate
emotional exposure. With adults, or even with older children, the
expectations of punishment so exceed those of reward that it is
risky to become involved; with them one must always be on guard,
ready to disengage.
It is, therefore, not surprising that as adults Hawaiian women
are highly motivated to have children, particularly very young chil-
dren, in their homes.
We would suggest that the social psycho-
logical dynamics are as follows: as a result of the reinforcements in
childhood for playing the nurturant maternal role, Hawaiian
women develop a strong need for babies, both as a source of per-
sonal gratification of intimacy needs and as a means of validating
their adult roles. As babies grow into children the dependency
behavior which adults have nurtured by their indulgence, and
which is now reflected in whining, clinging, willfullness, etc., be-
comes increasingly burdensome. Much of this behavior is regarded
as disrespectful and disobedient, and it is severely punished. In
addition, as children begin to develop autonomy and assume a
capacity to reject overtures of nurturance and affection, their
parents' intimacy anxieties are activated, and emotional disengage-
ment from the child is a frequent response. Parents begin to ignore
the child, become increasingly punitive, and avoid praise or other
overt expressions of affective commitment which would increase
their vulnerability. This in turn leads young girls to derive their
major interpersonal gratifications from younger children with
whom they can be nurturant, and so the cycle is repeated. It also
leads mothers to seek new infants to replace maturing children;
and if a woman does not give birth to another baby by the time
her youngest child ceases to be a source of fulfillment for her nur-
turance needs she will often seek a replacement through adoption.
Although men are not so directly socialized into the nurturant
role, they, too, find babies to be the safest creatures in an unpre-
dictable world and are motivated to have some in their house-
The Hawaiian subculture still emphasizes cooperation rather
TABLE 5 Age Positions of Adopted Children in Household at Time of Adoption*
Legal Adoption Hanai Fosterage Total
Community Youngest Other Youngest Other Youngest Other Youngest Other
Nanakuli 26 14 44 13 38 20 108 47
Waimanalo 7 3 16 0 21 12 44 15
Kewalo 28 1 10 2 37 11 75 14
Papakolea 5 2 4 0 3 2 12 4
TOTAL 66 20 74 15 99 45 239 80
% 77.2 22.8 83.1 16.9 68.8 31.2 74.9 25.1
*It was assumed that each adopted child who was younger than any natural child in the same household was the youngest child in the house-
hold at the time of his adoption. For example, if a household contained two adopted children, ages 8 and 10, and three natural children, ages
12, 15, and 17, then the adopted children were both counted as "youngest."
than competition, and it is within such a value structure that the
above developmental pattern lies embedded. Generosity and nur-
turance remain uppermost in the hierarchy of desirable personal
attributes, and a person still gains status for taking care of depen-
dent others, whether infant, child, or adult. It seems evident,
therefore, that the act of adopting combines personal gratification
with social reward in a compelling fashion for most Hawaiians.
Since love for children is also highly valued in the dominant
American culture, a further impetus is no doubt given to the prac-
This general interpretation is lent support by the homestead
data presented in Tables 5 and 6. They show that 74.9 percent of
the adopted children occupy the position of youngest child in the
household, and that 78.4 percent of the female adopters were
TABLE 6 Female Adopter's Age at Time of Adoption
Type of Adoption
and Age of Adopter Nanakuli Waimanalo Kewalo Papakolea Total %
Legal adoption
less than 25 yrs.
25 through 35
more than 35
less than 25
25 through 35
more than 35
less than 25
25 through 35
more than 35
less than 25
25 through 35
more than 35
9 16.4
10 18.2
36 65.5
5 8.1
10 16.1
47 75.8
2 2.7
5 6.8
66 90.4
16 8.4
25 13.2
149 78.4
older than thirty-five years of age when they took adoptees into
their households. It is also supported by a great deal of psycho-
logical data that we have collected by observation, interviews, and
social psychological experimentation during the past two years.
We suggest that such an explanation may also hold for the New
Zealand Maori and the Tahitians, both of whom have high adop-
tion rates, and perhaps for other Polynesian and Micronesian
populations with a similar social-psychological syndrome.
This appears to have been only one of several meanings for the concepts.
The term ho'okane is translated by Pukui and Elbert 'to behave as a male';
'masculine' (I 957: 119) and ho'owahine 'to behave like a woman', 'to imi-
tate the ways of a woman'; 'to grow into womanhood'; 'to have the
manners and ways of a lady'; 'to become a wife'; 'to obtain a wife' (rare);
'to take as wife'; 'feminine' (1957:349).
2 In this chapter the term "Hawaiian" is applied to that segment of the popu-
lation of the State of Hawaii that traces some ancestry to Polynesian
Hawaiians and who identify themselves ethnically either as Hawaiian or
J In order to inherit homestead land through the offices of the Hawaiian
Homes Commission a person has to be able to demonstrate that he is at
least half Hawaiian by genealogical ancestry. Some of the lessees, while
they themselves claim to be half or more than half Hawaiian, are married
to non-Hawaiians, and their children are ineligible to inherit the property.
Upon the death of the lessee, his family would have to move off the home-
stead land. One way around this is for such a couple to legally adopt an
eligible child from a close relative, with the purpose of keeping the prop-
erty in the family. When the foster child inherits title to the homestead
land, he is expected to allow his adopted siblings and other members of
the household to remain there, and thus they do not have to give up their
4 A sample of 31 Hawaiian mothers was interviewed to determine the ages at
which they expect or allow children to act in certain independent ways.
The median ages at which the Hawaiian mothers expect independent
behavior from their children on 17 out of 22 items are from one to two
years younger than a mainland Caucasian sample.
5 The median age at which Hawaiian mothers considered children "most
enjoyable" was six months. They frequently state that when children are
older they start going off by themselves and "don't need you as much."
Antony Hooper
Recent ethnographic studies of a number of communities of indig-
enous Polynesians in the Society Islands have shown that, in spite
of the wide variety of social and economic conditions which pre-
vail in the group, adoption is practiced in all of them. It is as com-
mon among the urban wage-laborers of Papeete as it is in the
peasant communities of the outer islands.
Many of the modern Tahitian
beliefs and practices concerning
adoption are, or were at one time, widespread in Polynesia.
children adopted are commonly close kin of the adopting parents,
their social and emotional ties with their natural parents are not
unequivocally severed, and the transaction is a highly contingent
one. As among the modern Maori (Hohepa 1964:37-40; Metge
1964:27) and the Cook Islanders (Beaglehole 1957: 164-66), the
system flourishes and persists quite apart from legal contracts and
In describing this system, I find it useful to make a distinction
between the strictly ideational aspects of modern Tahitian life and
the statistically established patterns of social action in various
interaction units. I have in mind the kind of distinction formu-
lated in slightly different terms by a number of writers. Oliver
(1958:803-804) for example, distinguishes between three "kinds"
of interaction. Those labeled "normative" and "suppositional" fall
within what I choose to call the ideational sphere; by contrast, his
"historical" kind of interaction is more directly subject to empiri-
cal observation. Geertz, following Parsons and others, has phrased
the distinction between "culture" and "social structure" as fol-
lows: culture is "the fabric of meaning in terms of which
human beings interpret their experience and guide their action."
Social structure is "the form that action takes, the actually exist-
ing network of social relations" (1957:33). Goodenough makes a
similar distinction between the anthropologist's model of a com-
munity's generalized culture (or the public culture) and the social
systems which "we observe in operation" and which are "created
and maintained as products or by-products of culturally guided
human action, and, as such, are artifacts of culture" (1963:
263-271 ).
I find this distinction useful because the ideational aspects of
the adoption system are common to all modern Tahitian commun-
ities. Members of any one community can (and a great number, in
fact, do) move freely and easily to another community knowing
"whatever it is one has to know or believe in order to operate in a
manner acceptable to its members" (Goodenough 1957: 167). On
the other hand, observation shows that communities differ in the
frequency with which adoption is practiced and in the frequency
with which different types of kin are adopted. Comparison of the
different communities thus allows us to isolate with some cer-
tainty the causes of this variation and to establish some of the
social effects of local variations in ecological setting, economic
opportunity, and demographic history.
1 shall describe the Tahitian norms, values, and expectations-the
"public culture" -concerning adoption, and then deal in some de-
tail with the adoption system in the community of Maupiti. Data
from a second rural community, Murifenua, will be used for com-
parative purposes.
The two communities with which we are concerned are located in
the administrative circonscription 'division' of the Iles sous-le-vent,
which embraces the five volcanic islands and four atolls lying to
the northwest of Tahiti. Together, these islands have a population
of a little over sixteen thousand. Outside the small administrative
and commercial center of Uturoa, at the northern end of the
island of Raiatea, the great majority of the inhabitants of this area
are indigenous Polynesians who live in small communities spread
out along the lagoon shores and narrow coastal plains of the larger
Although they are relatively remote, these rural communities
cannot legitimately be regarded as isolates. Two centuries of inten-
sive European contact have transformed Tahitian society and
culture. It is now well over a hundred years since the missionary
church effectively replaced the native religious organization and
much of the indigenous belief system. The whole of the intricate
native polity, which was built on a tribal system, and the concepts
of inherited rank and mana have completely gone. In even the
most remote communities, family and domestic economies involve
the production of cash crops and the purchase of building ma-
terials and foodstuffs (and outboard engines and motorcycles)
through a network of Chinese storekeepers and entrepreneurs
spread throughout the islands. To perhaps a greater extent than in
any part of island Polynesia except Hawaii, the indigenous popula-
tion has been gathered into the institutional structures of the ad-
ministering colonial power.
The Society Islands are an Overseas Territory of France. Tahi-
tians are French citizens, subject to French law. The rural
areas are divided into a number of administrative divisions in
which executive authority is in the hands of appointed officials.
Circonscriptions are divided into districts, which are the smallest
administrative units. In a district, administrative duties of various
kinds such as the registration of births, deaths, and marriages are
carried out by a salaried chef 'chief, whose position is achieved by
local election. He may be assisted by the five elected but unpaid
members of the conseil du district 'district council' and the sala-
ried mutoi, or 'local policeman'. The duties of a chef are not partic-
ularly onerous. 1 he position involves responsibility and some pres-
tige, but very little authority.
Through a system of universal adult suffrage, Tahitians vote for
the members of a Territorial Assembly which has wide powers for
the framing of local legislation. Political parties contest these elec-
tions, and their candidates visit and campaign periodically through-
out the rural areas. Most districts have factions which are linked to
various elected members of the Assembly or to the political party
organizations in Papeete, but these have to contend with a certain
amount of Tahitian indifference and cynicism about political ac-
tivities at the "national" level.
Some 80 percent of the population are members of /'Eglise
Evangelique de Tahiti, the independent Protestant church which
has developed out of the London Missionary Society's mission to
the islands. In the rural areas, parish and district boundaries are
coterminous. Each parish has a church and a number of small
meetinghouses for local religious groups (known as pupu) which
are scattered throughout the communities. Parish ministers
(mostly Tahitians or part-Tahitians trained by the Church in
Papeete) have a great deal of prestige, and the local religious parish
and pupu organizations play an important part in community
affairs. Cosmological ideas are derived wholly from the Tahitian
Bible and from Church teaching. They are neatly integrated with a
complementary system of indigenous beliefs about the efficacy of
the spirits of the recently dead. Native mediums are frequently
consulted, and there is an elaborate system of indigenous medical
practices of various kinds.
Education is free and compulsory up to the age of fifteen. The
administration-run schools in each district are staffed by Tahitian
or part-Tahitian teachers trained in Tahiti. French is the language
of instruction at all levels, but, in spite of this, Tahitian remains
the language of everyday use and of all religious and political
gatherings and affairs in the rural areas.
For the most part, rural land is owned by Tahitians. When the
Iles sous-le-vent came under effective French control at the begin-
ning of this century, land boundaries and claims of ownership
were codified. Since that time, the Tahitians' rights in land and
other immeubles 'real estate', and the manner in which these rights
may be transacted between individuals and other legal "persons,"
have been regulated according to French law. This has allowed for
land rights to be bought and sold, gifted, leased, exchanged, and
conferred by testament and succession. Although a certain amount
of rural land in the Iles sous-le-vent was bought and sold during
the earlier part of this century, most of the named sections of land
have been passed on according to the rules of succession, which
provide for the land rights of an intestate person to be transferred,
at his death, in specified proportions to various categories of le-
gitimate and natural consanguineal kinsmen. Legitimate children
succeed to the total estate in equal shares without regard to either
sex or birth order.
Rural Tahitians have not, in most cases, made testaments, with
the result that their land rights have been passed on to their chil-
dren in equal portions, by the laws of succession. This process has
not, however, involved the partition of the actual named sections
of land delineated and codified during the early years of this
century. The land divisions have remained, still in most cases
marked by the same boundary markers that were set up sixty-five
years ago. But the rights in these tracts have been divided, and
each section now has a large number of co-owners-often an ex-
tremely large number. Certain individuals from each body of co-
owners live on (or near) each section of land, and it is they who
are involved in its exploitation. Other co-owners live elsewhere,
either in other rural communities in which they or their spouses
have land rights, or where wage-labor is available.
To the extent that co-owners of land all trace descent from a
common ancestor (that person to whom ownership rights were
first granted according to French law) they may be said to form a
descent unit of some kind. Indeed, Tahitians themselves conceive
of bodies of co-owners in this manner and frequently name them
by reference to the common ancestor. Thus the Tahitian, and
widespread Polynesian, ideology linking land rights with member-
ship in a descent unit is maintained within the letter, if not the
spirit, of French civil laws.
The Tahitian term which I gloss as 'adoption' is fa'amura'a;
adopted children are termed tamari 'i fa 'amu 'children who are
fed'. When genealogical or biological connections are relevant,
adopted children may be distinguished from natural children,
termed tamari'i mau 'true children' or 'real children' or tamari'i
fanau 'born children'. In other contexts, both may be simply re-
ferred to as tamari 'i.
Adopted children live in the household of their adoptive par-
ents, who supply the children's food and other physical needs.
However, since Tahitian children not uncommonly live with adults
who are not their natural parents, it is important to distinguish
those who are tamari'i fa'amu from those who are not regarded as
such. The following examples, recorded during the course of a
household census, indicate some of the relevant contrasts that are
made by Tahitians:
A ten-year-old girl lived with her mother's half sister, a neighbor, for a period
of over six months. The girl helped care for young children while her aunt's
husband was in the hospital. The relationship was not one of fa'amura'a but
was characterized by the term tauturu 'helping'.
A six-year-old boy lived for over three months in the household of his
mother's father's sister's husband. He had run away from his own home when
he was severely punished by his stepfather, a man widely regarded in the com-
munity as 'bad'. He was found on the village road with weals on his body and
was taken to his relative's home and fed, housed, and cared for. Although
there is no indication that he will ever return to his mother and stepfather, he
is not regarded as an 'adopted child'; he was characterized simply as mea
taponi mai 'a thing hiding here'.
A four-year-old girl was left in the care of her unmarried mother's adoptive
father's adoptive son while her mother spent a year working as a waitress in
Papeete. The mother sent the payments of child allowance, which she re-
ceived from her employer, to those who were caring for the child. She in-
tended to return to the household herself, and the transaction concerning the
child was regarded as one of ha'apao 'looking after'.
A six-year-old boy lived with his mother's mother's husband's family for
three days, "because he enjoys it." He was mea haere noa mai 'a thing which
just came'.
Three siblings between the ages of seven and fourteen lived with their moth-
er's mother for fourteen months while their parents were working in New
Caledonia. The grandmother in this instance was regarded as being there to
ha'apao 'look after' the children, who would eventually be returned to their
natural parents.
These examples show clearly that, notwithstanding the length
of time which some children may spend living in the households of
people other than their natural parents, they are not for that rea-
son necessarily regarded as tamari'i fa'amu; they may be there be-
cause they are 'helping' or because they are 'hiding', being 'looked
after', or simply because they 'enjoy it'. On the other hand, there
are certain (admittedly unusual) circumstances under which an
adopted child may continue to live with his natural parents, while
still being widely regarded as the adopted child of somebody out-
side the household. One thirteen-year-old boy lived with his nat-
ural parents even though he had been adopted by an elderly
widow who lived alone nearby in a house built for her by her
neighbors, her kin, and her adopted son. Had the boy been mar-
ried it would have been his duty to take his adoptive mother into
his household.
Children for adoption should be taken from among the feti'i
'kindred' and not from those i rapae 'outside'. Children, like land
and personal names, are regarded as resources to which rights are
held by categories of people claiming a common ancestor. These
rights may or may not be used, depending on circumstances; but
to adopt a child from outside one's own or one's spouse's kindred
group is regarded by feti'i as a personal affront, since it carries the
implication that the group's resources of children are insufficient.
Kindred members are also obliged either to adopt children who are
left homeless, or to care for them for extended periods. As with
incest prohibitions and the obligations and rights to hospitality,
these obligations fall upon genealogically close consanguines more
heavily than they involve distant kin, in keeping with Tahitian
conceptions of cognatic kinship, which is calculated rather pre-
cisely in degrees of toto, or 'blood'.
Adopted children may not have sexual relations with any of
their natural kin within the prohibited degrees; to do so would be
to amu to to 'eat [one's own] blood'. It is maintained also that
adopted children should not have sexual relations with the natural
children of adoptive parents, since even if those children do not
fall within the prohibited degrees, they have been raised together
and are 'like siblings'; the same strictures apply to unrelated chil-
dren adopted by the same parents. I learned of one instance in
which the natural children of two unrelated women who had been
adopted by the same parents were allowed to marry, even though
there were some who disputed the propriety of their action. Sex-
ual relations with the adoptive parent who is not a consanguine
would not be regarded as incest but merely as 'disgusting'.
Children, even in large numbers, are valued resources, objects of
pride, love and affection. Childlessness is regarded as a serious
matter, both because of the loss of the pleasures of handling and
caring for children, and the prospect of an old age without care
and support. While it would be regarded as somewhat cynical to
suggest that the main reason for having children is to assure a com-
fortable old age, Tahitians explicitly and openly regard the care
which is given to children as incurring an obligation upon them to
pay back that debt in their later years by caring for their parents.
This debt falls equally on natural and adopted children.
The linguistic form tamari'i fa'amu belongs to a set of which the
only other members make reference to various domesticated and
captive animals:
hono fa'amu
pu 'ahorofenua fa 'amu
a 'pig [which is] fed'
a 'turtle [which is] fed'
a 'horse [which is I fed'
By far the most commonly used of these is pu 'a fa 'amu. It refers
to a transaction whereby a person contracts to feed a domestic
sow belonging to another person. When the sow farrows it is then
returned to its owner, together with a proportion of the litter,
while the person who has fed the pig keeps the remainder for him-
self. The transaction thus has certain similarities to the adoption
of children; the parallels are regarded as humorous and not vigor-
ously denied by informants.
Another commonly advanced reason for having children is to
provide mono 'substitutes' or 'heirs' who will bury their parents
correctly and make use of their house and land. For those who
have no natural children, adopted children are the universally pre-
ferred alternatives. Unmarried persons,
either male or female,
may adopt children.
It is expected that adopted children will be accustomed to life
with their adoptive parents, and that they will be bound emotion-
ully more closely to them than to their natural parents. These
expectations are based on the concept of matau 'familiarity' which
IN held to be an essential ingredient of all easy social relationships.
Older adopted children, and adults who were adopted as children,
feel 'ill-at-ease' with their natural parents when they have been out
of touch with them for long periods, perhaps through having been
adopted to another community. In cases where the natural and
adoptive parents live in the same neighborhood, adopted children
may be equally at home and at ease in both households, though
their filial obligations are always held to lie with their adoptive
parents, in repayment of the care given to them. There is, how-
ever, general agreement that adopted children, like natural chil-
dren, do not always fulfill these obligations.
Parents do not have the right to expect filial services from their
natural children who have been adopted. Adoptive parents do not
have the right to expect any special assistance from the child's nat-
ural parents. Neither sex is more highly preferred for giving or
receiving in adoption. The main consideration is the achievement
of a roughly equal balance between male and female children,
which is preferred in all Tahitian families.
Children have no legal claims to rights in their adoptive parents'
Rights may be conferred upon them by testament, but
this is done only rarely. Many adopted children in Tahiti are
grandchildren, who can in any case expect eventually to gain the
rights by the operation of the French laws of succession. Other
adopted children gain rights to the land of their natural parents by
succession. In cases where the child is an orphan and the adopting
parents have natural children, it would be most unusual for the
child to be given any rights by testament on the part of his adop-
tive parents, since this would diminish the rights of the natural
children. In special circumstances, however, testaments are made,
as in the case of an elderly widower who had adequate land and no
natural children and who adopted a child from the kindred of his
deceased wife. This child did not expect to obtain sufficient land
from his natural parents by succession, so land was willed to him
by his adoptive father. Testaments are never drawn up, or even
agreed to, at the time that adoptions are initiated, since such an
action would imply too great a commitment in a relationship de-
pending on so many circumstances, such as the possibility of the
child not fulfilling his obligations in later life.
An adopted child may return to his natural parents of his own
free will, or be taken away by them if he is being cruelly mis-
treated, or even, in some cases, if he should just happen to prefer
living in his natural parents' household. When children leave in this
manner, it is maintained by some that the adoptive parents have a
claim upon the natural parents for monetary compensation for the
time and expense which they have invested in the child. However,
parents abandoned by adopted children are, in almost all cases,
too overcome by shame to press these claims in any public or legal
manner. In the words of one woman informant, it is the 'ultimate
shame' ua hope te ha 'ama to have an adopted child run away for
any reason, and it is perhaps for this reason that many adopted
children are treated with special indulgence.
Adoptions nearly always take place between close kin. The ter-
minological system for kin is of a simple generational type, so an
adopted child simply continues to apply the same kin terms within
his adoptive family. The terms of reference, papa fa 'amu and
mama fa 'amu, may be used when necessary to distinguish adoptive
from natural parents, even when the adoptive parents are the
child's natural grandparents.
It is not uncommon for Tahitian women to have one or two
children before beginning an approved and open conjugal relation-
ship. It is an ethically neutral act for a woman to give such chil-
dren to her parents or to other close relatives for adoption. The
woman is thus freed to continue ohipa taure 'are 'a 'the business of
adolescence' if she should wish to do so. There is no social stigma
attached to a person who is, in the eyes of French law, illegiti-
mate, or even to those persons whose genitors are unknown. Such
en/ants naturels are, however, discriminated against by the law in
matters of succession. Most conjugal unions are somewhat un-
stable during their early years, for at least the period until one or
two children are born. When such unions break up, the children
are commonly adopted or else cared for by kin until their parents
have established other unions.
Children may be adopted at any age up to the onset of adoles-
cence, when individuals are expected to be (and generally are)
highly independent of adult control. But adoptions either at birth
or during early infancy are strongly preferred, since they ensure
that the child will not 'think of his mother'. Feeding bottles and
canned condensed milk are commonly used in the rural areas by
mothers who prefer not to suckle their children. Arrangements to
udopt a child may be made before the child is born.
Tahitians have a variety of personal names conferred formally
and informally upon them, but the legally correct name is that
which is entered in the civil register of the Administration. Since
declarations of birth must be made within three days, and since it
is usually the socially-recognized genitor who makes this declara-
tion, children to be adopted are usually given a name chosen by
the natural parents. An individual or couple to whom the child is
promised in adoption may make the declaration of birth if they
wish to and also bestow names; but the matter is generally re-
garded as of little importance, since the child will in time acquire
various nicknames, a baptismal name, and a marriage name.
Naming only assumes importance if a child should be sickly and is
diagnosed by a local medium as suffering because of the anger of
the spirit of a dead person for having been 'incorrectly' named.
Tahitians recognize as a general truth that grandparents are ex-
ceedingly fond of their grandchildren, and they treat their grand-
children with a great deal more indulgence than that with which
they treated their own children. Explanations and rationalizations
are not offered for this state of affairs; it is simply a fact. Grand-
parents gain obvious pleasure from having grandchildren in their
households. It is held that they have the right to "demand" their
grandchildren in adoption, and that it is a filial obligation to agree
to such demands.
Maupiti is a small, infertile, and comparatively isolated island some
forty miles to the west of Raiatea. The 657 inhabitants are
grouped into 102 households which are, for the most part, strung
out along the eastern lagoon shore of the island. Since there is in-
sufficient fertile soil for the production of vanilla, the people have
not shared in the prosperity which this crop has brought to neigh-
boring islands in recent years. The per capita annual income from
cash crops is approximately sixty-three U. S. dollars, about half
that of the more prosperous com1nunities.
Adopted children are relatively numerous on Maupiti. Using as a
measure what may be termed a crude adoption rate (the number
of adopted children as a percentage of the total population) the
figure for Maupiti is 8 percent, by comparison with 3. 6 percent for
the prosperous vanilla-growing community of Murifenua. It may
be shown with some certainty that the higher crude adoption rate
of Maupiti is due to a complex of reasons stemming ultimately
from the island's ecology and geographic setting. This requires,
however, a consideration of data on the motives for adoptions and
the kinship relations between children and their adoptive parents.
Of the 102 households on Maupiti, 39 contain adopted chil-
There are 5 3 adopted children in the whole community,
and these are classified below according to two criteria:
The expressed motive for the adoption. Adoptions require the consent of
both the natural and adoptive parent(s), but, since adoptions are commonly
initiated by the adoptive parents, their reason for adopting is taken as the
primary one.
The kin relationship of the child to his adoptive parents.
I use the term adoption unit, abbreviated AU, to denote either a
man and his spouse, or any individual-widow, widower, divorced,
or never married. Adoption units are the adopters.
REASON A To bring into the household a child who will stand
in a grandchild relationship to the AU, and who will extend to
three generations the genealogical span of the household; 21 AU,
32 children, as follows:
ChiChi 17 AU, 28 children. Of these 17 AU, 13 have adopted
lite children of children who have left Maupiti to live elsewhere.
These 13 AU account for 21 of the 28 children. Four of the 17
AU have adopted children of children who live on Maupiti. These
4 AU account for 7 of the 28 children.
SibChiChi 3 AU, 3 children. These 3 AU adopted SibChiChi
Instead of ChiChi. There are different reasons in each case, as
AU I A widow with no natural children who had adopted a
SibChi and then a SibChiChi. The SibChi ran away to his natural
purents, leaving the widow with only a SibChiChi.
AU2 An elderly man and his wife with four grown sons who
huve never married because they have practically no land. The wife
udopted an illegitimate SibChiChi from Papeete.
AU3 A man and his wife who had adopted the man's sister's
illegitimate child. When this child reached adolescence she went to
Papeete and sent back an illegitimate child of her own.
ParSibChiChiChi 1 AU, 1 child. This AU is an unbalanced old
woman who has never married and who lives with her two unmar-
ried grown sons. Years ago she adopted a ParSibChiChi. The latter
has now grown up and, although she has never married, she has had
numerous children, of which one is the adopted child of the
REASON B Childlessness; 7 AU, 9 children, as follows:
Friend's Chi: 2 AU, 2 children.
SibChi: 3 AU, 5 children.
SibChiChi: 2 AU, 2 children.
REASON C To bring into the household a child not wanted by
his natural parent(s); 8 AU, 8 children, as follows:
ChiChi: 3 AU, 3 children.
SibChi: 4 AU, 4 children.
ParSibChiChi: 1 AU, 1 child.
REASON D Unclassified. (Reason for adoption not recorded);
3 AU, 4 children, as follows:
Friend's Chi: l AU, 2 children.
ChiChi: 1 AU, 1 child.
SibSpoParSibChiChi: 1 AU, 1 child.
The above information is presented in summary form in Table 7.
TABLE 7 Reasons for Adoption in Maupiti
Children Households
Adoption No. % No. %
A 32 60 21 54
B 9 17 7 18
c 8 15 8 21
D 4 8 3 7
TOTAL 53 100 39 100
The reasons for the comparatively large number of adoptions on
Maupiti may now be made clear. The people are comparatively
poor. They are conscious of their poverty, and they regard poverty
as a hardship. As a result, considerable numbers of young adults
leave the island to settle in other parts of the Society group, par-
ticularly Tahiti, where wage-labor employment opportunities are
available. In spite of a high rate of natural increase, the total popu-
lation of the island has remained stable since 1951.
As a result of
this emigration, the Maupiti population has proportionately fewer
persons in the "young adult" and "adult" categories than other,
more prosperous rural communities. Table 8 shows the relevant
data for the two communities of Maupiti and Murifenua.
Maupiti also has high dependency ratios for both children and
the aged.
Figures for the two communities are given in Table 9.
The demographic facts presented in Tables 8 and 9 have a
number of important consequences for various social characteris-
TABLE 8 Population of Mau pi ti and Murifenua
Categories Maupiti Murifenua
Child 490 418
(0-14 yrs.)
Young adult 154 254
( 15-24 yrs.)
Adult 290 291
( 25-59 yrs.)
Aged 66 37
(60-80 yrs.)
TOTAL 1000 1000
TAliLE 9 Dependency Ratios for Maupiti and Murifenua
( 'hildrcn
tics of the Maupiti community. Their effect on adoption is seen
most clearly, however, through a consideration of some of the de-
tails of household organization. In the rural areas, households are
the principal social units of production and consumption in the
subsistence economy (which embraces fish, vegetable foods, and
such store-bought goods as bread, sugar, and kerosene). For the
provision of these goods, there is a marked intrahousehold division
of labor. The only aspect of this division of labor that concerns us
here is the retirement of elderly men and women from productive
roles to enjoy the support of younger household members, nor-
mally their children. These economic considerations account for
the marked uniformity in the composition of household groups
and the changes which they undergo over time. Over 96 percent of
all rural households may be placed within the developmental
schema which is set out in Figure 1.
The logic of this developmental sequence is plain. The mainten-
ance of the domestic subsistence economy requires active and
able-bodied participants. As Tahitians grow older, they must begin
to make provision for their well-being in later years. Having raised,
fed, and cared for their children while they were young and help-
less, the parents can expect their children to repay the obligation
in kind as the children establish their own marital unions. This ob-
ligation falls equally upon all children, yet not all of them can re-
main within the household after they have established their own
marital unions. Some will undoubtedly leave the community al-
together, particularly those who belong to sibling groups which do
not have adequate land. Others settle within the community,
either in new households or in those to which their spouses al-
ready belong. Yet, at least one must remain within the natal
household, bringing in his or her spouse, so that the household
acquires the three-generational span which it will maintain until
the death of the members of the older generation.
Over 90 percent of rural Tahitians over sixty years of age are
established in households of three-generational span; that is, in
households of types B-G (see Fig. 1 ). The differences in the age
composition of the populations of Maupiti and Murifenua have
clear effects on the frequencies of the different types of house-
holds in the two communities. These frequencies are shown in
Table I 0. Mau pi ti has a considerably greater percentage of three-
B c
o-o 0-0
OorD=D H
0 mI or femel
I] clepen.d from houllhold
Figure 1. Household development cycle
TABLE 10 Types of Households in Maupiti and Murifenua
Type of Frequency(%)
Household Mau pi ti Murifenua
A 43 65
B 12 IO
c 14 6
13 IO
2 3
F 8 6
H 8 0
IOO 100
generation households than does Murifenua ( 49 percent as against
35 percent), but, most significantly, this difference is due very
largely to the higher percentage of type C households in Maupiti.
This would seem to indicate that in Maupiti, where the emigration
of young people is so common, the adoption of grandchildren is a
mechanism through which the economically necessary sequence of
household development is maintained. The people of Maupiti
adopt grandchildren because large numbers of their own children
have left (reason A).
It is somewhat more difficult to account for the specific fre-
quencies of adoption for reasons Band C. Childlessness (reason B)
is not uncommon in the rural areas; venereal diseases may have
something to do with a number of cases, but I do not have suffi-
cient information to elaborate upon this suggestion. The fre-
quency of the adoptions which are made to provide care for
children who are not wanted by their natural parent(s) (reason C)
also has complex determinants. Three of the eight children
adopted for this reason were taken because of maternal death; two
other children were taken because their mothers were simple-
minded. The remaining three, the children of unknown (though
reputedly Chinese) genitors, were sent back to the island by their
mothers in Tahiti. One may presume that in the urban environ-
ment of Papeete the socially recognized genitors of illegitimate
children are not under the same pressures to acknowledge and care
for their progeny as they are in the rural communities.
For Tahitians, children are valued resources, both for reasons of
sentiment and for economic reasons. It is a feature of the Tahitian
ideology of the parent-child relationship that a long-term balance
of care, help, services, and affection should be achieved between
parents and their children.
A number of the propositions which make up what I have here
called the "ideology of adoption" indicate that the adoptive
parent-child relationship is very similar to the natural parent-child
relationship. There are the same long-term obligations and rights,
many of the same sexual prohibitions, and the same expectations
about the quality and strength of the emotional content of the
relationship. In these circumstances, the adoption system in the
Society Islands has a number of clearly adaptive features:
I. Individuals who have no natural children may by adoption
validate their adult status and provide for their own later security.
2. Illegitimate children, orphans, and the children of the unbal-
anced and incapable are cared for.
3. The labor demands of household subsistence economies are
provided for in the face of disruptions such as emigration.
Since the ideology provides for a very flexible, adaptable means
of "social substitution," it is probable that the system will persist
with the inevitable further "modernization" of Tahitian society.
The field research on which this report is based was supported by a United
States Public Health Service Fellowship, MH-13,805; a Research Grant
from the U.S. National Institute of Mental Health; and the South Pacific
Research Programme at Auckland University.
2 Tahiti being the largest and most widely known of the Society Islands, I
use the term "Tahitian" to refer to "the indigenous Polynesians of the
Society Islands."
J Firth ( 1936: 588-596) provides what is still the best summary of these cus-
toms. His summary is largely based on Bishop Museum ethnographies,
which describe, for the most part, vaguely "traditional" practices on the
various islands.
4 This word is derived from fa (causative prefix)+ 'amu ('to eat')+ ra'a (sub-
stantive suffix). Davies {1851:61) gives only fa'a'amu. In modern Tahitian
the causative prefix has two acceptable forms,fa'a and fa.
5 There is in rural Tahiti a socially acceptable, extralegal form of enduring
conjugal union. By "unmarried persons" I mean those who are not living
in any form of socially-recognized conjugal union.
6 Unless, of course, they happen to be legally adopted. There were no cases
of legal adoption in the two communities which I studied.
7 A "child" is here defined as a person under sixteen years old.
8 Official census figures are 1951: 649; 1956: 665; 1962: 657.
9 The ratio between the number of children and the number of economi-
cally productive males is frequently used by demographers to express the
degree to which the economically productive section of a population sup-
ports those who are not productive. It is called the dependency ratio. In
the Society Islands, both children and the aged are dependent, and it is
thus necessary to calculate two dependency ratios:
DR for children= No. of persons 0-14 yrs./No. of males 15-59 yrs. X l 000/ l
DR for aged = No. of persons 60+ yrs./N o. of males 15-59 yrs. X 1000/ l
Robert l Levy
In some regions of the globe, however, adoption is practiced on a scale
wholly disproportionate to any rational ground therefor . ... Oceania as a
whole represents a main center for adoption earned to unusual lengths.
(Lowie 1930:460)
In an article on Tahitian adoption, Kay states:
Tahitians appear to have strong preferences for fostering close relatives, which
operate irrespective of the social structure of the community in question.
l lowever, within these overall preferences, there are finer distinctions which
c.lo appear to be affected by the local social structure. The most dramatic of
these preference differences is that of the preference in the relatively rural
communities, where agricultural land is of greater importance, to foster chil-
c.lren of a not too distant generation (1963b: 1038).
Having demonstrated economic factors in the tendency to
choose kin of a close generation for adoption in rural commu-
nities, Kay asks, "Why then does not the total number of foster
children drop off in the urban community," where the particular
l'conomic motivations he had suggested did not operate. In an
curlier treatment of the problem (l 963a: 165) he remarks, "Our
discussion seems to be asserting that urban Tahitians continue to
foster children in the absence of any good economic or socio-
logical reasons for doing so-apparently due to the 'blind force of
custom'." He rejects this residual explanation, and goes on to sug-
gest other rational economic motives for urban fosterage. (Thus,
parents or other older relatives of the parents of the children
assume their care, "this being exchanged for direct economic sup-
port," as "young adults in Papeete [the urban center] are less
likely to see children as a future economic asset ... than as a
present deterrent to the enjoyment of the pleasant distractions of
the town" [ l 963a: l 65f] .)
Since Firth's review of the early ethnographic literature, a lim-
ited variety of "good economic or sociological reasons" has been
suggested for Polynesian and Micronesian adoption. He wrote:
It is obvious that the custom is common in practically every Polynesian com-
munity and that the adoption most frequently practiced is that of kinsfolk.
But within this general pattern there are many variations. I think it could be
shown, if material were available, that such variations are not merely hap-
hazard but each fits into the institutional configuration of the particular
society ( 1936: 595).
Beaglehole epitomized the types of explanations which had
been given:
Adoption was ... a common practice in Polynesia, sometimes for the purpose
of securing an heir, sometimes to cement an alliance between two lineages,
sometimes to adjust matters of property-inheritance or to secure a share of a
limited food-supply; and at all times probably because children were always
welcome in a Polynesian household that knew not the terrors of disintegra-
tion of unemployment and starvation ( 1940: 54 ).
Similar factors have been stressed in later work. Danielsson, for
example, considering the highly frequent adoption in Raroia in the
Tuamotu Archipelago, stressed economic motives. The static econ-
omy "naturally makes it difficult for parents with many children
to provide for them, and the natural way out is to give them away
to couples with few children and plenty of land. On the other
hand, childless couples are pleased to adopt other people's chil-
dren .... Since ... the older generations are dependent on the
younger ... adoption is the Raroian's special form of life insur-
ance" (1955:218).
Lambert, in an article on adoption in the Gilbert Islands,
stressed kin bonds. "The majority of fosterage relationships grow
out of long-standing friendships between the natural parents and
the future foster-grandparents .... Fosterage is a means of trans-
forming the tenuous link that ordinarily connects distant kinsmen
into a far stronger and more significant one" (1964: 238).
Weckler writing about adoption on Mokil in the Eastern Caro-
line Islands suggests various motives. These include ideals of com-
pleted nuclear family membership, "without concern for and
sometimes contrary to the adopter's personal economic advan-
tage" (1953: 557); the formation of bonds with unrelated people
on or from other islands; the manipulation of land; and the hope
of economic gain for the child or for his patrilinial extended
family. There are instances in which the adopters are seeking
"social security" from the services of the child. Weckler suggests
that the stresses on economic uses of adoption in Mokil are recent,
and due to economic acculturation. "Ancient cultural ideals and
expectations still motivate people to participate in adoptions. But
the economic facts of life on Mokil today often counteract tradi-
tional values. There is more jockeying for advantage by partici-
pants, and more concern about the economic aspects of adoption,
than there used to be" (1953:576).
Adoption persists in some quite modernized Polynesian com-
munities long after political, social, or economic uses of the types
described for more traditional communities seem relevant, as the
study by Howard et al. (chapter 2) of a suburban Hawaiian
"homestead" community on Oahu so clearly indicates. For this
community, the more immediate personal motives for adoption
become more salient: the charitable taking of someone else's un-
wanted baby, and the pleasure of having a new baby around the
house as one's own children start to grow up. As Howard (chapter
13) puts it for Rotuma, where the traditional political and prop-
erty transfer functions of adoption have disappeared, "fosterage
... has become an exclusively domestic affair."
In this chapter I wish to suggest another aspect of fosterage,
which may be related to its persistence in a variety of Oceanic
social, economic, and political settings, and which gives it an im-
port beyond "a blind cultural force," or an "exclusively domestic
In his book on human adaptation, Rene Dubos summarizes recent
work on sensory deprivation:
Experiments on sensory deprivation indicate ... that the very structure of
the mind is determined by environmental stimuli. We know the mind only
through its expressions in an accustomed social environment, particularly one
which we consider to be normal. Like the body, the personality is being con-
stantly molded and held in shape by environmental forces. It is this fact
which gives literal truth to Emerson's remark that "men resemble their con-
temporaries even more than they do their progenitors" (1965: 26).
Jerome Bruner, in an article, "The Cognitive Consequences of
Early Sensory Deprivation" (in Solomon et al., 1961 :2040 wrote:
It would seem ... that not only are there critical problems of the develop-
ment of adequate models of the environment and adequate coping strategies,
but that there are also maintenance problems [my italics] of an order of
delicacy that were not even imagined before the pioneering experiments of
Hebb and his associates at McGill. ... What is this maintenance problem? I
would like to suggest that it perhaps relates to a kind of continuing feedback-
evaluation process by which organisms guide their correction strategies in per-
ceiving, cognizing, and manipulating their environment.
Renewed interest in the open, field-sensitive aspects of psycho-
logical systems reminds us that cultural forms affect individual
personality not only through processes of internalization during
the personality development of children, but also through their
continuing maintenance effects on older individuals. Although
most work on the importance of sensory complexity has dealt
with the importance of complexity in itself, Bruner's remarks sug-
gest that the particular structure of the complex environment
which is being constantly scanned has its own implications.
Thus, the codelike, formal regularities which Levi-Strauss finds
in various aspects of culture may represent patternings essential
for the maintenance of individuals' cognitive integrity.
Related is
the idea of specific aspects of culture as "models" that condense
useful information, which can be "extracted" by active participa-
tion of individuals with the models. Anderson and Moore referred
to the class of such models that were isolated as "play" or "art,"
activities "cut off, in some suitable sense, from the more serious
aspects of the society's activity" as "autotelic folk-models"
(l 960: 3). They stated: "We believe ... that every society will be
found to contain in its culture certain autotelic folk-models, with
the help of which it socializes the young, and through which the
adults understand and 'play at' the workings of the society itself'
( 1960:4 ). This idea has received empirical support in relation to
games in the work of J.M. Roberts and his associates.
It seems evident that certain specific cultural forms, which are
not isolated as are "play" or "art," persist partially because of
their function as important psychological statements. Such an
assertion involves all the problems entailed by "functional" propo-
sitions, particularly the dangers of circularity. Nevertheless, I will
consider the Tahitian variety of Oceanic adoption (particularly in
contrast to some aspects of American and Western European adop-
tion) as a cultural "message," which involves not only the partici-
pants in specific adoption transactions but the community as
At the time of my fieldwork in a rural Society Islands village,
which I will call "Piri," adoption was a frequent practice. In
November, 1962, Piri had 284 people living in fifty-four house-
holds. There were 42 children living in households other than
those of their biological parents. Of these children, 34 were iden-
tified by others in the household as fa'a'amu
'adopted' children;
3 were grandchildren with neither parent living in the household; 5
were borderline cases who were living in the household for indefi-
nite periods, but who were identified as a relative's child, rather
than as afa'a'amu child.
The oldest individual identified as a fa 'a 'amu child in a house-
hold was nineteen years old. There were, in all, 166 individuals
nineteen years old or younger living as children in households.
Thus, 25 percent of this age-group in Piri did not live in the house-
hold of a biological parent.
These adopted children lived in twenty-five households. One
household had 4, three had 3, five had 2, and the remaining six-
teen households each had 1 non-biologically related child. Of the
lwcnty-nine households which did not have adopted children,
some had had in the past, and many had given children for adop-
tion by other households. Thus the majority of households in Piri
had been involved in adoption in one way or another.
Children in Piri were transferred in two situations. First were
those cases in which the main pressure for transfer was the natural
parents', usually the mother's, wish to give the child away. Sec-
ond, were those cases in which the main pressure came from the
adopting parents' wish to take the child. Many cases involved some
combination of these two. Rarely, a mother who wished to give a
child away might have some temporary trouble placing him. This
might be resolved by "temporarily" leaving a baby with a relative
in another village and never taking the child back. A more fre-
quent asymmetry involved those transfers in which the child was
taken against the parents', again usually the mother's, will. This is
the situation in which, as Firth described it for Tikopia, "Some
member of the wider group of relatives bears off a relative from
the married pair and brings it up in his own household" (l 936:
The immediate motives for giving children away are unam-
biguous. Mothers wish to give away children if they are still in the
period of adolescent, experimental, shifting sexual relations; if they
do not want to settle down with the father of the child; if they are
in an early settled relationship but do not wish to become com-
mitted to it; if they have been in a more permanent relationship,
perhaps with a number of children, but at the time of the birth of
the new child have had an argument with the father, or have sepa-
rated from him.
The immediate motives for seeking an adopted child (which
should be distinguished from the accepting of a child that some-
one else wants to give) are not as clear, nor as clearly articulated
by people in the village. The most common explanations given are:
"They liked the parents of the child, and therefore wanted to have
one of the couple's children", ''They did not have any children of
their own"; "Their children were growing up, and they wanted a
new baby in the house." The latter explanations were supported
by the household census. In twenty-three of the twenty-five Piri
households with adopted children, the adopted children were
either the only children living in the household at the time (thir-
teen households) or the youngest child or children (nine house-
I never heard anyone state economic motives for an adoption
that he was directly involved in. There are reasons for not making
economic motives explicit. The ideology applying to adoption is
that it is an act of 'love and charity' arofa on the part of the taker.
The child, when he has grown, has an obligation to the adopter for
this selfless act, and community opinion will sanction this obliga-
tion. But, if the adoption is seen as a purely calculating act for the
new parents' profit, the child, supported by the community, has
the right to escape from this exploitation. This is a general feature
of the ideology of exchange in Piri. Goods and services are given
"selflessly." One returns them to avoid the shame of not recipro-
cating the "freely given" gift.
Statements about the economic motives of others-"People take
children because later the children will help them, and they can
share in the profits of the children's property" -were occasionally
offered by more marginal, cynical people, particularly by some
members of the newly established Mormon Church in Piri, which
had been attacking the custom.
It was frequently said, particularly by men, that it was not diffi-
cult or particularly troubling for the natural mother (in those cases
in which she had not initiated the adoption) to give up the child.
Men tended in general to play down all possible difficulties of the
female role, such as childbirth and menstruation. Some women
who had given up children expressed some cautious resentment
and recounted having made ineffectual attempts to hold on to the
child. Sometimes the demanding couple had had to repeat their
demand over a period of months; but, if they were adamant
enough, in time the mother had to yield. The more commonly
l' xpressed response was, "So and so asked me for the baby. I gave
it. l don't think about it any more." As Barnett writes for Palau,
"Parents are expected ... to take the changes of status [of their
children given for adoption] as a matter of course. This is the ideal
altitude and the one which is normally in evidence. There are,
however, suggestions that although people acquiesce to the de-
mands of custom, they are not always happy about it. These con-
trary sentiments are never overtly expressed except in individual
cases and in unguarded moments" ( 1960: 55).
An adoption in Piri was usually arranged for during the moth-
l'r's pregnancy. If the mother died in childbirth, or was ill, the
infant was transferred at birth. Otherwise, infants were given
either after they were five or six days old ("strong enough to leave
the mother") or, more commonly, after five or six months, when
they were weaned from the breast to a cup. Several women stated
that it was a good thing to give the infant away at an early age,
before one year, so that the mother would not become too at-
tached to it.
The majority of statements about the relative treatment of
adopted and biological children stressed equality of treatment.
"After all, they took the children because they wanted them and
because they love children." One villager said that in the rare cases
in which the adopted child was not a relative, he would more
likely be mistreated. Reports concerning mistreated adopted chil-
dren tended to support this.
Village observation showed no gross differences in parents'
treatment of adopted and biological children. There were, how-
ever, some subtle differences.
Biological children would sometimes confront the adopted child
with his special position. For example, a woman in her sixties talk-
ing about her childhood said: "We had an adopted child in our
house then. She was older than I. My father loved her very much.
When we would fight and she would say 'I'll beat you up,' I'd say,
Oh yes? Does this father belong to you?' ... It made me angry;
my father really loved her." She laughs in recounting this.
A frequently reported aspect of Polynesian and Micronesian
adoption is that it is not a complete transfer, The child has various
residual relationships to his biological family. ln the Tokelaus, "in
the simple village life, where the children roamed in and out of
every house, the separation from parents was not absolute. A child
always knew who his true parents were and understood his rela-
tionship to his foster parents" (Macgregor 1937:38). In Palau,
"adopted children always know who their real parents are, for
there is no effort made to conceal this fact from them. Indeed,
concealment would not be feasible because the real father retains
the privilege of using the patrimony of his children that comes
from their foster father" (Barnett 1960: 54 ). Firth sums up for
Polynesia: "Actual severance is usually fairly complete as far as
residence is concerned but a number of social contacts are usually
maintained with the true parents" (1936:595).
In Piri, there were various degrees of transfer in the adoption
relationship. In some cases, when the natural parents had casual
contacts with a child whom they had given for adoption they
acted particularly coolly towards him. Such a child was relatively
cut off from his natural parents, although in some cases, after the
death of the adoptive parents, the child would return to them. In
the majority of cases in which the natural parents lived in the same
village, or a nearby one, the child maintained contact with them.
In some cases the child moved back and forth occasionally be-
tween the two families. The changes were often due to the child's
decision to move, following some punishment or disturbance in
the household where he had been staying. The emphasis on the
child's decision to move is important. It would be considered
wrong for the natural parents to coax or influence the child to
return, and this would stir up considerable community disap-
proval. When a child is given for adoption, the natural parents
theoretically relinquish their rights of control or influence over the
child. If, however, the child is badly mistreated by the adoptive
parents, the natural parents may try to take him back.
The child living in his biological family could, like the adopted
child, leave his household for that of another relative if he felt mis-
treated, although this is only rarely done now. According to infor-
mants, changing of households is more likely to be done by an
adopted child. In most respects, however, in the ordinary course
of a child's life, the differences between the experience of children
in adoptive relationships and those in biological family relation-
ships seemed minor.
As revealed in life history interviews, there were no discernible
differences in the village status, in behavior, or in psychological
characteristics between adults who had been adopted as children
and those who had not. (Seven adults, adopted as children, were
among fifteen adults given detailed life history interviews.) This
does not mean that surveys on larger populations or the use of
finer psychological studies would not have shown differences. It is
l(Uite likely that they would have. If special effects of having been
adopted did exist, effects which would noticeably differentiate the
adopted person from others in the community, they did not show
up in the values, interpretations of the world, interpersonal tech-
niques, and defense structures elicited by psychodynamic inter-
viewing. In his article on adoption on Mokil, Weckler states, "The
adoption of children is a major means by which adults can alter
the membership of that basic social unit, the elementary family.
We may consequently expect that, where adoption is frequently
resorted to, it will have important effects on the structure of
society and on the personality dynamics of the people involved"
(195 3: 5 55). If Weckler means socially important specific effects
on the adopted individual, this may be questioned.
The question of the locus of the effect of the practice of adop-
tion comes up in relation to Firth's treatment of 'the adhering
child' tama fakapiki, in Tikopia:
This institution of the tama fakapiki has the effect of providing a child in a
house where otherwise there is none to help in the work, but the natives do
not always regard it primarily as a device for assisting barren couples or in-
creasing the household strength.This is shown by the fact that though a man's
eldest child is often taken by his younger brother who has as yet no offspring
of his own, in other cases the child is added to an already existing set. It has
no inferiority in the family to the real children .... [The practice is] a mark
of respect to it and its parents. [If a man is generous] his children are sought
by his kinfolk. At the back of this is the idea, quite clearly expressed in fre-
quent statements to me by natives, that it is bad for a child to adhere only to
its parents; it belongs to the larger group, the kano a paito, and must stand in
an equal relation to all therein.
. . . Two mechanisms are employed to part the child from its parents ....
The child may be severed from the household of its father and mother at an
early age and attached to another; failing this, or supplemented by this, its
interest is attracted by other members of its family circle who thus seduce its
budding affections. The parents, it may be noted, regard this with ap-
proval ...
. . . The basic social motive in this is to preserve as far as possible uni-
formity of conduct and attitudes within the larger social group and not allow
the bonds of the individual family to become so strong as to threaten the
wider harmony .... [Parents) approve ... that the child shall undergo a
social weaning [original italics] as well as a physiological one (1936:205f).
Firth found only 18 cases of 'adhering children' and some bor-
derline cases among 587 children and adolescents. If the impor-
tance of the practice for "not allowing the bonds of the individual
family to become so strong as to threaten the wider harmony" had
been confined to the families involved, or to the children directly
involved, it would have been quite limited.
These considerations suggest the possibility of a more general-
ized, and elusive, effect of Polynesian and Micronesian adoption.
In a tape-recorded interview, the village chief of Piri had been
talking about his adoptive mother, and her methods of discipline. I
asked him if he got angry with her when she punished him. (My
questions and comments are in parentheses):
I never got angry at her, even if she whipped me, because I was aware that she
had taken me. She took me. So I thought I was the one who was at fault, that
is why she hit me. I wasn't angry at her. (Explain that to me ... ) She took
me in my infancy ... [pauses]. (And, therefore ... ) Therefore, I loved her
although she may have been wrong in punishing me ... I didn't get upset. I
didn't get angry at her, because she took me from my infancy and kept me.
So, I thought, it wasn't right to be angry at her. ( ... Suppose she had been
your biological mother, would that have been different? Is one more grateful
to an adoptive mother than to a biological mother?) There are two different
situations. If you are not adopted, you are grateful to your biological mother,
because she gave birth to you. On the other hand, when you are taken in your
infancy by somebody, it isn't worthwhile to think anymore about your
mother. The woman who took you is just the same as your biological mother.
Your gratitude is great because you were an infant, and you were taken.
Another man in his mid-forties, who had been adopted as an
infant, expressed the same idea in talking about his adoptive
mother who was still living: "I'm not like some children, when
they are grown up they reject their parents ... because I came to
realize if it weren't for her, who would have taken me in my in-
The natural parents of both these men had moved to other
islands, and they had been brought up entirely by their adoptive
parents. Some of those who were adopted, but whose parents lived
in the same village, also remarked on the kindness of the foster
parent in taking them. People who had as children moved back
und forth between households often compared the relative good-
ness or badness of their different parents.
Responses such as "I am grateful to the persons who took me,"
or "what would have happened to me," or "I am free to move
between and compare two families" have in common the implica-
1 ion, stressed by Firth for Tikopia, that the adopted child is re-
lated to his adopted and biological families only contingently and
not categorically.
The chief of the village had noted that children who are not
given for adoption are, or should be, grateful to their mother be-
cause she had given birth to them. The covert formula may be
assumed to be, "she gave birth to them, and she kept them." It
would seem that if there are a number of adopted children in the
community, particularly if these seem to the village observer to
have been taken more or less randomly, then the social reality
should be created: children are kept by their parents, not because
of the natural, given order of things. but because the parents
happen to wish to, and are allowed to by others in the com-
munity. That is, all parent-children relationships will tend to be
seen as contingent.
Older children are sometimes told, particularly if there is a
visitor from another village, "We will give you to so-and-so." This
is a teasing threat, which would tend to reinforce other messages
about the possibility of separation.
It is useful to contrast some features of modern European and
American adoption with the Polynesian and Micronesian cases.
Distinctive features are associated with the adoption of non-
relatives, which, for the United States, represents about 50 percent
of adoptions. Adoption by "relatives" usually involves a step-
parent in remarriages or, presumably, close relatives where one or
both parents have died or become incapacitated (American Acad-
emy of Pediatrics 1959:2).
According to Kirk, "Considering the total stream of family be-
havior, the phenomenon of adoption is indeed numerically rather
insignificant. The best available evidence suggests that about 2.5
percent of the child population of the United States is adopted"
( 1964: l f). As the Fischers noted in an American village study,
"Adoption is a very rare phenomenon in North Village. It hap-
pened once during our year of residence, and no other cases were
recorded by us. Legal adoption, however, is such a long and diffi-
cult process that it is only undertaken by those who are truly
desperate to have children" (1963:889).
As Kirk has stated:
In modern industrial societies, adoption is practiced as an emergency meas
ure. Courts and agencies generally take the view that the original family
relationship should be preserved for the child whenever possible. The practice
is thus encircled by protective steps that signify the official reluctance to
tamper with bonds of consanguinity ( 1964: 1 ).
In a United Nations study on the adoption of children in vari-
ous Western nations is the statement: "Involving as it does a break-
ing of the ties which bind a child to his natural parents and substi-
tuting new ties with the adoptive parents, adoption has been
deeply influenced in law and in practice by the more or less sacred
character attributed to family bonds by society" (United Nations
Consonant with the "sacred character" of the family bonds is a
stress on secrecy of transfer:
Although all adoption workers are convinced of the necessity that the child
himself should know he is adopted, it is also his right, as well as that of his
natural and adoptive parents, that the adoption should not be advertised
through public proceedings, but heard in camera, and that the records should
be treated as strictly confidential. Such safeguards enable the child to be
more closely identified with the adoptive family (United Nations 1953:9).
The ideal of secrecy was limited by the requirement in every
country in the United Nations study that incest between the child
and a member of his biological family be avoided. Ideals of West-
ern adoption, then, as reflected in these reports, are that it should
be limited, relatively secret, and should involve a complete transfer
(limited by incest rules) from one family to another.
According to the World Health Organization meeting of experts
un the mental-health aspects of adoption, "the object in adoption
is to establish between the adopting parents and the child relation-
ships which coincide as nearly as possible with those between
parents and natural children" (World Health Organization 1953:
One may speculate that an object of Polynesian and Micro-
nesian adoption is to establish between parents and natural chil-
dren relationships which coincide as nearly as possible with those
between parents and adopted children.
An ideal-type contrast can be set up between Polynesian and
Micronesian adoption, and Western adoption in terms of form, of
the meaning or message contained in the form, and in terms of the
psychological implications that may be supposed to follow from
the impact of the message.
In form, Polynesian and Micronesian adoption is relatively fre-
quent,6 public, casual, and involves only partial transfer of the
adopted child to the new family. Western adoption is relatively in-
frequent, private, formal, and involves an almost complete trans-
fer, limited in principle only by incest considerations.
The message in the Polynesian and Micronesian adoption is that
relationships between all parents and children are fragile and con-
ditional. The form of Western adoption mutes from community
awareness the impact of the breaking of the biological relationship
and thus protects the essential Western orientation that relation-
ships between all parents and children are categorical.
Observations and life history interviews in the Society Islands
suggest the following psychological orientations as generalized
implications of parent-child contingent relationships, orientations
which would seem to contrast with those influenced by primary
categorical relationships. It is felt to be generally difficult and
unsafe to establish close, trusting, committed, emotionally in-
volved relationships with individuals-age peers, mates, children,
political leaders. These relationships should be kept shallow, unin-
volved, tentative. One should not care enough, or be engaged
enough, to be vulnerable if anything goes wrong. One should be
able to get out, to detach oneself. The nature of a relationship is
judged not by the abstract, naturally-given definition of the rela-
tionship, but practically and concretely by the actions of the
people involved, and often only by the latest actions (a considera-
tion which makes courtesy and good manners especially impor-
tant). In general one is ready to shift allegiances as interpersonal
and group situations shift. One searches for clues to the local
rules for acceptance, for the essential consensus, in the shifting
situation in which one finds oneself. Tentativeness extends to
goals in general, to one's sense of control over essential situations.
The thesis that Polynesian and Micronesian adoption is associ-
ated with, and implies, certain psychological orientations through
its pervasive message qualities does not state that it is a sufficient,
or even necessary, cause for these orientations. One may assume
that functionally important psychological patterns will be shaped
in a variety of ways towards roughly equivalent functional forms.
But the pervasiveness and durability of the practice suggest that it
has particular importance.
In Piri the message that relationships are contingent and inter-
changeable, and must not be taken too seriously, is repeated in
many ways. Children, from the age of four or five, are pushed by
parents into sibling and peer group relationships and out of close
parent-child dependency relationships; childrens' possessive attach
ments to objects are systematically shamed and interfered with;
adolescent romantic love relationships are vigorously attacked by
other members of the adolescent peer group; magic beliefs hold
that mourning and intense desire are dangerous in that they attract
dangerous spirits, which lose their power if one is casual enough;
articulated world view teaches that intense concern with an out-
come will produce its defeat; the lexicon has a paucity of terms
having to do with loneliness and feelings of deprivation involving
broken intense relationships, and such states are interpreted as ill-
ness, or supernatural uncanny states. (This is in contrast, for ex-
ample, with a developed, differentiated vocabulary for angry emo-
tions, which are interpreted naturalistically as forms of anger.)
Given the renunciation of the possibility or desirability of intense,
deeply involved relationships, there is in Piri a "reconstruction."
Mild, sentimental, verbal attachments are made. It does not seem
difficult for those people who express 'verbal gratitude' maururu
and 'compassion' arofa for their parents or others to leave them
and ignore them completely.
It may be asserted, without demonstrating it here, that the com-
plex of casual, tentative orientations is of great importance for
smooth interpersonal and social performance within the social and
cultural structure of a Tahitian village and that it is also of impor-
tance as an integrating element for culturally prevalent personality
structures. But observations in Piri indicate that the set of influ-
ences tending to produce casual orientations are in conflict with
tendencies toward intense relatedness. Mothers often wish to hold
on to their babies; adopted children sometimes report that they
get 'peculiar feelings' when they see their biological parents; ado-
lescents occasionally fall in love and start to form ideals; it some-
times takes a good deal of community pressure, including spirit
doctor treatment, to get people over mourning behavior. In con-
trast with the dominant reality statements of the cultural language
are the "natural" tendencies of complex derivation. It is the con-
llicts between these realms that in some sense makes the availa-
bility of techniques for reaffirming the cultural statements dynam-
ically important, and presumably of different quality from those
involved in the transmittal of non-conflicted patterns.
It is presumed here that the nature of the parent-child bond is
perhaps the most powerful statement to members of the group on
both the affective and cognitive levels as to the nature and possi-
bilities of person-other (including non-human object) bonds. It is
hard to imagine a way more potent than Polynesian and Micro-
nesian adoption of delivering the message that there is no relation-
ship which is not conditional.
The impact of this message is on
adults and older children.
I assume that without adoption, casual, tentative attitudes
would be much more difficult to maintain, and that if adoption
disappeared there would be a change in such attitudes and associ-
ated social forms, or else some other generating influence would
be markedly hypertrophied.
It has been suggested in this chapter that a central aspect of Poly-
nesian and Micronesian adoption is the expression and transmis-
sion of personal orientations which are related significantly to a
variety of cultural forms. An attempt is made to extend Firth's
thesis about the function of adoption in Tikopia from considera-
tions of the effects on the people immediately involved to con-
siderations of more generalized community effects. The kinds of
behavioral forms assumed to be associated with Polynesian and
Micronesian adoption are those suggested by fieldwork in the
Society Islands in French Polynesia. The generality of their associ-
ation with adoption throughout Polynesia and Micronesia is an
analytically derived hypothesis which remains to be demonstrated.
The fieldwork on which this chapter is based was a general study of
psychological behavior made during June and July 1961, and from July
1962 to June 1964, in two Society Islands communities. It was supported
by Research Grant M-5567-A from the National Institute of Mental
Health, United States Public Health Service, and by Research Grant
NSF-G23476 from the National Science Foundation. The study was made
in collaboration with a general study on social change in the Society
Islands under the direction of Douglas Oliver.
In this paper the adjective "Tahitian" refers to the Society Islands in
The author, a psychiatrist, prepared this chapter while he was Visiting
Associate Professor of Public Health, School of Public Health, University
of Hawaii. I wish to thank George Grace, Edwin Cook, and Vern Carroll
for helpful comments on an earlier draft.
2 Cf. also Colby 1966.
3 See, for example, Roberts and Sutton-Smith 1962; Roberts, Sutton-Smith,
and Ken don 1963; and Sutton-Smith, Roberts, and Kozelka 1963.
4 Fa'a'amu means 'to feed'. It may be noted, as an apparent exception to
the usual Polynesian metaphor of 'feeding' for adopting, that Davies' A
Tahitian and English Dictionary (of terms collected in the 1830' s) also lists
the now obsolete fa'atavai 'to adopt another's child', which is derived from
tavai 'to tend'. 'to look after', 'to look out for'.
5 Early ethnographic statements about Polynesian and Micronesian adoption
do not give exact, and often do not give even impressionistic, frequencies.
Adoption was said to be "common" in the Tokelaus (Macgregor 1937),
"exceedingly common" in Tonga (Gifford 1929), and "quite common" on
Truk (Gladwin and Sarason 1953).
Kay's data for his two Society Islands communities (1963a) indicate
that about 25 percent of the children in each one were fostered.
The lowest frequencies reported were for Futuna and Tikopia, both of
which had elaborated norms and rules for adoption. Burrows, in a four-
month stay in Futuna, wrote: "In spite of the various ways in which adop-
tion may take place, it is not very frequent. In a census of two villages
involving 4 7 4 individuals, only one was specifically labeled as an adopted
child" (1936: 60). Firth ( 1936) found eighteen cases of "adhering chil-
dren" and some borderline cases among 587 children and adolescents, in a
population of 1,281 people divided into 218 households.
Weckler ( 1953), on the basis of genealogies, concluded that nearly one-
third of all the children born on Mokil since 1775 had been adopted.
Danielsson (I 955) found that 40 percent of the children on Raroia
were adopted. Danielsson enumerates both the number of adopting house-
holds and the number of donating households. Of thirty-five households
and sub-households (relatively distinct nuclear units in a household), eight-
een had adopted children and twelve had donated children. Four house-
holds had both given and taken children, so that twenty-six of the thirty-
five households in Raroia had been directly involved in adoption.
6 Cf. footnote 5. The number of children involved and the number of house-
holds involved are both significant.
7 Anger and aggression in the Society Islands is discussed by Levy in Caudill
and Lin 1969.
8 There is a balancing aspect of the message which has not been considered
here. One can also take children. One aspect of the ethos of easy renun-
ciation is the belief in a balance of easy return of some other good to make
up the loss.
Paul Ottino
Rangiroa Atoll is located in the Western Tuamotus, about two
hundred miles from Tahiti. It is the largest and most populous
island of the archipelago; about seven hundred people live in two
villages, Tiputa and Avatoru. The data presented here were col-
lected at Tiputa, which, with its population of about four hun-
dred, is the more important village. However, it seems to me that
the conclusions are valid for the whole cultural and linguistic re-
gion which includes the neighboring atolls of Tikehau, Mataiva,
'Arutua, Kaukura, Apataki, and also the island of Makatea-a re-
gion sometimes referred to collectively as "Mihiroa."
Since ancient times adoption has been very common in the
Western Tuamotus. Together with marriage rules, postmarital resi-
dence rules, and other practices such as women exchange, it facili-
tated the allocation and redistribution of people among small
localized descent groups called 'iiti. These institutions made it pos-
sible to compensate rapidly for the unequal statistical distribution
of births by redistributing people among combinations of two,
three, or four traditionally related 'iiti, thus maintaining a total
number of individuals and a sex ratio more suitable to familial and
economic needs, including those of defense (see Ottino 1967). It is
suggested here that these past conditions continue to influence the
present practice of adoption and account for its very high rate of
It is necessary to define briefly some terms- 'iiti, 'opu, 'opu ho 'e,
'opil feti'i-which will frequently appear in this chapter. Formerly
the 'dti were 'lineages' of undifferentiated recruitment and exclu-
sive affiliation (see Ottino 1967), which were distributed all
around the periphery of the atolls. They lost this character of ex-
clusive affiliation when the population gathered in villages. Today
the inhabitants of the Tuamotus can affiliate themselves through
cognatic kin ties to a sometimes considerable number of 'iiti scat-
tered throughout the archipelago. But the 'iiti which really matter
for any individual are the 'iiti of the atoll in which he or she re-
sides and of which he or she can claim the status of 'inhabitant by
ancestry' feia tumu, since rights to land depend on affiliation with
these "indigenous" 'ati. At Tiputa only three ati are recognized:
Marere, Fariua, and Hoara. The 'iiti are divided into 'genealogical
branches', called 'opu, which are themselves subdivided into sub-
branches of two sorts, 'opil feti 'i and 'opi1 ho 'e.
Narrowly, 'opu ho 'e (literally, 'family', 'womb') applies only to
a set of siblings, suggesting terminologically that these constitute a
distinct sociological entity. The 'opu ho 'e disappears at the death
of the last surviving sibling. More broadly, an 'opi1 ho 'e is com-
posed of brothers and sisters and their off spring in the two follow-
ing generations. The 'opu feti'i is a more inclusive category which
is defined in relation to a grandparent (probably dead) and in-
cludes as many 'opu ho 'e as the number of distinct sibling groups
in the generation of this person's grandchildren. 'Opu ho 'e and
'opil feti'i are clearly cognatic categories. The social groups which
express these categories can, like individuals, be attached to several
'iiti at the same time; but in practice such groups are considered
attached to the 'iiti from which they get their lands.
Ordinarily, the term 'opil ho 'e applies only to natural siblings.
But in fact, adoption can create the same sort of relationship as
that between true siblings. Children raised together in the same
household feel very close to each other, as if they were true sib-
lings, with perhaps more affection and less antagonism. Preexisting
adoptive relationships within the family of orientation are very im-
portant in matters of adoption and adoptive kinship; they figure
prominently in subsequent adoptions involving family members.
The groups formed by parents, their siblings, and the descen-
dants of the latter are known as 'opu metua 'family of the par-
Usually, adoptive children are taken from or given to only a nat-
ural or adoptive father or mother, natural or adoptive siblings of a
parent, or the immediate descendants of the latter.
Table 11 is based on the results of an exhaustive census of
Tiputa Village taken at the end of 1964. Later on we will consider
some other cases of adoption which do not figure in this table
TABLE 11 Relationship of Adoptees to Adopters
Lineal grandchild
Adopted grandchild
Sibling's grandchild
Sibling's child
Sibling's adopted child
Children of adoptive parents
Deviant cases
Subtotal: 'very close kin' 'opii ho'e
'Descendants of parents' siblings' 'opii metua
Subtotal: 'close relatives' 'opii feti'i
Descendants of related ancestors other
than 'opu metua
Subtotal: all relatives
since they belong to an earlier period of time. The 1964 census
shows 164 cases of adoption, but as 37 village children who had
been adopted are counted twice, the actual number of children
transferred is 127. The kin relationships relevant to these 127
cases are summarized in the table, which indicates the frequency
of adoption for each type of relationship and also the kin cate-
gories within which adoption takes place.
The four "deviant cases" in Table 11 include three cases of the
adoption of a wife's child by her second husband, and the case of
a young brother adopted by his eldest adoptive sister. These rela-
tionships would be classified as 'opu ho 'e.
Table 11 shows that adoption occurs generally in the circle of
close kin. Of the 127 relationships noted, 99 (78 percent) are be-
tween 'opu ho 'e 'very close kin', and 107 (84 percent) are between
'opu feti'i 'close relatives'. By contrast, adoption of 'mere rela-
tives' feti'i is rare. It is very likely that many of the adoptions be-
tween 'mere relatives', and perhaps also some of the adoptions
between 'non-relatives', in fact involve children of close adoptive
relatives whose exact adoptive relationship has been forgotten.
Adoptive ties may easily be forgotten, consanguineal ties are never
forgotten. This then is one of the differences between the two
types of kinship. While natural relationships may be indefinitely
extended lineally, adoptive ones do not reach beyond the persons
related by adoption; they are operative only, as it were, to the first
degree. This explains the practice of maintaining this type of kin-
ship by adopting close kin (generally children) of one's adopter or
The adoption of children known as 'non-kin', those who have
no recognized genealogical relationship to the adopter, usually
involves an antecedent adoptive relationship. Of seven 'non-kin',
two were adopted on the grounds that they belonged to the same
'llti as the adopters. If the adoptees were effectively related geneal-
ogically to the adopters' 'ati, then their genealogies could be
traced; the fact that their exact genealogical relationship to other
members of the 'iiti could not be traced indicates that almost
without doubt these adoptees were the offspring of persons al-
ready adopted into the ati. The remaining case of the adoption of
non-kin concerns a young Chinese girl, who is her adoptive
mother's goddaughter. She is not exactly considered a 'stranger'
even though she is not genealogically 'kin'. According to Rangi-
roan ideology, she is "very much like close kin."
An important fact to point out is that among twenty-four
young couples who have adopted for the first time, nineteen are
childless. Sterile couples are extremely numerous; about one third
of all couples are childless. We are dealing here with a demographic
fact which we cannot disregard. Since all Rangiroans want chil-
dren, a childless couple has a tendency to ask to adopt a child
before his birth, even before a young bride becomes pregnant. Of
thirty-eight children adopted from deliberate and free choice,
eighteen were tamari'i tapa 'o 'asked for before their birth', and of
the remaining twenty, five were asked for on the very day of their
birth, as soon as the sex of the child was known. For the tamari'i
tapa 'o the sex of the child is not relevant; the adopter agrees to
take the child he has reserved whether it is a boy or a girl.
Among different forms of adoption, the adoption of the off-
spring of natural or adoptive children seems to be perceived as
somehow different from the adoption of children in the first de-
scending generation. In case of the mqther's death, the grand-
parents feel bound to take the children with them; it is an implied
obligation (of which we have eighteen examples in our census).
Another difference between ordinary adoption and the adoption
of grandchildren is that, while adoptive parents are always called
'father' metua tiine or 'mother' metua vahine, even if they are very
old and their adopted children very young, adoptive grandparents
remain 'grandparents' metua ru'au. In such a case, adoptive rela-
tionships do not prevail over consanguineal ones. The special status
of the adoption of grandchildren no doubt explains its persistence
in families of the Sanito religion, in spite of the determined oppo-
sition of this church to other forms of adoption.
The two predominant religions in Rangiroa are Roman Catholi-
cism and the Sanito religion (Reorganized Church of Jesus Christ
of the Latter Day Saints); both churches are opposed in principle
to adoption. However, the strongest opposition comes from the
Sanito church. Among church-going Catholics, adoption, like in-
terfaith marriage, is encouraged if gains for the church (such as the
conversion of non-Catholics to Catholicism) are expected, but it is
vigorously fought against if no such gains are foreseen. Examples of
a change in religious affiliation following adoption are numerous
enough to reinforce this suspicion. There is no case of adoption in
the family of the Catholic catechist or in that of his son.
The policy of the Sanito church in matters of adoption is even
more strict. A conversation with a Sanito catechist, He'enui, made
its main points clear. He'enui has ten children by his wife Mauari'i;
she is the adopted daughter of a woman from whom she got her
land and the house in which she lives with her husband. Although
his wife's adoption redounded to He'enui's advantage, he remains
strongly opposed to the institution of adoption, though he is
otherwise traditional in outlook. He'enui explained to me that he
almost reneged on his sectarian principles when one of his brothers
repeatedly asked for one of his children; but he decided finally
that his brother's marriage was not stable enough and that "it was
better not to give the child away." "Of course the brother was of-
fended," He'enui mentioned, "as anyone would expect." The de-
cision to refuse was well thought out and made with deep convic-
tions. The Sanito church, which preaches familial stability and har-
mony, formally forbids the adoption of other people's children
(even children of close relatives). According to the teachings of
this church, everyone must take care of his own children. Sanito
pastors fight vigorously against premarital and postmarital sexual
freedom. Without doubt, they want to discourage, in an indirect
way, accidental or experimental unions, which are so much easier
when there is no difficulty in placing the children of these unions.
However, Sanito pastors do not even attempt to discourage the
adoption of grandchildren by their grandparents. To attempt to do
so would run contrary to the belief that especially strong ties exist
between grandparents and grandchildren. He'enui thought that
this type of adoption was "somehow different" from other types
of adoption. He told me that most Sanito comply with the teach-
ings of their church, and adoption is confined mostly to grand-
children. This is true: of twenty-four adoptions by Sanito couples,
nineteen are adoptions of grandchildren. The special status of this
type of adoption in the culture accounts for its persistence.
A census made in 1964 of the seventy-four households of the vil-
lage of Tiputa showed that twenty-two had given children in adop-
tion, twenty-six had adopted children, and six had both adopted
children and given children in adoption. Thus fifty-four house-
holds, or 73 percent of the households in the village, had been
involved in at least one adoption. After eliminating from consider-
ation the households of bachelors and other individuals living
alone (such as widowed men and women who live alone in one of
the houses of their parents' compound), and also the households
of newly-married young couples, the total number of households
which could have adopted or given children in adoption, but did
not, is reduced from twenty to eight. Adoption is thus very com-
mon, Its occurrence is the norm; it is the absence of adoption in a
household which requires an explanation.
We shall begin our discussion of the circumstances of adoption by
examining a few concrete cases of adoption which will enable us
to sort out the different forms which it takes.
The following examples have been chosen because they illus-
trate nearly all of the forms of adoption on Rangiroa and thus
make apparent the ideology which commands decisions and
choices in this domain.
The starting point of the sequence of events which led to the
relationships illustrated in Figure 2 is Potini's adoption by Ro'o.
Potini is considered a 'son' of Ro'o (literally, a 'son related
through the 'iiti'), who belongs, through his mother, to the same
'iiti as Ro'o. Subsequent to his adoption of Potini, Ro'o adopted
Potini's eldest son. In this case it was not because the child was 'a
grandchild through the 'iiti' but because he was the son of Ro'o's
adopted son. Generally, people call the children they adopt from
their sibling's children or adoptive children mo 'otua fa 'a 'amu
'adoptive grandchildren', to distinguish them from 'lineal grand-
children' mo 'otua or mo 'otua fiinau. Fa'a'amu, meaning 'adopted',
is composed of fa'a causative prefix and amu 'to eat'.
We find the same sequence of successive adoptions-the adop-
tion of both a child and the child's firstborn-in the case of Maue,
Potini's daughter, who was adopted by Konea's husband and
raised by the couple. Maue's firstborn was adopted by Maue's
adoptive mother, who had 'marked the child before its birth'.
Eugonie V.B.
En gin
Tauiarri Konea Ro'o

I 11
I ..,.I 1
I ..,...,. ,
I Potini JJt _,. .,, -'
I \ ,,,,'- ,,,"'
I ,,, ,,,,\""' ,,, "
I ,,,,,""' \

1 I I l"-" //
// I \
/ I \
_______ /_ I \
I \
... ...
I --'M Bel
--- 'strangar'
I \
"" \
. Karo '...
Figure 2. A portion of the 'a ti Marere

- 'strangor' adopled
4-" _. into the "iti Marare
Maue's second child was adopted by Maue's 'natural father'
( metau fanau), Potini.
Potini's youngest daughter was adopted by his wife's sister.
From the latter's point of view, the adoptee was a sibling's child.
Potini's second son was given to a 'stranger', a woman who is said
to have "adopted outside of her 'opufeti'i ('kindred') out of spite"
after having asked several times without success for one of her sib-
lings' children. Potini's fourth son was adopted by the daughter of
the brother of Potini's adoptive mother. This adoption was made
on the basis of the prior adoptive ties and not because of the ac-
knowledged but distant genealogical relationship. This child was
taken back by Potini during sickness and died in the house of his
natural parents. This shows that the ties between a child who has
been adopted and his natural family are not irreparably severed.
The remaining adoptions noted in Figure 2 concern Ta'ati and
To'eto'e, and result from a dissolution of marriage through the
death of the spouse in the case of the former and by divorce in the
case of the latter. After his wife's death, four of Ta'ati's children
were adopted by his mother, Eugenie Van Bastolaer, who, a few
years earlier, had already adopted Engina, the eldest child (see p.
95). The sixth child, a girl, was adopted by her deceased mother's
father. To'eto'e's situation is similar: his two children were
adopted by his mother's brothers, Pari'i and Potini. This situation
is common: in many cases in which a marital union is dissolved,
the children of the union are quickly adopted by their lineal
'grandparents' or one of the siblings of the latter.
Besides illustrating various forms of adoption, these cases show
that adoptions are related to preexistent adoptive ties and are
made on this basis rather than on the basis of consanguineal ties.
This very important fact is even clearer in the case of Mama Teipo,
a middle-aged, childless woman who adopted nine children.
Mama Teipo had been adopted twice-which is exceptional-
first by Tearo (her father's sister) and Tearo's husband, and then,
after they both died, by Tane, a stranger. The various relationships
in which Mama Teipo participates are diagramed in Figure 3. The
numbers under the adopted children indicate the order of the
adoptions. At the time of the study, 1, 2, and 3 had established
their own families. The same was true of Mama Teipo's adoptive
siblings, a sister who lived in the village and a brother, Maxime,
who lived in the same compound as Mama Teipo, on the other side
of the yard.
The case of Kate, Tearo's natural daughter, is worth noting. In
spite of the fact that Kate was Mama Teipo's father's sister's
daughter, her adoption was decided with respect to the adoptive
tie (adoptive mother's daughter); the existing natural ties were
totally ignored. The adoption of Tereia offers a second example of
an adoption of one's adopter's natural child.
The adoption by Mama Teipo of children 5, 6, and 9 (her adop-
tive siblings' son, daughter, and granddaughter) was explained to
me as likely to "strengthen the [formerly existing] adoptive ties."
Considered in this light, some of the preceding examples illustrate
the same desire to strengthen and artificially expand adoptive ties
by adopting either a child of the adoptive child, a child of one's
adoptive siblings, or a natural child of the adoptive parents.
Let us consider more systematically the question of who adopts,
the conditions under which adoption occurs, and some of the con-
scious motivations which seem to control these decisions. It seems
reasonable to me to impute "motivation" only when the adoption
is freely decided upon without any exterior pressures. Taking care
of grandchildren or the children of siblings after the death of their
mother is culturally prescribed. The degree of freedom to refuse
such obligations is practically nil; thus, any study of "motivation"
in such cases would be meaningless.
In Rangiroa the initiative in adoption is taken by the prospec-
tive adoptive parents. It is rare to find natural parents who wish to
give away a child. I know of only two cases similar to those re-
ported in this volume by Levy (chapter 4) for the Society Islands,
where the practice occurs of "temporarily leaving a baby with a
relative in another village, and never taking it back."
Not all of the very young children who live in a household other
than that of their natural parents are necessarily adopted children.
Teura Pou, a widow who lives with her daughter, explained that
her grandson Teri'i was not adopted, that she 'only looks after
him' ti'ai noa. To my question about why she did not adopt him,
she answered in surprise, "But I am a widow." This answer ex-
presses two separate ideas: on the one hand, a child must be
adopted by an established family; on the other hand, the family
which adopts must be able to provide for the adopted child's
needs. Many people noted that the adoptive parents were well off.
This is an ideal requirement, but I know of two old people who
are needy and who nevertheless adopted four children, even
though they are barely capable of providing for their own needs.
In two other cases of this sort, I suspect that the purpose of the
adoption was to get control of the adopted child's land. In one of
these cases the adoptive family lives on their adopted child's land.
The quest10n of the status of a child by a former marriage of
one of the spouses is complex. Adoption of such a child depends
on his age, and, if he is the wife's child, on her attitude. It is a
well-known fact that in matters of divorce and remarriage mothers
are very possessive regarding their own children and do not always
allow their new spouse to take care of them (cf. Bohannan
1965: 122, citing Bernard 195 6). This is especially so when the
man is not a native of the atoll on which he lives. In this latter case
the care of the child is left exclusively in the hands of the mother
and her siblings. Thus, the extent of the role of the woman's
second husband e.ssentially depends on her attitude and on circum-
stances of residence. But women are always ready to adopt their
husband's children and to treat them as their own children.
Formerly, children were adopted by an individual and not by a
couple; more exactly they were adopted by a resident member of
the lineage who was part of the core assuring the perpetuation of
the 'iiti. Nowadays children are adopted by a couple. However, the
situation is obscured by two related factors: the adopted child's
'iiti or 'opil affiliation, and hopes of succession.
The discussion of the cases diagramed in Figure 2 has brought
to light the notion of 'those related through the 'iiti' (feti'i 'iitiJ.
Since husband and wife most often do not belong to the same 'iiti,
and, owing to incest rules, never belong to the same 'opu of the
same 'iiti, adoptions can never be made of children equally related
to both adoptive parents. Adoptions must be made either on the
husband's side or on the wife's side; if possible they are made
alternatively on each side 'to achieve equitable distribution'. The
preferred form of adoption, i.e., 'on alternative sides', is possible
only when both spouses are feia tumu 'indigenous to' the place in
which they reside. When one of the spouses is a ratere 'outsider',
most of a couple's adoptions will occur in the family of the spouse
who is native to the locality of the couple's residence. Thus it fre-
quently happens that the Tuamotuan who finds no child to adopt
in the circle of his close kin prefers to adopt from his spouse's
side, adopting one of his spouse's siblings' children, rather than
adopt one of his own relatives from a distant place. To adopt such
a relative would cause difficulties at the time of a property trans-
fer, because from the point of view of the adopter's family group,
the adoptee would be a complete stranger. "The faufa'a ('proper-
ties', 'riches', especially land) must not leave the 'iiti. "
The decision to adopt is made by a couple, but a request for
adoption is normally presented by only one of the spouses, who
acts as a spokesman for the couple. There are, however, some
exceptions to this principle. Eritaia Tepava recalled that Mama
Teipo (his wife's sister by adoption) together with her husband
had come, dressed in their best clothes, to ask for one of his
daughters (number 6 in Fig. 3).
As a general rule, no difference was traditionally made between
one's children, regardless of whether they were natural, adoptive,
or stepchildren. It also appears that, in Rangiroa, disposition of
property had no influence on the affective ties between adopted
children and their adoptive parents. The child was adopted by a
couple, although he might inherit from only one of his adoptive
Among voluntary adoptions-which we contrast with quasi-
obligatory, culturally prescribed adoptions-some are really desired,
while others result from fortuitous circumstances. Adoption gener-
ally being an assymetrical process in which initiative belongs to the
adopters, one might easily fail to appreciate the motivations of the
natural parents. In all studied cases, with perhaps two or three ex-
ceptions, the supposed desire by young people to give their chil-
dren away was simply assumed on the basis of the conviction that
young people were "naturally inclined to devote themselves to
amusement at this particular time of their life." This might some-
times be true.
X (a woman): "We have a little girl who was born in 1957. She is the
daughter of Y who left the island to follow his wife to 'Arutua (an atoll in the
Tuamotus). At that time Y and his wife were our neighbors. I asked for the
baby after she was born. She was two months old and the father immediately
consented. They were young and wanted to be carefree. They have not in-
quired about her since that time. When they left Rangiroa for 'Arutua, their
daughter was eight months old, and they came to see her before they left."
It seems curious, at least at first glance, that people rarely give
the reasons for which they adopted a child; they find such a ques-
tion naive. If one insists upon a response to a question concerning
motivation, informants will answer that they did not have any
children and wanted some. If one recalls that nineteen couples out
of twenty-four who adopted for the first time were childless, one
must admit that this explanation is not entirely irrelevant.
It is worth noting that in no case do young adoptive parents
mention the help that they can later expect from their adopted
However, allusions to the assistance anticipated from
adopters may be made by grandparents who, alone again after
their children have established families, want to adopt one or more
of their grandchildren. Eugenie Van Bastolaer (see Fig. 2) stated:
"I adopted Engina (Ta'ati's eldest daughter) because I wanted a
grandchild to help me. I took her when she was nine months old."
Such a clear statement on the part of adopters is very rare, but it
echoes some vague and malevolent oft-repeated insinuations con-
cerning "those people who adopt children in order to be waited
upon." At least once, this formulation was made in a more
pointed fashion. It came from Y, an angry Don Quixote, who was
on the worst possible terms with his relatives:
"No, I do not want to give my children away. My sister-in-law broke the leg
of my son who was given to her in adoption. The child complained for many
years later [I doubt very much that this incident is true] . For this reason I
am not willing to give away my children in adoption. All my wife's brothers
and sisters ask me to give them children-one of her brothers, who lives on a
distant island, has already asked me seven times-but I don't want to. I do not
want my children to be 'servants' away from Rangiroa. My wife is willing but,
as for me, I am not willing."
Adoption may sometimes result from fortuitous events-the
opportunity makes the adopter.
Kaua Tevaria: "My eldest daughter is not my 'kin'. I took her when she was
one year old. She had already been adopted by a woman in Pape'ete. This
woman traveled a lot and twice left the child with us here at Tiputa. My wife
asked to keep the little girl, and her adoptive mother consented. Our other
adopted child is my brother's son. My brother and his wife lived in my house
for a year, and when my brother left I asked him to leave me his child. The
baby was then seven months old."
In the preceding case the adopted girl was transferred from the
first adopter to a second adopter without the natural parents being
consulted. Generally, however, the successive adopters are close
To'eto'e: (see Fig. 2) "I had three children by my first wife, and I kept them
after we separated. The eldest is still with me. The second one was asked for
by Pari'i, one of my mother's brothers, and I agreed right away. The third one
is Karo. Maue asked for her. When Maue and her husband went to settle in
their new house, Potini (Maue's father) wanted to keep Karo."
Maue's version: "It was at the beginning. I had come back to my father's
house with my husband, and I wanted to take Karo. We all lived together in
my father's house while waiting for the new house to be finished. My mother
took care of Karo a lot. When the house was ready, my father told me, 'you
leave Karo here'."
Another case is that of Tengaki Papata: "My eldest daughter was
adopted. She was a year old when Tera (MoMoBro) asked for her.
Later he gave her to Fanau ... Fanau is my mother's husband who
lives in Pape'ete."
Then there is the case of Eritaia Tepava's father, who lived in a
neighboring village. When Eritaia Tepava came back from Papeete
with his wife, who had delivered her baby there, his father refused
to give them back their daughter, Paulina, whom they had left in
his care.
The successive transfers of adopted children, in the cases pre-
sented, were made for understandable reasons. However, it
happens that some people adopt children not for themselves, but
for their close relatives. This is particularly odd since these agents
inform neither the child's natural parents nor those to whom they
propose to give the child of their intentions.
Culturally prescribed adoptions occur among very close lineal
relatives, or siblings, for compelling reasons such as divorce, sick-
ness, and death. The age of the children is not relevant in such
cases. Any young member of the 'opu ho 'e who is looking for a
roof will be fostered by someone or other.
When the parents have sicknesses which are believed to be incur-
able, the children are adopted, since, as Maratino Tehei expressed
it, "the parents are forced to give away the children." Adoptions
following separation or divorce are also common. In the case of
the death of the mother of small children, the orphans are immedi-
ately taken by the parents of the father or of the mother. It is
customary for grandparents to take their grandchildren 'coming
back from the funeral ceremonies'.
Grandparents are inclined to think that young people do not
know how to take care of their children. This might explain the
number of adoptions of firstborn children by grandparents. This
is especially true in the case of grandchildren born to unmarried
"It is my daughter who gave me the baby. We don't know who the father is
but he is not a 'Caucasian' popa'a, he's a true 'Polynesian' ma'ohi."
Mahuru Tetoka: "My son-in-law and my daughter wanted to give us their
child because he was sick often. The mother and the father were very young
and did not know how to take proper care of him."
I cite these two cases together because they reveal the fact that
grandparents usually maintain that the fa 'a 'amu have been given to
them. The parents on the other hand usually maintain that their
children were taken from them by the grandparents. The latter is
probably closer to the truth. Grandparental solicitude toward
grandchildren knows no bounds: the grandparents insist upon
having a sick grandchild at their house.
In 1965 this sort of concern was instrumental in ending a very
serious quarrel between a man and his son-in-law. The latter left
his father-in-law's house after a serious argument and went with
his wife to live at his brother's house. Then he moved to a piece of
land which belonged to his mother-in-law, and he built a tempo-
rary house of coconut fronds. After a few months, one of the
couple's children became very sick and the grandfather went to
look for his grandchild, reestablishing the harmony of his family
by so doing.
A last example of the concern of persons of the grandparental
generation toward young children is presented:
In 1965 a fifty-year-old woman of high social position went to fetch a grand-
niece who had not been treated well by her natural mother (the wife of a
nephew). The little girl, who was nine or ten, had been made responsible for
all of the housework. The child's mother did not protest the older woman's
action. It was said that she "kept quiet because she was ashamed."
Although the social pressures on the natural parents to accede
to an adoption may be considerable, adoptions are in fact some-
times refused, as the following case indicates:
Eliane Petis: "Three years ago I wanted to adopt the son of my brother. He
did not want to give the child to me because he had only two children. Six
years ago I asked my sister for her child, she said that it was difficult; she
wanted her son to go to school at Pape'ete, so she refused. Generally when
one asks, people agree, but they don't have to agree. "
In the end, however, Eliane Petis adopted the son of a sister and
the daughter of a brother, in addition to a little girl adopted from
a 'stranger'. She suggested to me that between very close relatives
in the 'opu ho 'e, especially between brothers and sisters, it is
easier to ref use, because "between cousins it is more of an of-
Potini (see Fig. 2), who had already given away his first three
children, explained why he refused to give away his fourth, while
underlining the dangerous consequences of such a refusal: "I did
not give the fourth away; I did not want to. There were no chil-
dren with us. My wife's sister had asked for a girl but it was a boy.
Of ten people who have been refused a child are angry. " Potini
gave away his fifth child, a little boy whom, as we have seen, he
took back when the boy was sick, and who died in Potini's house.
Finally he gave the last of his children to his wife's sister, the
woman who had asked for his fourth-born. Maue, Potini's daugh-
ter, was no luckier with her natural children. She had to part with
all three of her children.
Her husband: "The eldest was adopted by my wife's adoptive mother
(Konea) who asked for him after his birth and took him when he was three
months old. The second was adopted by my father-in-law, Potini; she is being
raised with his adopted son. She was asked for after her birth and taken when
she was five months old. My wife did not want to give the youngest away
since we have no child with us, but I wanted to. One must give when kinsmen
ask. "
This statement agrees with that of another informant who sug-
gested to me that it was difficult to refuse distant relatives. It
would seem that adoption beyond a certain degree of kinship be-
comes a form of alliance, that is, it becomes a matter of politics or
prestige rather than a matter of familial relations. Repeated re-
fusals invariably bring about the breaking off of a relationship, as
the following case will show. A woman asked her sister for one of
the sister's sons from among the sister's seven or eight children.
Her sister absolutely refused to give her a child to adopt. So the
woman got angry and adopted outside the family. Some infor-
mants commented that the woman had asked for a child time and
time again before she got angry and decided to adopt outside the
family "out of spite."
Adoption establishes relationships between at least three persons:
the adopter, at least one of the natural parents of the adopted
child, and the adoptee. We can study these relationships from the
point of view of any of three dyads: the adopter and the adoptee;
the adoptee and the natural parents; the adopter and the natural
parents of the adoptee. Generally, the emphasis in the literature is
put upon the relation between the adopting couple and the
adoptee, whereas the other relations are more or less ignored.
To omit the relation between adoptee and natural parents is
ethnocentric. In Western culture the ties between the adopted
child and his natural parents would be completely ignored if the
fear of incest did not make it necessary to remember them. It is
completely different in Polynesia. Adoption is not secret in char-
acter and adopted children are well acquainted with both their
adoptive parents and their natural parents, and have ties with both
sets of parents. In Tiputa a great number of adopted children live
in the same compound as do their natural parents and siblings,
only a few yards away from their natural parents' houses.
The relationship between adoptive and natural parents is, in the
same way, often disregarded in the literature. This creates a serious
lacuna since, in matters of child transfer, the decision belongs to
these persons, who think, decide, and act, probably with the pos-
sible consequences of their actions in mind. This is especially so in
matters of inheritance. Adoptions by siblings of their respective
nephews and nieces result in a 'strengthening of the 'opil' by dis-
tributing the children among the siblings' families and preventing
an inequality of patrimony division in the second or third descend-
i11g generation. These processes are considered more fully in my
dissertation on the contemporary social anthropology of Rangiroa
(Ottino 1969); it is sufficient to stress here the formal solidarity
which often exists between the adoptive and natural parents of a
child. The institution closest to Polynesian adoption in respect to
establishing an enduring relationship between individuals of the
same generation is no doubt the Catholic institution of god parent-
hood, especially the compadrazgo as it is found in Latin America
(see, for example, Deshon 1963).
To these three fundamental relationships we might add a
fourth, the relationship between the adopted and natural children
of the same person. When we consider the relations between adop-
tive siblings, the crucial point is the possibility of union between
otherwise unrelated natural and adopted children of the same
parent. Such relationships are created each time that spouses, by
definition 'strangers' to each other, both adopt a child from their
respective 'opil ho 'e. Three informants from Tiputa claimed that
adopted children who are not feti 'i 'related' "could marry, because
they did not have the same toto 'blood'." One of them thought
that the union offa'a'iimu was "not bad" becausefa'a'amu 'are not
united by true blood' (aite 'ino, 'e 'ere te toto mau), meaning by
this that they could not be considered 'opu ho 'e. Although it is
true that incest and the prohibited degrees of kinship are some-
times defined in terms of 'blood', the theory advanced by these
informants would be considered false by other members of the
society. It is an entirely idiosyncratic view which was advanced by
three fervent Sanito and represents one aspect of the conflict of
contemporary ideologies. Obviously religious dogma has won out
here over native ideology. The indigenous theory which contra-
dicts these assertions is perfectly clear. It was elucidated to me for
the first time in the course of a discussion on household exogamy:
'[only] one eating place, [only] one plate, [only] one household'
(ho 'e vahi tamara 'a, ho 'e mere ti ma 'a, ho 'e utuafare). The same is
said of children (excepting adolescents) who have been raised to-
gether from an early age; although not related, they are considered
feti'i 'relatives' and cannot later marry. One notes that the validity
of this theory is even more general, applying, for example, to chil-
dren other than those who have been adopted. It applies to chil-
dren who are the products of other unions of their parents, even
when these children have not been adopted by the parent's current
spouse. On Rangiroa, any union of such children raised in the
same household would be regarded as incestuous.
The general Rangiroan theory of close relationship applies only
to children who have been raised together; these children generally
belong to the same generation. In such cases, the ties are as strong
as those between true siblings. I know of no case of a property
dispute between such persons. Conflicts about inheritance, when
they occur, are between cousins or between adoptive children who
have not been raised together.
It should be noted that adoption occurs at two different points
in one's life: the first point is when one is young; the second point
is later on when one's own children have left one to marry and
start their own families. Generally, the children adopted at the
first point are considered tamari'i 'children of the first descending
generation'. The children adopted at the second point are mo 'otua
'grandchildren' (or 'children of the second descending genera-
tion'). If the adopted children of these two periods are not related
by blood, ties between the two groups are practically nonexistent.
Conflicts of interest can easily break out, and union between them
is possible since there can be no incest between such persons. The
kind of kinship or relationship which comes from the fact of being
raised together cannot be created through the relationship that
two parties maintain with a third person, even a mutual relation-
ship with an adoptive parent.
The case of Potini, who took back his seriously ill son, (pp. 96 and
I 04) illustrates the principle that adoption in no way breaks the ties
between the 'child given elsewhere' and his natural family. The re-
lationship can only be severed if the geographical distance is such
that it precludes any contact. This happens when adoptive and
natural parents reside in different archipelagos. But even here the
break is not necessarily final. After a certain number of years, the
i.:hild, now an adolescent or adult, may come back. He is always
i.:crtain that he will be welcomed by his parents and by those of his
parents' siblings who live on the same island as his parents.
It frequently happens that, for reasons concerned with school-
ing, the adopted child comes back for a while-maybe for some
years-to live with his natural parents. This sort of thing has
tended to decrease on Tiputa following the establishment of an
elementary school program there, but this school does attract to
the village the children of the neighboring village of Avatoru, and
even of other nearby atolls. This of course brings back, for a while,
some of the Tiputa children who have been adopted out. In such
cases the natural parents act as parental representatives with regard
to their own children.
Eliane Petis: "My adopted son went to town to go to school. He lives with his
natural mother and comes here during the vacations. It is I whom he considers
his true mother. He calls his natural mother Mama Karo, but he calls me
Mamii. I am his true mother."
Adoption creates affective ties which are lifelong. An old couple
may make a long trip in order to visit the aged adoptive mother of
one of them. However, it happens often enough that an adopted
child chooses to 'return' to his natural parents when he becomes
an adult, especially when the adoptive parent dies, or when the
adoptive relationship is otherwise dissolved. Inspection of the life
histories of twenty persons about fifty years old, all of whom had
been adopted when young, shows that half stayed where they had
been adopted, while half went back to the village where they had
been born. Four or five elderly persons, originally from other is-
lands, are considered today, at least by the younger generation, to
be "from" Rangiroa. They have in fact become so: after their
adoptive parents' death they stayed on the latter's land. Other per-
sons who are natives of the village but who have been raised by
their adoptive parents on other atolls of the Tuamotus, or on
Makatea or Tahiti, have later chosen to 'come back'. For example,
Matahi, who had been adopted by his mother's mother after his
mother's death and raised on Raroia Atoll, came on a schooner
one day, identified himself, and said he "was going to live with his
father," who had since remarried and lived at Tiputa.
A woman who had been adopted by her maternal grandfather
also came back to Jive 'at her father's side'. Sometimes, the re-
turn is caused by an argument with the adoptive parents. M was
born in Tahiti and was adopted when a year old by a 'brother' of
his mother, following the separation of his parents. He and his
adoptive father never got along, and after a violent argument, he
"went back home, to his natural father," who was born also in
Tahiti but whose parents came from Tiputa.
In addition to these 'returns' to the fenua 'land of origin', (not
necessarily the land where one was born), 'returns' occur within
the village, from one household to another.
These 'returns', already frequent, will no doubt increase in the
future, since adopted children are not legally recognized as such
by French law and have no legal right to inherit from their adop-
tive parents unless-and this is sometimes done-their adoptive par-
ents bequeath a part of their property to them.
In the absence of
a will, the fortune of such children depends on the good will of
the other heirs. French Polynesians who live on atolls practically
never legally register an adoption. I know of no case of legal regis-
tration on Rangiroa. Some friends from Tiputa confessed that the
thought of doing so never occurred to them. Moreover, according
to custom, the rights of use that the adopted possesses are only
given for his lifetime; they cannot be transmitted to the adoptee's
direct descendants unless the latter are in turn adopted. The result
of this is that the natural parents feel increasingly responsible for
the future of their children given in adoption. If these adopted
children were turned out or anticipated being turned out of their
adoptive home, they would tend to come back to their natural
parents for shelter and subsistence.
"My son came back to live with me. He was given in adoption by his mother
after we separated, but it was I who recognized him, not his adoptive father .
. . . As for the baby, the daughter my present wife gave me, my mother's
brother asked me for her so I gave her to him, but he did not recognize her,
so legally she is my daughter."
Stressing of the Western legal conception of legitimacy as
opposed to traditional norms of the adoptive relationship is char-
acteristic of the present changing situation. Concerning the "regis-
tration" alluded to in the above case, it was formerly not un-
common for adoptive parents, at the birth of the child they had
spoken for, to register him under their own name. I know of three
such cases.
Changes in adoption customs are swift. A case which caused a
scandal is indicative of the weakening of ties between the adopter
and the adoptee (in this case the adoptee was a 'stranger' to the
adopter). H, a young boy of sixteen, had a serious accident on a
motorcycle which did not belong to him. B, the adoptive father,
asked A, the natural father, to pay for the repair of the broken
motorcycle. This claim shocked people; they thought that B's rela-
tionship with H was wholly exploitative. Finally the natural father
and the adoptive father each agreed to pay half of the expenses. A
was off ended. He reckoned that if he and B were 'relatives' 'opii
feti'i this would never have happened. Curiously, B reasoned that,
since he and A were 'opu feti'i, he was entitled to ask for 'help'
tauturu from his relative, A.
In the past, adopted children had the same status as natural chil-
dren, and most of them inherited from their adoptive parents. A
survey of a representative sample of inheritors shows that of nine-
teen middle-aged adopted persons, eighteen had inherited from
their adoptive parents. This has changed recently and is more and
more rare nowadays because there is always 'trouble' with the
natural heirs. One example is the case of a man who first adopted
his brother's son, and then, much later, adopted a grandnephew
who was a 'relative' of the same 'iiti. The grandnephew planted
coconut trees on the old man's land and took care of him until his
death. The old man was grateful and left the second adopted child
a piece of land located in the southern part of the atoll. After the
old man's death, the first adopted son, who was then in the Cen-
tral Tuamotus, hastened back to Rangiroa and joined two of his
brothers who lived on the island in forbidding the youngest to
walk on that piece of land. To avoid friction, the youngest adop-
ted child of the old man gave up his rights, of which he could
easily have forced recognition.
The oldest adopted cnild explained his position to me by point-
ing out that he acted as a 'close relative' of the old man against a
'distant relative' feti'i atea of the same person, and thus was not
obliged to take into account the adoptive relationship which re-
lates each of them to their adoptive father. He added that the
younger man was 'well married' and preoccupied with the care of
the land of his parents-in-law [an observation which was correct].
Another informant said jokingly that this sort of difficulty be-
tween fa'a'iimu who are raised apart is frequent, "because they do
not belong to the same nest of eggs."
I have already mentioned the singular practice of registering the
adopted child under the name of his adoptive parent rather than
under the name of his natural parent. This is particularly easy in
the cases of tiimari'i tiipa 'o 'an adopted child who has been spoken
for before his birth'. This practice, entirely illegal from the point
of view of French law, is fully consistent with the indigenous
ideology and was certainly much more satisfactory than the pres-
ent practice of providing for adopted children by bequeathing
them land. Indeed, the old system made it possible to fully incor-
porate the adoptee into the group of consanguines. In matters of
inheritance, this protected him from litigation initiated by those
of his adoptive brothers and sisters who were natural children of
his adopter. At the same time he had no advantage over the latter
since the fraction of rights which he inherited from his adoptive
parent was of the same magnitude as the fractions which the other
siblings inherited.
Properties thus remained familial properties,
the management of which remained within limits neatly defined
by the culture and under the control of all of the siblings. By con-
trast, the adopted beneficiary of a formal legal bequest inherits
complete individual rights subject to the control of no one. This is
quite contrary to Tuamotuan norms.
At present the property inherited from natural parents remains
familial property, in which only the natural children have use
rights, unless-and this is very rare-these siblings opt for legal divi-
sion of the property. On the other hand, the property of an
adopted beneficiary of a bequest has been removed from the
joint-ownership of the sibling group; such property is individually
owned and can be freely disposed of. In other words, the bequest
of a fraction of inheritance to a fa 'a 'amu is equivalent to a defini-
tive partial division, to the ultimate advantage of the adopted
This fraction will, in all probability, be later passed on to
Conflicts of interest in land are rare while the beneficiaries are
still alive. Such conflicts tend to break out in the ensuing gener-
ation. Since the adoptive legatee's rights are more extensive, they
are that much more irritating to his adoptive siblings, especially in
the case where the latter are only distantly related genealogically.
Friction is maximum when there is no known genealogical rela-
tionship between them, i.e., when they are 'strangers'. However,
when the fa 'a 'amu is a close relative, he does not behave differ-
ently from the way he would behave if he had simply inherited a
use right. In this case the restraints in exercising his rights that he
imposes freely on himself can be explained by his status of 'close
relative'. It is rare that 'close relatives' who have been adopted,
especially members of the 'opu ho 'e of the adopter, are benefi-
ciaries of separate bequests, for the simple reason that rights given
to these persons are usually not contested. In contrast, bequests
are usually made to adopted children who are 'distant relatives' or
'strangers', to protect them from future claims.
In adoption, the principle that "riches must not leave the 'iiti"
accounts for differential treatment of children in wills according
to which of the families of their adoptive parent they belong. A
man of the 'iiti Marere who left part of his property to one
adopted daughter because she belonged to the same 'iiti left noth-
ing to his other adopted daughter, taken from his wife's side,
"because she belonged to the 'iiti Fariua." In former times, what-
ever their origins, the adopted were wholly assimilated into the 'iiti
where they resided. The first stage in this assimilation was their
adoption; a frequent second stage was their marriage to a member
of the 'iiti. Another way for an adoptee to assimilate himself to
the group into which he had been adopted was to adopt members
of the 'ati or 'opil of his adopter.
Modern changes in the practice of adoption have modified the
ideas which are related to it. More and more people hesitate to
leave their property to fa 'a 'amu who are 'strangers', unless the
latter are settled in the village and give other evidence of their will
to be assimilated, such as marrying and adopting within the 'iiti or
'opu. Any bequests to 'strangers' must be strongly, and usually
emotionally, motivated. One might recall here the case of the
woman who adopted "out of spite" and made her adoptive son
her heir (pp. 96 and 105).
It is common to hear people say that adoption 'strengthens'
(fa 'apuai) kinship. What is even more interesting is the remark
made by two informants that adoption serves to 'maintain'
(ha'amau) kinship. Unhappily, neither informant was able to spec-
ify exactly what was meant by these terms. Some persons felt that
adoption had something to do with 'the absent', that is, those
members of descent groups who are native to the atoll but who,
for various reasons, reside elsewhere.
For example, until very recently the Pou branch of the 'iiti
Marere was in eclipse on Rangiroa because its descendants, scat-
tered throughout the Central Tuamotus, had been away from the
atoll too long (for nearly three generations). Owing to the adop-
tion of Pou children back into the established branches of the 'iiti,
the Pou are gradually recovering their status in the Marere 'iiti.
Several cases already presented have thrown into relief examples
of adoptions in a sequence such that one adoption accounts for
the one which follows it. Since adopters and adoptees live together
for at least one period of their lives, and since the latter sometimes
inherit houses from the former, adoptive filiation, like ordinary
filiation, accounts for certain facts of residence. Especially signifi-
cant in this regard is continuity in occupancy. The successive occu-
pants of houses or residential units are as often related to each
other by adoption as by natural filiation.
Adoptive filiation, just as much as cognatic filiation, insures the
perpetuation of localized descent groups. The opu ho 'e (or opu
feti 'i) is a widely scattered extended family whose sociological
reality (as opposed to conceptual reality) is at best tenuous, ex-
cept in the case where a core of its members resides in the place
with which the family is traditionally associated.
Only a domicile (this word is here given its full legal meaning
and contrasts with residence) allows the group to retain its iden-
tity as such. Thus one understands the necessity of maintaining a
permanent representation, a core of resident 'natives' (feia tumu)
who occupy the 'familial houses' (fare tupuna) and take care of the
property (especially land), which, if the groups were to disappear,
would immediately be scattered among other similar groups,
claimed through cognatic ties which cut across all the groups. To
give adoption the same weight as natural filiation is a means of
maintaining a sufficient number of close relatives in the place of
domicile. According to the data collected on Rangiroa, permanent
residents of Rangiroa are the ones who retain the initiative in most
adoptions. Very often, when a native of the atoll leaves, he feels
the need to leave one of his children behind-"to console my
feti'i" as one young woman said. (Note the use of the general
word feti'i 'relatives', even though the child was left with the
l l 3
woman's father and mother.) Similarly, those who stay on the
atoll feel a need to adopt a young child of the feti'i "who is leav-
The Rangiroans quite obviously feel that the perpetuation of
the 'opu is contingent upon the maintenance of this residential
core. A friend, whose brothers were off amusing themselves in
Papeete, said that it was a good thing that he was in a position to
adopt his brothers' children, "or else what would become of the
'opu?" Clearly there was no danger that the 'opu, considered as a
group of related people, would die out: the fact that the brothers
had children precluded this. "What would become of the 'opu"
clearly implies that for the 'opu to "exist" some of its members
must live 'at home' as a residential family. This is shown by the
case of a group of siblings which includes their adoptive brother
Tiho Harrys. Of nine siblings (including Tiho Harrys) only Ruben,
Punua, and Tiho Harrys reside in the village. The others are in
Tahiti: Maro, because he was ad0pted there as a child and stayed
there; two of the siblings for health reasons; and three out of per-
sonal choice. The three resident brothers have no children, and
were it not for adoption, the local representation of their group
would risk extinction. Adoption saves the situation: of the fifteen
children born to the group of siblings, five have been adopted by
Ruben, Punua, and Tiho Harrys. These numbers do not express
fully the sociological realities of the situation, since Maro Tepehu,
who had been adopted and was permanently resident in Tahiti,
and his children are considered 'lost' and have no more ties with
Tiputa. The proportion of adoptions, once Maro's children are ex
cluded, is five out of eleven. Considering the four siblings who had
children, seven of their eleven children were 'left.behind'.
Ruben said, concerning Ioane, his eldest brother's son, who al
though living with him is the fa 'a 'amu of Tiho Harrys, "whether
he lives here (at Ruben's) or there (at Tiho Harrys') it's all the
same." In the same way, the young woman in tears about to go on
board the schooner had left a child "to console her feti 'i. " These
cases lead one to wonder about another problem: in what sense
does adoption really relate to individuals, rather than to larger
Perenate Fuller (in charge of the interests of her brothers and sisters on Ran
giroa): "At that time I had only one daughter, Vatina, my eldest; so my
eldest sister asked my brother Raihea, who already had many children, 'to
send me a boy'. This is how I took Teofiro who was then seven months old.
Four years later Raihea had his eighth child on Takapoto [an atoll of the
Eastern Tuamotus]. My mother, who lives in Pape'ete, made the trip to
Raihea's island and, without consulting me, asked him to give me another
boy, the last born. Raihea agreed, and my mother brought back Ernest, who
was then five months old."
The case of Punua Tepehu reported by a third person: "Among the Tepehu,
some of the siblings have many of each other's children, while others like
Ruben and Punua have none. Ruben had many other fa'a'amu children, but
Punua never ventured to ask. So the brothers and sisters talked among
themselves, and one day one of them who lived in Tahiti came to Tiputa with
his first son, who was then nine months old. He went directly to Punua and
asked him to take the child, since Punua had no son."
These cases, which could be multiplied by citing similar events
that occur in the history of every descent group on Rangiroa
which has rights in land on the atoll, suggest that relating adoption
to individuals is perhaps to take the effect-or more exactly, the
consequence-for the cause. It may be an illusion. Adoption quite
possibly should be related to the group of siblings; this group is
the key to Polynesian familial organization.
I do not think that adoption can be studied only from the
synchronic point of view. One must take account of situations in
the past; these very often explain situations in the present.
Owing to the existence of an underlying ideology, which ac-
counts for most of the decisions taken concerning adoption, the
presentation of a statistical table in which the observed cases are
logically ordered and classified on the basis of the nearness of the
consanguineal kinship relationship would not have, in itself, very
great significance. In the very numerous cases in which adoption is
decided on the basis of a previous adoption, such a presentation
would be irrelevant, since consanguineal kin ties are never relevant
to the actors in such cases. Moreover, when consanguineal kin ties
happen to coincide with adoptive ties, only the latter are deter-
minative. Preexisting adoptive ties alone explain why a person
adopted a particular child. These observations parallel those of
Goodenough, who when troubled by a similar problem with re-
spect to residence, stated:
It is clear, then, that more than census data may be needed for even the sem-
blance of pattern in residence to emerge. It is also clear that after a pattern
does emerge, interpretation of particular residences with respect to that pat-
tern requires additional sociological and cultural information (italics added,
Inspection of the cases presented by Goodenough shows that he
obtained this "additional information" by taking into account the
historical dimension. This is the only way to follow the events
and, to take Goodenough's example, to understand the trans-
formations of a unitary residence rule which can be described as
either patrilocal, avunculocal, or neolocal. In the same way, adop-
tion must be studied as a sequential process. Its pattern cannot be
grasped by focusing on the events of only a brief moment.
In the low islands of the Tuamotus, residence factors are at least
as important as kinship and operate simultaneously (cf. Titiev
1943). Decisions concerning adoption are made with reference to
an unconscious model, the notion of the 'opu ho 'e. These deci-
sions result in the gathering into one household of 'siblings' who
are not siblings by birth but who, nevertheless, are very closely
related to each other by the fact that they have been raised to-
gether. The composition of this family of orientation is often
highly relevant to the adoptions which will occur at a later time.
The study of residence histories and of 'familial' or 'ancestral
houses' (fare tupuna) is essential to an understanding of localized
descent groups. Since any one person is, independent of his birth-
place and residence, supposed to be 'a native' (feia tumu) of some
one place, these groups are composed of only a fraction of their
potential members, the ones who live in the place of which they
are native. Adoptive filiation, like natural filiation, appears to be
one way to maintain at a satisfactory level the effectiveness of this
resident nucleus, and to thus insure its perpetuation. In the village,
such groups correspond to groups of households, sometimes scat-
tered but more often close to each other or even enclosed in the
same 'liua fare 'compound' around the ancestral house. Finney has
termed these groups "household clusters" (1965: 303). These
groups are of great importance in that they exercise control over
the properties, and especially the land, of the entire 'opu ho 'e. The
adoption of the children of scattered members is evidently the
best way to preserve the reality of this somewhat fictive notion of
'opu ho 'e, which appears to be a kind of extended family that is
widely scattered geographically and that retains its identity only
by its connection with a place and through the nucleus of its mem-
bers who, living in that place, attest to its continuity. Adoption
counterbalances the disadvantages of dispersion (necessary for
other reasons) and indirectly maintains the potential rights of the
absent through their children who have been adopted. The fact of
adopting a child "because one likes to have a child at one's side"
has no bearing on the structural requirements of the social system,
the mechanics of which become more apparent as soon as one ob-
serves what actually happens over a period longer than the lifetime
of a single individual.
The research upon which this paper is based was conducted under the aus-
pices of the Office de la Recherche Scientifique et Technique Outre-Mer in
Papeete from the end of 1962 to the end of 1965. A total of eighteen
months, distributed over this three-year period, was spent on Rangiroa. I
am indebted to Vern and Raymonde Carroll who translated this contribu-
tion from the original French. Their searching comments on earlier drafts
have led to extensive revision.
2 Finney (1965: 291) suggests that taure'are'a are considered an asset.
3 For more than a century, the French legal code has nominally applied to
all of French Polynesia. In fact, however, the impact of this legal code was
scarcely felt in the Tuamotus until the past several decades. To this day
one finds there a very limited understanding of its niceties. The frequency
of inter-is!and movement is explained by the economic opportunities
offered on the high islands in contrast to the lack of such opportunities on
the atolls.
4 These fractions were use rights rather than ownership rights. Ownership
was vested in the sibling group as a whole, which in respect of its rights in
the land of a parent might be considered a "corporation." The concepts
suggested by Goodenough of "full ownership" as opposed to "divided
ownership" are also applicable here. He defines these terms as follows:
Two basic forms of ownership must be distinguished. One of these will be called full
ownership, be the owner a person or a corporation. It confers on an owner what will
be called a full title. The other divides a full title assymetrically between two parties,
either or both of whom may be individuals or corporations. This will be called di-
vided ownership. It confers on the two owners two distinct titles respectively, each
characterized by different rights and duties. One will be called a provisional title and
the other a residual title (1951:33-34).
5 Cf. Leach, especially the following passage:
Property ... is owned only in two ways. Rights over property are either vested in
perpetual corporations or in individuals. Individual property is a private matter, the
owner acts as he chooses. Corporate property is more complicated, for the individuals
who control such property may either be acting in a representative capacity-in
which case the actors' behaviour is almost wholly predetermined by the rules of the
corporation-or in a private capacity. In the latter case the rules merely specify the
limits of what is permissible t 1962: 131).
Vern Carroll
Nearly every adult Nukuoro has had first-hand experience with an
adoptive relationship. A majority of the individuals living on the
island in 1965 had been adopted at least once in their lifetime and
had thus experienced having adoptive parents.
Most children have
had the experience of growing up with adoptive siblings.
The ex-
perience of adoptive parenthood is also shared widely: almost all
of the married adults on Nukuoro have either given or taken a
child in adoption, or helped to raise a spouse's adopted child.
There were only two married adults-representing just 2 percent of
the resident population-who had po experience at all with adop-
tive parenthood.
Adoption on Nukuoro bears some relationship to barrenness: all
childless married couples who had been married for five or more
years (in 1965) had adopted at least one child. There were noun-
married childless women over twenty-nine who had not adopted a
child, and all but three mature single men had also adopted chil-
A post-World War II increase in the fertility rate has been
accompanied by a decrease in the rate of adoption,
but infertility
does not begin to explain the frequency of adoption, or the details
of adoption procedures, or the nature of adoptive relationships.
The aim of this chapter is to present the relevant ethnographic
data concerning adoption on Nukuoro and to explain these data
by presenting certain propositions about Nukuoro kinship that
may be of some relevance to the other societies considered in this
volume. We shall begin by indicating what the Nukuoro mean by
An 'adopted child' in Nukuoro is a tama ssala 'a child who has
been sought'. This term is used especially for children who have
been adopted from rather distant relatives, or to replace an heir
who has died, or to provide an heir for a barren couple. For ex-
ample, a grandchild who has been adopted to 'replace' a son or
daughter would clearly be considered a tama ssala. It is this sort of
adoption which is made most explicitly and which is accompanied
by the greatest degree of formality. A child adopted in this fashion
is least likely to be relinquished by his adoptive parents. He is also
least likely to return to his natal home of his own volition. A child
becomes 'sought' only after he has gone to live with his adoptive
parents; in the contexts in which it is used, the term explains why
the child is not living with his natural parents. Thus, a child who
has been asked for in adoption but not given up by his natural
parents cannot be a tama ssala. Also, since the term connotes the
completion of an adoption transaction, any termination of the
adoption (through the child's returning to his natal home, for ex-
ample) establishes retrospectively the fact that the child was not
tama ssala at all.
An adopter is sometimes said to 'keep' (in the sense of 'retain'
or 'restrain') daohi a child. The term is especially appropriate
when the adopter is a parent or other close ascendant of one of
the child's parents. Less commonly it is used in connection with
an adoption by a sibling of the child's parent, especially an older
sibling. It is interesting that the Nukuoro use a term for 'adoption'
which is used in other contexts in the sense of 'keeping somebody
(or something) from doing what he (or it) has a natural inclination
to do'. The only way this usage can be explained is by supposing
that the term alludes implicitly to the fact that children "natur-
ally" gravitate toward their natural parents (who "naturally" tend
to assume responsibility for them) but that close kin of the par-
ents have certain rights in any child born to the latter, and that an
adoption of the sort under discussion is considered as an exercise
of these rights. This supposition is supported by the many cases in
which a child's grandparents take an active part in arranging for his
adoption. Such an adoption is viable only for as long as the needs
which motivated it exist. For example, an aged grandparent living
alone may 'retain' a grandchild as household help and then give
the child back to his natural parents when there is no longer a use
for him, for example, when the grandparent moves into a house-
hold where he is provided with daily care.
Less formal still is the process of 'making a child one's own' hai
maana. This is not necessarily more than a mere frame of mind, of
which all parties save the "adopter" may be ignorant. When a child
is 'made one's own' it does not necessarily mean that the child is
taken from his natural parents, or that the relationship supplants
in any way the relationship between a child and his natural par-
ents; rather there appears to be an implication in this usage that
the adoptive relationship is one established in addition to the rela-
tionship between the child and his natural parents. An adopter in
this sense may be prepared to confer upon an adoptee all of the
rights of a natural child without requiring that the adoptee fulfill
the obligations of a natural child. This sort of adoption allows for
the possibility of a child being adopted by several persons at the
same time, and this indeed occurs when the various adopters are
not aware of each other's declaration to adopt. This is the sort of
adoption which leads to the most confusion, since none of the
relevant parties are clear as to who has done what or who intends
to do what. The confusion becomes apparent at certain times
(such as in drafting a will) when the principals must make their
relationships explicit. The Nukuoro do not, however, appear to
mind the ambiguity of adoptive relationships except on occasions
when inexplicitness cannot possibly be maintained.
It will be noted that the three sorts of adoption which the
Nukuoro discriminate linguistically have been arranged in descend-
ing order of explicitness. It is fair to say that the more explicit the
adoption, the greater the likelihood of it proceeding to the point
where the adoptive parents supplant the natural parents for most
practical purposes; the more vaguely stated the adoption, the more
likely that it will come to naught. What is surprising from the
European point of view is the relatively small number of adoptions
which are ever made fully explicit. While the majority of the
people on the island appear to have been 'adopted', there are rela-
tively few adoptions which have not been reversed before the
death of the principals. This suggests that the reasons for our mak-
ing adoption ties explicit-to insure the welfare of the child-are
not matters of great concern to the Nukuoro.
In what follows, whenever the English term "adoption" is used,
I am referring indifferently to any of the phenomena described
above. Although it might make the account less confusing to dis-
cuss each type of adoption in turn, it is not possible to do so be-
cause the Nukuoro do not see adoption as falling into distinct
types to which different norms attach.
It is convenient to begin an ethnographic account of Nukuoro
adoption practices by considering the contexts in which adoption
requests are made.
A child is normally adopted, in the first instance, by a single indi-
vidual who negotiates the adoption with a close relative.
It is
only a 'relative' who feels that he or she has a right to ask for a
child, and it is only a 'relative' to whom a child will be given.
There is no cultural preference for male or female adoptees.
Adopters who are serious in their intentions are expected to
provide some clothing, soap, and other small presents to the child
at its birth. A failure to manifest concern at this time is construed
by the natural parents as lack of sincerity. Even more important is
the provision, by potential adopters, of the child's mother with
gifts of food during pregnancy and lactation.
The adopter may not take the child before he is weaned. In the
few cases where grandparents have insisted upon taking a grand-
child in infancy, the infant is reported to have died soon there-
after. Weaning occurs on the child's first birthday and is often
accomplished by sending a child to live with relatives.
A relative is never asked to conceive a child for the express pur-
pose of making a child available to a potential adopter, and only
rarely is a relative asked to promise a child before it has been con-
ceived. It is common, however, to ask for a child as soon as preg-
nancy is evident. Adoption requests made after the birth of the
child are more likely to be resisted. This appears to result from the
extreme reluctance of parents to part with their children once
they have become attached to them.
It is difficult to refuse a relative's request for a child-especially
if the request is made before the child's birth-since to do so is to
deny the existence of a bond of kinship with the person who is
making the request. Relatives should be willing to share all of their
resources, including children, especially when one of the relatives
is in need. It is especially difficult to deny a request from a mar-
ried relative who has no children or who lacks children of the
appropriate age or sex to provide essential household services.
Even where need is not entirely clear, proper kinship spirit is dem-
onstrated by acceding to an adoption request. The closest ana-
logue in our society to a Nukuoro adoption request would be an
invitation to a dinner party: to refuse for no reason is to risk rup-
turing whatever relationship has been established. Most Nukuoro
are unwilling to take any risks of this kind; instead they resort to
various stratagems to avoid giving children to relatives with whom
they do not feel close, without actually refusing requests in so
many words. The only recognized "excuse" for not giving up a
child in adoption is that the child's parents have already agreed to
a prior request. Parents who flatly refuse to allow any of their chil-
dren to be adopted cut themselves off from all of their kin and are
completely dependent on their own resources in securing a liveli-
hood and satisfying their various other needs. Only one of the
households on Nukuoro in 1965 had so isolated itself.
An adoption request from a parent is least likely to be evaded;
siblings and close cousins are also hard to deny.
Parents or
siblings would be outraged if ever they were refused. On the other
hand, one's closest relatives are those who are least likely to termi-
nate a kin relationship merely because an adoption request has
been evaded. Thus, failure to satisfy a distant relative involves
more risk to the continuance of the relationship.
It must be stressed that "closeness" of relationship, from the
Nukuoro point of view, is a subjective state which is not wholly
predictable from genealogy. In many important ways, a full sibling
from whom one is estranged is less of a 'relative' than is a distant
cousin who is also a good friend. The immediate content of a kin
relationship is as important as its genealogical form; so, for ex-
ample, it is much more difficult to refuse an adoption request
from a very distant relative with whom one has been on good
terms than to refuse a request from a close relative with whom one
has had a falling out.
Adoption requests from 'relatives' with whom one feels little
kinship are usually managed by assenting to the request and then
evading compliance with one's promise. In theory, the first relative
to ask for a child should have it. To discriminate between relatives
on grounds of "closeness" is to violate the principle that relatives
should be treated equally. A distant relative whose adoption re-
quest is denied in favor of a closer relative would be incensed to
learn that a closer relative is 'favored', except of course when real
or classificatory 'parents' have asked for a child. 'Parents' must be
obeyed; their requests take priority over those of siblings and
Neither a father nor a mother can give away one of their chil-
dren in adoption without the consent of the other; both have
equal and inalienable rights to the child (except that an unmarried
mother may deprive her child's genitor of claiming these rights by
refusing to admit that he is the actual genitor). It is generally rec-
ognized, however, that an adoption request is very hard to deny,
and one would hesitate to embarrass one's spouse by refusing an
adoption request originating with a spouse's relative. It is not un-
common, however, to claim that one's spouse will not agree to an
adoption, as a means of fending off an adoption request from a
relative to whom one does not wish to give one's child. A relative
will never approach the spouse to verify a report of this nature:
most Nukuoro are 'afraid' to deal directly with non-relatives. An
adoption request can often be dodged merely by suggesting to
one's relative that he take his request directly to one's spouse.
The difficulty most Nukuoro have in dealing directly with non-
relatives makes them aware that it is impossible to get an honest
expression of sentiment from someone to whom one is not re-
lated. Thus a prospective adopter will usually suggest that his kins-
man ascertain the wishes of the latter's spouse concerning the
prospective adoption, "lest the spouse be embarrassed and hide his
true feelings." This makes it all the easier for the pivotal relative to
provide a report which expresses his own views on the matter. It
may be arranged between kin for the child of one to go and live
with the other, with the tacit consent of the child's other natural
parent, but that parent can always claim that the child was never
really 'adopted' since permission was never formally granted.
The Nukuoro are well aware of the possibilities for misunder-
standings among the parties to an adoption, and I shall later dis-
cuss their motivation for often allowing adoption arrangements to
remain so ambiguous. It is clear to everyone that the only secure
adoption is one in which the assent of both of the natural parents
has been obtained. Occasionally such an adoption is arranged, but
the natural parents do not jointly assent to the adoption, rather
each of them makes a separate pact with the adopter, adducing
different reasons for agreeing to the adoption. For example, if A
and B are the natural parents of a child who has been asked for in
adoption by C, then A may tell C that he is giving him the child
"because we are relatives," while B may say something like, "I am
not giving you my child because you are my husband's relative,
but rather because you and I are [distant] relatives."
The spouse of the adopter is also relevant to any adoption pro-
ceedings. A child must have a mother to feed him and care for
him, thus a male adopter must first obtain his wife's consent to
any projected adoption. If the natural parents of the prospective
adoptee do not feel that the woman will be an adequate mother,
or if they strongly dislike her, then they may attempt to avoid the
adoption of their child.
A stepparent has no responsibilities to his or her spouse's
adopted child: a child will be mistreated by his stepparent, it is
thought; and certainly a stepparent cannot be relied upon to pro-
vide for a child by leaving him land. Unless a female adopter's hus-
band joins with her in the adoption, there is no reason to believe
that he will not eventually attempt to have sexual relations with
her adopted daughter. Similarly, the motives of a not yet elderly
man wishing to adopt a nearly adolescent girl would be suspect
unless the man's wife also became an adopter. Thus, for a great
many reasons, natural parents who really mean to give up their
child in adoption prefer that the adopter's spouse joins in the pro-
ceedings. Here again it is not a simple case of a couple jointly
adopting a child: each of the natural parents asserts that he is giv-
ing his child to each of the adoptive parents, and four sets of rea-
sons (involving a postulated relationship of kin, or a long-standing
friendship) will be cited to validate each of the four dyadic rela-
tionships involved.
Notwithstanding the efforts made in some cases to involve each
of the adopters and each of the natural parents in dyadic adoption
relationships, it is clear that the fundamental relationship in adop-
tion is the close consanguineal relationship between one of the
child's natural parents and one of the adopters. If the adopting
couple divorces, the adopted child will go with his "primary
adopter" (i.e., in the ordinary case, the adoptive parent who is
most closely related to one of the child's natural parents). The
natural parent who is most closely related to the "primary
adopter" can also be counted upon most to live up to the adop-
tion agreement by encouraging a child to return to his adoptive
parents, should the child show an interest in returning home. The
natural parent who is not closely related to the primary adopter
will usually attempt to get the child back sooner or later. In the
event of divorce, a parent will often try to get back his natural
children who have been adopted by his spouse's relatives. A pri-
mary adopter tends to treat an adopted child differently from the
way in which the child's other adoptive parent treats him. One is
less reluctant to beat an adopted child to whom one is already
closely related consanguineally.
If it becomes known that a child has been asked for in adop-
tion, then others may also express a desire to adopt the child. This
appears to result from a desire to test the quality of a relationship
by comparing what one is able to obtain from a relative or friend
against what others are able to obtain. Often several individuals
will have asked for a child by the time it is born, but some of these
will have changed their minds meanwhile and will not pursue the
matter further, either because they have truly lost interest in what
was only a passing fancy, or because they are reluctant to compete
against other potential adopters. In the face of determined efforts
to obtain a child, an undesirable adopter is told that another rela-
tive has asked first. In extremis a close relative may be asked to
adopt a child in order to provide a pretext for refusing another
If a child's parents have not succeeded in discouraging an unde-
sirable adopter by the time customary for the child to be taken
from its natal home, then a child may be sent to a relative "for
weaning" and the adopter later told that the relative has refused to
give back the child.
It never happens that an adopter relentlessly pursues an adop-
tion request despite all subterfuges designed to frustrate him. The
Nukuoro are very sensitive to rebuff, and it usually takes little to
discourage them in matters of this kind.
Once a child has been adopted by someone, then other prospec-
tive adopters tend to lose interest. It does happen, however, that a
child is given to an adopter from the family of one of his parents
and a prospective adopter from the family of the other parent
feels injured thereby. In such cases, the hurt feelings of the dis-
appointed party may be smoothed if the parent to whom he is
related tells him that "my half goes to you." Since each parent has
full rights in their children each of them may assign their rights to
different individuals. This does not happen frequently, but a dis-
appointed adopter who has been mollified in this way really feels
that the child is "his," even though the child has gone to live else-
Owing to the competition between families on Nukuoro, often
when an adoption has been made on the child's father's side, the
mother's family will feel that the couple's next child should be
given to them. Similarly, when a child has been adopted by some-
one from his mother's family, a member of the father's family
may later claim their "turn." A couple's natural children are often
adopted just before the couple's divorce, usually at the behest of
one parent who wants to keep the child in the family and not lose
it to the 'other side' (that is, to the spouse's family).
Not all adopters are anxious to take their adopted child from
the home of its natural parents. A man may have reasons for
adopting a child but feel that his wife will not take good care of it,
or that there will be friction between the adopted child and the
couple's natural children. An unmarried man is in no position to
take a child away from its natural parents, but may have reasons
for wanting to adopt the child. Similarly an aged person who is
unable to care for an infant may leave it with its natural parents.
In the eyes of the Nukuoro, a child who is not taken away from
his natal home is no less 'adopted' than a child who is taken away
from his natural parents. Nukuoro children move back and forth
freely from the household of one relative to that of another; even
so, they are rarely more than a few hundred yards from their natal
home. A child's residence at any moment is not construed by any-
one as indicating the location of his 'parents'. To a certain extent
all of a child's senior relatives share responsibility for his care.
Also, all of a child's relatives of the parental generation, or older,
have equal authority over him-for example, all are in theory free
to beat him. Thus the question of specific and exclusive authority
over a child (including the right to determine his place of resi-
dence) is not crucial to the Nukuoro notion of 'parenthood'.
A child may be adopted "for" someone else, usually by the
adopter's parent. The "adopter" is not even required to be
present. One man of my acquaintance-unmarried at the time
returned from school to find that his aged mother had adopted his
sister's son as his 'child'. Occasionally, such an adoptee is adopted
"for" an entire group of siblings. It also happens that a childless
couple will adopt one child as principal heir and then adopt other
children (of roughly the same age, possibly full siblings of the first
adoptee) as 'children' to the adopted child. All such arrangements
seem to satisfy parental interest in fixing the course of inheritance
transactions in future generations. Adopted children whose natural
parents have died are bequeathed, like plots of land, to one's heirs,
especially if they are not willed land by their adopters, or if they
have inherited considerable land from their natural parents.
Under no circumstances will 'strangers' be permitted to adopt a
child. Despite the high status on Nukuoro of American govern-
ment officials, a child was refused to a District Educational Ad-
ministrator. Only four children have been adopted off the island
by someone who was not a close relative of either natural parent;
all of these alien adopters were formal 'friends' who had been
assimilated to kinship roles, and all of these adoptions were recent.
Nukuoro living on the atoll are continually being asked for their
children in adoption by friends from other islands, especially from
Ponape. These requests, which are usually for older children, are
generally thought to be motivated by a need for domestic laborers.
Parents are extremely apprehensive about their children leaving
the island: "What will become of the child who is abandoned by
his adopter on another atoll?" they ask, even though no one can
remember a case in which any child was in fact "abandoned." Not
giving up children to 'strangers' is simply a refusal to accept these
strangers as kin. To the degree that the Nukuoro become accus-
tomed to extending their kin networks into other societies (i.e ..
learn to make 'kin' of 'friends') and have some reason to do so,
one would expect that adoption out will increase.
From the foregoing discussion we can conclude that adoption
requests on Nukuoro are extremely frequent, rarely denied out-
right, and usually acceeded to when made by close kin. The large
number of Nukuoro children who are adopted testifies to the pres-
ence of impelling reasons for adoption. So far we have related this
to Nukuoro notions of appropriate behavior among kin-a matter
to which we shall return when we consider in more detail the
Nukuoro ideology of consanguinity and its relationship to adop-
tion. The unwillingness of the Nukuoro to refuse adoption re-
quests outright has been related to Nukuoro norms of social inter-
action: it is impolite to refuse anything to anyone (except
'strangers', who, to be sure, stop acting like 'strangers' when they
begin asking for things), since the asking asserts a kin relationship
and a refusal denies this relationship. Although no Nukuoro is (or
wants to be, or can be) 'kin' to everyone on the island, few are
willing to publicly assert that they are 'not kin'.
It remains then to account for the frequency with which adop-
tion requests are made. Why is it that almost every child born on
Nukuoro is asked for in adoption by someone, and many children
are asked for by several individuals?
Asking for a child and agreeing to an adoption are expressions
of appropriate kin sentiment which give momentary pleasure to
both parties. As brief symbolic acts they are perhaps closest in
spirit to participation in feasts and funerals. To ask for a child is to
assert tentatively a kin relationship, but this does not necessarily
imply that the relationship suggested by the request has already
been clearly established. Many adoption requests are made impul-
sively, often when it is first noted that a female relative (or a male
relative's wife) is pregnant. Such requests appear to be motivated,
at least in part, by a desire to "test" the relationship-to see if the
individual who is asked is really a 'relative' and will act as a relative
should by agreeing to the adoption. Potential adopters make tenta-
tive requests which they do not pursue, in much the same way
that adolescent boys continually assay their attractiveness by casu-
ally propositioning girls. A person making an adoption request is
not committed to pursue the matter, just as the parent who as-
sents to an adoption request is not necessarily obligated to relin-
quish the child. Nothing on Nukuoro is ever quite settled until it
has been reiterated many times.
This compulsion on Nukuoro to test the status of interpersonal
relationships expresses itself in a constant stream of "asking activ-
ity" in which all but a few misanthropes are constantly engaged.
Asking for another's child is perhaps the most effective test of a
relationship, because a child is one of a person's most valuable pos-
There are a few factors which inhibit adoption requests: one is
the reluctance to ask for a child from very distant kin for fear that
the close kin of the natural parents might let it be known that one
is presumptuous. Another restraint on the unlimited solicitation of
others' children is that not asking for a particular child can be con-
strued privately as a denial of kinship with that child's parents,
and thus may be used to slight secretly those relatives with whom
one is not on good terms.
Such restraints somewhat limit the number of adoption requests
for any one child, but the extremely high incidence of adoption
suggests two questions: Are the Nukuoro perhaps anxious to be
rid of their excess children? Do they lack affection for them?
It is generally assumed on Nukuoro that no affection is stronger or
more enduring than the affection of natural parents for their chil-
dren. An adopted child finds unlimited evidence to support this
assumption. If the adoptive parents are not considerate, the child
may assume that it is because he is "not really theirs"; if adoptive
parents are indulgent, the child may conclude that he is not being
treated "naturally." One girl, who was quite well taken care of in
her adoptive household, said that because her adoptive parents did
not frequently beat her she was not really 'loved'!
Even the unwed mother, for whom the burdens of child-rearing
are greatest, is not anxious to get rid of her child. Food is plentiful
on Nukuoro and is readily shared; no one goes hungry for long.
Friends and relatives can be counted upon to help with those few
tasks that one might not be able to do for oneself. However, it is
true of Nukuoro, as it appears to be of Banaba (see Silverman,
chapter 8), that food anxiety is great and people would prefer not
to put consanguineal relations to the test by relying exclusively
upon kin for subsistence. I know of only one case in which a par-
ent expressed an interest in giving up his child in adoption as a
way of avoiding parental obligation. In this case, a man asked the
daughter of his half brother to adopt his son, since his wife had
complained that she was "sick of looking after his children."
Nukuoro informants interpreted this statement more as an expres-
sion of the woman's discontent with her husband than as a state-
ment of unhappiness with her obligations as a mother. There are
cases, especially when a parent is living away from the atoll, in
which a relative is asked to 'look after' dilo ange a child, but this is
not 'adoption' from the Nukuoro point of view.
Nukuoro culture stresses the responsibility of parents toward
children. The birth of a child is a time of joy for all concerned, a
'time when the roofs leak' (since ordinary household chores such
as roof repair are neglected).
Mothers never take the initiative in arranging for a child's adop-
tion, except to prevent a less desirable adoption. It happens occa-
sionally that politically ambitious fathers will hint to close friends
and relatives of high social standing that a child is available for
adoption. It is said however that, in general, the father is as at-
tached to his child as is the mother, but it is only males who com-
pete in community struggles for status and power. I have been told
that there are certain Nukuoro males-perhaps five or six in all-
who do not appear to mind giving up some of their children in
adoption, although I failed to interview any of them concerning
their sentiments on this matter.
In keeping with their political role, males are said to take a
greater part than females take in the arrangements for an adoption
which has been decided upon. However, the actual adopters are
most usually females.
This is partly because child care is wo-
man's work, and no mother will allow her child to be adopted un-
less she is assured that another female will assume the responsibili-
ties of child care.
Thus it appears that there is no way in which the frequency of
adoption on Nukuoro can be linked to a lack of normal parental
affection, or to practical circumstances which make child-rearing
difficult for some. Nukuoro parents appear willing and able to
care for their offspring. Clearly, the motivation for giving up
Nukuoro children in adoption is unlike the motivation for giving
up American children in adoption. At this point we shall turn to
the statements of the Nukuoro themselves concerning their motives
for adoption.
In reviewing the detailed histories of more than a hundred cases
of adoption, one is struck by the absence of any simple pattern of
motivation and circumstance. Each case seems unique in significant
aspects, and there is little agreement either between the parties to
an adoption or among outside observers as to why a particular per-
son adopted a particular child. There are nevertheless several broad
trends which can be recognized.
A Nukuoro is hardly considered an adult until he has at least
one child under his authority. Most adolescent girls (but not boys)
are anxious to have infants in their charge, just as they are anxious
for their secondary sexual characteristics to develop- these are
signs of 'growing up'. In a society which stresses the categorical
authority of the 'mature' over the 'immature', it is not surprising
that young people seek to identify with the generation of 'par-
ents'. Most men who remain unmarried into their thirties also feel
the need to adopt a child. There is also the matter of shame at
childlessness: a sterile man or woman is mocked as 'barren' (lead-
ing some sterile persons to change marital partners often in an
effort to disprove their infertility). Adopting a child seems to allay
the anxiety from this source and in some cases seems to promote
Childless adopters often say that they want a child to 'help
them in their old age', since it is generally known that it is only
one's children who will perform this service. Also frequently ex-
pressed is the feeling that one needs an heir to take over the care
of one's land when one dies, and thus 'keep it in the family'. Land
on Nukuoro is individually owned, and there generally arises a
rivalry between middle-aged siblings, when the loyalty that once
attached to the family of orientation shifts to the family of pro-
creation-a rivalry which is based at least in part on the fact that
one's siblings' children also belong to the families of non-relatives.
To die without a child as heir is to let one's land fall into the
clutches of 'another family'.
Children on Nukuoro are useful in their households. At an early
age a girl can be taught to clean the living area, wash clothes, run
errands, take care of smaller children, help fetch taro, prepare the
earth oven, and keep her mother company. There are only a few
ways in which a very young boy can be helpful, for example, run-
ning errands; but a male child holds out the promise of a house-
hold with all of its huts in good repair, always enough coconuts
and fish, and cash from copra. An infant, of course, is of no help
whatsoever, but he brings joy to all by his mere presence. As one
Nukuoro said, "people like an infant for the same reason they like
a dog-it is something to play with."
When a child has died there is a desire to 'replace' him, espe-
cially if the couple are of an age where it is not likely that they
will have many more children. A household without children of
both sexes is somehow felt to be "incomplete," and Nukuoro
couples with a single child will frequently adopt a child of the op-
posite sex 'so that their child will have a sibling'. Children are
known to get bored unless there are other young children at home.
Parents may adopt other children in order to provide playmates
for their children so that they do not wander.
Although adopters, when asked about their motivation, tend to
stress their need for children to 'help them' later on, those who are
not immediately involved in an adoption stress the significance of
matters involving land. Land should be 'kept in the family', but
there is a great deal of maneuvering to get control over land be-
longing to other families.
'Keeping land in the family' figures prominently in several sorts
of adoptions. There is first of all the case in which a woman at-
tempts to adopt the children of her husband who have been born
to other women, either before or during her marriage, lest the hus-
band, through guilt at neglecting the welfare of his natural chil-
dren, later will them a disproportionate share of his land. Adop-
tion in this case is to protect the interests of a woman's children.
There is also the danger of a childless adult leaving his or her land
to a spouse or to an adoptee from 'another family' (i.e., an adoptee
who is not a sibling's child). The parents of a childless adult will
often arrange before their deaths to adopt one of their grandchil-
dren 'for' their child, to insure that their lands will remain in the
possession of their lineal descendents. The siblings of a childless
adult may also attempt to maneuver one of their children into the
position of heir, lest land 'pass out of the family'.
Adoption may serve to 'get land away from other families'. In
one case, a man, A, adopted the son, Y, of his adoptive brother, B.
The child had been given the name of Z-a land-rich man, now
deceased-the man who had adopted both A and B. The adopter,
A, said that he "felt affection for Y because when he called the
child by name the child reminded him of his beloved adoptive
father, Z." (It is likely that the child was given this name by his
natural father, B, to stress the child's relationship to Z and thereby
to stress B's claims on Z's estate.) A, who had virtual control of
the estate of Z, was disposed to will at least a portion of this estate
to his adopted child, Y, because he knew that the land belonged in
fact to Y's father, B. It was better that it go to "his" (A's) child,
Y, than to B. B, who was not inclined to assert his rights in his
adoptive father's land, obviously felt that it was a coup to arrange
things so that "his" son, Y, (rather than A's children) would get a
major portion of the estate.
A child who is adopted may eventually inherit land from both
of his adoptive parents. He may also inherit land from both of his
natural parents. Most Nukuoro feel that the prospect of one's
child obtaining land from a second set of parents is a strong
motive for engaging in adoption. Both the natural and the adop-
tive parents feel that the child is 'theirs', and whatever land the
child receives from one set of parents is a source of satisfaction to
the other set of parents, who feel that a member of 'their family'
has gotten land away from a member of 'another family'.
Although many observers assert that adoptions between land-
rich and land-poor families are promoted only to obtain land from
the wealthy, this motive would not explain why adoptions in both
directions are equally valued. Adoption of one's child by an im-
portant person shows that a close relationship exists between the
adopter and at least one parent of the adoptee; adopting the chil-
dren of the rich asserts the exact same thing. In the pre-World War
II period when wealth and status were stabilized in the hands of a
few, it was not uncommon for all of the natural children of an
important person to be adopted and for that person to adopt
many children from others. In the earliest portions of the gene-
alogies we were able to collect-those dealing with the nineteenth
century-such wholesale trading of children appears to have been
the norm. This seems to have been important in maintaining at an
optimum size the group of kin whose members were closely inter-
related through cross-cutting adoptive relationships, since many
couples were childless or had only a single child. Now that elemen-
tary families are larger, there appears to be less need for maintain-
ing such closely-knit relationships between them.
Neither the adopter nor the natural parents of an adopted child
feel indebted to each other, indicating that an adoption request is
considered a valid claim, and that complying with the request dis-
charges the obligation established by the claim. An adopter may
feel that he is discharging a long-standing debt by taking a child,
while it is equally possible to give up a child in adoption in order
to compensate a benefactor. Following an adoption the principals
do not exchange goods and services to any degree greater than
would be expected of relatives related clo:;ely enough to adopt
from each other.
Although the group of siblings who have inherited land from a
parent tend more and more, as they get older, to work their land
separately, they still regard all of their lands as belonging to 'their
family'. A group of this sort is extremely possessive, not only
about their lands and other valuable goods but about the natural
children belonging to any of them. The ethic of sharing within the
kin group is balanced, as it were, by a continual effort to purloin
the land and children of other kin groups.
If a child remains with his natural parents, he begins life with
equal ties to both his father's family and his mother's family. But
it is generally supposed that he will eventually align himself more
closely with one side than with the other, so a competition ensues
between the two families for the loyalty and affections of the
child. If the child is adopted by a member of one of these com-
peting families, then his connection with the family of his other
parent is weakened. The side which "loses" in this kind of tug-of-
war can console itself with the fact that one of its 'own' will stand
to inherit an additional portion of the competing family's land,
that is, the portion which the child's adopter is expected to leave
to him.
Sometimes a child prefers to live with relatives (perhaps follow-
ing a period during which the child and his parents were living
with these people). If such arrangements go on for any great
length of time, the parents may assert that the child has been
'adopted'. In theory a child may live anywhere he pleases, but it is
embarrassing for a child to prefer others to his own parents. Claim-
ing that an 'adoption' has taken place mitigates the shame of the
parents of having a child who has abandoned them.
Any Nukuoro is proud to have dependents, since a large house-
hold shows that one is able to provide and keep others happy.
Often, households of important persons contain far more adopted
children, foster children, and dependent adults than can be eco-
nomically utilized for household tasks. It is only to such persons
that a child may be proposed for adoption in the occasional case,
for the reasons already elaborated upon (see p. 136).
'Keeping children in the family' is a motive in placing children
with family members just prior to a divorce. Such children must
be 'adopted' (rather than 'fostered') if the parent is to be able to
deny the possibility of reversing the child's residential status when
asked to do so by the other parent, or a relative of the latter. Thus
giving a child up in adoption is often the only practical way to
'keep it'. Expressed differently, it is important to Nukuoro parents
to maintain physical possession of a child (and his loyalty etc.)
within the group of close kin, even if this means giving up actual
custody. For as long as a child's natural parents retain custody, the
child is subject to the claims of the other parent and the other
parent's kin. An adopter cannot, however, transfer custody of his
or her adopted child to a third party without the permission of the
natural parent who gave up the child in adoption (if he or she is
alive). The situation here is very much like that concerning any
valuable possession. If I (as a Nukuoro) have a musical instrument,
it is difficult to refuse close kin who ask to borrow it; if however I
give it to someone outside my immediate family I can retain the
occasional use of it and also prevent my kin from abusing it: if
someone asks me for it, I tell him to ask so-and-so because it is
'his', while if so-and-so is asked he claims that he has no authority
to lend it since 'it is not his'.
An illustration will perhaps make clearer the kinds of contexts
in which 'keeping a child within the family' may be a relevant
motivation in adoption. The wife of a sterile man became pregnant
during his absence from the atoll. The man was much older than
his wife, who was one of the island's leading beauties, and was not
prepared to divorce her for her obvious infidelity, since he de-
spaired of finding another wife who was equally attractive. More-
over, the man's elderly mother reminded him of their church's
teachings against divorce. The man decided to keep the child as his
own, especially since giving it up to the genitor's family would
have given his wife an excuse to go visit her child at their place. It
was anticipated, however, that the genitor or a member of his
family might come to claim the child. The man therefore arranged
for his brother to adopt the child. After some years, the brother
left the atoll, and the child, now confirmed in his status as a mem-
ber of the family of the two brothers, came to live with the man
and his wife (the child's mother) where he has remained ever since.
There are innumerable other reasons advanced for adoptions in
specific cases. Infants who have been nursed back to health by a
woman other than the genetrix are usually offered to the nurse
since 'had it not been for her the child would not have lived'. In
ancient times it was customary for a couple to summon a diviner
to ascertain which person would be most successful in performing
certain magical acts required to keep a newborn infant from harm.
The magician in such cases often 'adopted' the child for whose
wt!ll-being he was thought to be responsible. In still another case,
which took place at the turn of the century, an infant was adopted
because the chief had ruled that parents must either provide feasts
for the whole island or kill their children. Since it was not known
who fathered the child, he was "adopted" by several unrelated
women (i.e., food was provided for the required feasts 'out of
Of equal interest to the motivations for adopting are the moti-
vations for not adopting. Very few people claim to have taken a
child in order to relieve its parents of the burden of child care;
informants laugh at the idea of such altruism. I once heard the
adoptive brother of an unwed mother say that he wanted to adopt
her child, "to leave the mother free to marry." But a child is cer-
tainly no impediment to marriage for a girl who is good-natured,
hard-working and, by local standards, unpromiscuous. The adop-
tive brother, it should be noted in this case, was under increasing
pressure from his wife to produce a child from somewhere. His
wife had acquired the habit of complaining to others that 'his
semen was rotten' (since they had been married several years and
she had not yet conceived).
An illegitimate child who is not immediately claimed by its
genitor is usually adopted by the mother's parents or siblings to
forfend the possibility that the genitor will claim the child later;
and a woman with illegitimate children is often not unwilling to
allow relatives to take such children in the event of her marriage,
lest their stepfather mistreat them.
In all of the preceeding discussion it may be noted that the
stated goals of adoption are long-term: provision of care in old age,
arranging the course of inheritance, fixing a child's status in just
one of the families to which he is related. It is interesting to note
that these goals are rarely ever achieved in any specific case. Any
reasonable interpretation of Nukuoro adoption must cope with
this fact that the stated goals of Nukuoro adoption appear largely
delusional. But before explaining just how and why most adop-
tions on Nukuoro fail to meet the goals which informants state, it
is important to know precisely how a proper adoption should take
An explicit adoption in which the prospective adopter has secured
from both natural parents a statement to the effect that the child
is 'being set aside for you' dugu maau, in theory, transfers to the
adopter(s) the distinctive rights of natural parents. This does not
mean that al/ 'parental' rights are given up since any kinsman of an
ascending generation has rights and authority over every child of a
kinsman of his own generation; he can, for example, beat any of
them with impunity, and they are free to ask him for things. If a
child's paternal uncle adopts him, the uncle, as it were, becomes
the 'father' but the father remains an 'uncle'. Since the American
categories of 'uncle' and 'father' are terminologically and concep-
tually merged on Nukuoro, the status of a child who has been
adopted, vis-a-vis members of the paternal generation, changes
only slightly. The major shift in an adoption is to shift primary
responsibility for the child's welfare, supervision, inheritance, etc.
from one person to another. For example, adoptive parents should
will land to their adoptive children, just as though they were nat-
ural children, and they should not discriminate in other ways be-
tween their natural and their adopted children.
Similarly, the adopted child is primarily responsible for obeying
and assisting his adopter rather than his natural parents. For ex-
ample, when an adopted child marries, he should first ask permis-
sion of his adopter(s) and then of his other close relatives, includ-
ing his natural parents. Natural parents may ask the adoptive
parents for the loan of the child's labor, just as any relative might;
they are not entitled to ask merely because they are the natural
It is considered morally wrong for natural parents to attempt to
effect the return of a child who has been adopted. It is the adop-
tive parents who are 'tired from raising the child'; the only com-
pensation for this effort is enjoyment of the child's filial affection
and his labor when he is old enough to be of use.
No effort is made to keep an adopted child from learning the
identity of his natural parents, and the child usually has frequent
social contact with his natural parents, since the entire population
lives in a single village which stretches for about two-thirds of a
mile along the lagoon shore of the largest islet of the atoll.
If an
adopted child shows any tendency to dally at the household of his
natural parents, they should send him away, saying: "Is this your
home? Are we your parents? Go away to where your parents are!"
It is also wrong for the natural parents to interfere with the edu-
cation which their child receives in his adoptive home by, for ex-
ample, giving the child precepts or instructions which may be
contrary to those received from the adoptive parents.
Although this is how a proper adoption should work, few adop-
tions work out as they should. One part of the problem (to be
considered at length below) is that very few adoptions are
"proper" in the sense of their having been arranged explicitly (i.e.,
in such a fashion that all parties to the adoption share a common
set of understandings). But there are other difficulties in making
an adoption conform to the ideal, and it is these that we shall now
Although many children on Nukuoro are 'adopted', some adopters
never succeed in taking their adopted child away from the house-
hold of the child's natural parents. There are of course situations,
as we have mentioned, where an adopter does not desire to take
the child to live with him, or it is not practical to do so. But there
are other cases in which an adopter's efforts to take a child are, for
one reason or another, frustrated. Some adopters maintain a con-
tinuing interest in a child they feel they have adopted, even if the
child remains with his natural parents. One woman who had failed
in her efforts to obtain custody of a child she wanted to adopt was
heard to say to the child's mother (with whom the child resided)
"if you continue to beat my child, I will come and beat you!"
Even if a child is brought to live with his adopters, there is a
strong possibility that he will return to his natural parents some-
time before late adolescence.
Children have motives of their
own for returning to their natural parents at adolescence. In an
effort to escape parental discipline at an age when domestic chores
interfere with romantic activities, an adopted child may use visits
to his natural parents' house as a pretext for staying away from
home overnight. The natural parents are usually so happy to see
their children that they provide little or no discipline during these
visits. Several older boys of our acquaintance were able to remain
free agents by playing off two sets of parents in this fashion: when
one set began to make demands, the child would leave for the
house of the other set. Girls are as eager as boys to get away from
parental discipline, since parents usually try to prevent excessive
sexual activity by their daughters, at least for the first two or three
years of adolescence. Many natural parents make a great effort to
maintain close relations with their children, becoming especially
solicitous, it is said, when the child approaches adolescence and is
able to contribute useful labor to the household. Perhaps more
important than cynical calculation is the general appreciation that
a child who remains with adoptive parents through adolescence is
not likely to return to his natural parents later on. A parental
maneuver for a child's affections at this age is motivated by a feel-
ing that it is "now or never." This is one of the many indications
that it is very difficult for parents to give up their children alto-
Adoptive parents cannot and will not make any obvious effort
to restrain a child from visiting his natural parents (although the
child may be beaten for "returning home late"). Since children are
in theory free to live wherever they choose, the natural parents
can always claim that a child has of his own choice returned to
them-the implication being that he was unhappy with his adop-
tive parents.
It seems curious that Nukuoro adopters should believe that
adopted children will grow up happily in their adoptive house-
holds and take good care of the adoptive parents in their old age,
since it is very plain that this is not the usual turn of events. Even
those children who stay with their adopters may be notably lack-
ing in filial affection. There is no certain evidence that adopted
children are any more negligent of their aged parents than are nat-
ural children, but a great many Nukuoro feel that this is so.
The ties of adoptive parenthood which should be equivalent to
those of natural parenthood are clearly not always perceived as
such. Several men are said to have had sexual relations with girls
whom they adopted, while there are only two authenticated cases
of a sexual relationship between a father and his natural daughter.
The frequency of sexual relations between adoptive siblings and
between an adopted child and the spouse of his or her adoptive
sibling is also said to be greater than between natural siblings and
their spouses. I am not at all sure that these generalizations are
true, but it is interesting that many Nukuoro believe that adoptive
relationships are somehow less efficacious than consanguineal rela-
tions in inhibiting sexual relations.
When an adopted child lives with the natural children of his
adopters, invidious distinctions may arise between them. An
adopted child may be taunted with the fact of his adoption. In a
moment of pique a child may say to an adoptive sibling, "Why do
you stay here? Are these your parents? Is this your land?" Most
adopted children seem to acquire the notion at some time that
their natural parents would cherish them more than the people
with whom they are living. If an adopted child who has many
natural siblings tries to return home, the latter may suggest that he
'go back where he belongs' (i.e., to the home of his adoptive par-
ents). It might also happen that just one of his natural parents-the
one most closely related to his adopter-will encourage him to stay
with his adoptive parents, while the other parent subtly encour-
ages him to return to his natal home. Only rarely do parents re-
frain from attempting to provide, during visits at least, some educa-
tion for their natural children, especially concerning whom they
are to treat as 'siblings' and 'elders'. All of these factors tend to
remind the adopted child of his anomalous status.
A parent whose natural child is negligent or disrespectful will
not disinherit the child, since the child is 'after all, of the same
blood' but it happens that an adoptive parent will leave only a
token amount of land to an adopted child who has grown up in his
or her household, if the child does not live up to expectations.
Parents can never be sure that the adopter of their child will
leave him land, and adopters can never be sure that the natural
parents of their adopted child will not leave him land. If a parent
approaches death while the child's adopter is still living, the parent
may leave the child only a token amount of land "just in case." In
this event, especially where the adopter has other natural children,
he may decide that the child has already been provided for by his
natural parents and there is no reason for the adopter to do like-
wise. If an adoptive parent approaches death before the natural
parents do, then the former may assume that the latter will pro-
vide land for the child's future welfare. Both sets of parents are
reluctant to leave land outright to a young child, lest the other set
of parents assume control of the land through the exercise of par-
ental authority over the child. In this sort of situation the child's
share of parental lands is willed to an older guardian, who often
appropriates it for his own children. A child who remains with his
adopter until the latter's death will often receive a fair portion of
the adopter's estate,
but the fact that so many adoptees return
to their natal homes means that many adoptees fail to inherit.
The only circumstance in which adoption works out well is
when a childless woman adopts a child, and her husband (also
childless) adopts the same child, neither of them adopting again
unless the original adoptee dies. Adoption by only one parent, or
by adopters who have children of their own, or adoption of more
than one child work against the success of any adoption. The death
of an adopted child's natural parents is of course another factor
which promotes the stability of adoptive relationships. Only about
14 percent of the population has been "successfully" adopted.
The difficulties in Nukuoro adoption are generally recognized
by the Nukuoro themselves. In interviews designed to elicit local
opinions about the custom, many of those interviewed said that
adoption is a 'bad custom'. These informants cited the unfortu-
nate outcome of most adoptions and felt that it has a detrimental
effect on the children who have been adopted; some said that
'only a child's natural parents know how to raise him properly';
others suggested that 'if God doesn't want some people to have
children then He must have a reason' -although there is no explicit
teaching against adoption in the Nukuoro (Protestant) church.
Those who found both good and bad in the custom felt that adop-
tion is a good thing for everyone if it conforms to the ideals im-
plicit in an adoption transaction (i.e., if the child really becomes
like a natural child to his adopters, and they leave him a fair share
of their land) but that it is a very bad custom in those numerous
cases which do not conform to the ideal. A minority of informants
thought that the custom is a good one: they stressed the fact that
it helps the childless to have children, provides parents for illegiti-
mate children, and may help 'to keep land in the family'. About
half of those polled said that they would be in favor of a law
which prohibited adoption in any and all circumstances. Perhaps
the slowly diminishing frequency of adoption is at least partially
due to an awareness that it is rarely to the advantage of anyone
except the childless, of whom there are a decreasing number.
How is it that the Nukuoro adopt in the face of the knowledge
that the adoption probably will not turn out as planned?
Clearly the mere act of adoption must lend a value to the trans-
action which is independent of later events. If this were not so
then we would have to assume that all of the parties to an adop-
tion transaction were merely hoping against hope that the adop-
tion would achieve its long-term objectives. For this and many
other classes of behavior-most notably marriage-it seems much
more reasonable to imagine that the initiating act itself is of such
importance that the events following are of little immediate import.
All of the difficulties encountered in actual adoptions can be at-
tributed to the difficulty in reconciling the contradictory implica-
tions of Nukuoro kin5hip axioms. Adoption ideals imply that one
set of parents can replace another, but Nukuoro notions of con-
sanguinity imply that there is no possible substitute for one's nat-
ural parents. In this section we shall elaborate on how Nukuoro
notions of consanguinity lead to this contradiction.
Kinship on Nukuoro implies a host of vaguely defined recip-
rocal obligations, the most important of which perhaps is the
obligation to share. One should be always willing to assist a kins-
man with one's time, effort, money, knowledge, skill, shelter,
tools, canoe, food, the products of one's land, and-not least im-
portant-one's children. The ethic of sharing is continually rein-
forced by offers to share, even when there is no apparent need. It
is merely polite in Nukuoro to offer food to close kin on each and
every occasion on which one is eating and kinsmen are present.
One's children are continually being offered in much the same
way-to run errands, to help prepare food for feasts, etc.
One of the ways in which an individual discharges the diffuse
obligations of kinship to parents and siblings is by teaching his
children to treat these kinsmen with the same respect due to the
parent. Children are reminded daily that grandparents, aunts, and
uncles should be treated as parents. Children are often sent to help
their relatives, but it is also stressed that the latter have obligations
to treat these children as their very own. The child is made to
understand that one or more of these close relatives may someday
replace one of the child's natural parents, in the event of incapac-
ity or death. The children of siblings are encouraged to treat each
other as full siblings, against the day when they may need each
other's help. Thus it is part of the daily experience of every Nuku-
oro to be reminded that the kin group provides the most security
for all when its norms of sharing and solidarity are subscribed to:
two cousins were admonished before they left the island to put up
a front of mutual solidarity whenever they were in the presence of
foreigners, for "if others see that you are united they will be afraid
of you, while if they notice you are divided between yourselves
they will not be afraid of you and may even kill you."
It is within this matrix of sharing and social identification that
adoption takes place. Close kin have the right to ask for one's
child, just as they have the right to ask for the loan of a cooking
utensil. One has an obligation to a close kinsman to give up one's
child in adoption; to refuse to do so is an assertion that the kin-
ship tie does not exist. Sharing of children is facilitated by an ide-
ology which stresses that the primary kin of one's parents are "the
same as" one's natural parents. Thus, while children are frequently
adopted, they are almost invariably adopted by someone whose
status prior to the adoption approximates that of a natural parent.
Ideal behavior between kin is so self-sacrificing and altruistic
that no one can consistently live up to the ideals. If the Nukuoro
were to judge their kin on the accumulated evidence of the fre-
quent daily encounters which are inevitable in a small village, they
would have to conclude that no one treats them as a kinsman
should be treated. Nukuoro family life is such that children grow
up with a burden of hostility toward parents and siblings, while
the norms of appropriate style in social interaction make it ex-
tremely difficult to express any of this hostility overtly. The
promise of kinship-affection, loyalty, cooperation-is great, but
most individuals need constant reassurance that any particular
kinsman is in fact a person who will provide these things. Consan-
guinity defines who should be a kinsman, but actions indicate who
one's kin really are.
Since it is customary in the anthropological literature for "kin"
or "kinsman" to be reserved for statuses defined on the basis of
consanguineal relationships, we shall, at the risk of some confu-
sion, appropriate the term "relatives" to indicate the individuals
whom a Nukuoro thinks of as acting as a kinsman should act. In
the ideal case, all "kinsmen" would be "relatives" and vice versa.
But for various reasons there is only partial overlap between these
two groups from the viewpoint of any individual Nukuoro. Many
"relatives" are not "kin" and many "kinsmen" are not "relatives."
Whereas we might say "he is my uncle" (on the basis of geneal-
ogy), "but he doesn't act as an uncle should act," the Nukuoro, in
a similar circumstance would say, "he is not my uncle" (that is, he
does not treat me as an uncle should), even though it may be ad-
mitted that "he is my uncle" (admitting the consanguineal rela-
tionship which is involved). Similarly, a Nukuoro might say, "she
is not the woman who gave birth to me, but she is my 'real
mother'," in speaking of a woman who has treated him better than
his own mother did.
To the above generalizations must be added the notion that
one's natural parents will love and cherish one more than any
other human being. No matter how much they neglect and abuse
one from day to day, it is thought that in time of need they will
be on hand while others may not be. Thus, however much a child's
adopters are like his natural parents, they are not exactly the same
thing. Adoptive parents may discharge their obligations with even
more concern and exercise their rights with even less insistence
than do natural parents, but they lack the immediate biological
connection which culturally defines the closest sort of kinship,
and which validates symbolically the most compelling degree of
It is therefore inevitable that adopted children should want to
return "home," to live with the only people who, by definition,
'love' them, i.e., their natural parents. It is only in the few cases
where the natural parents show no parental concern whatsoever,
and the adoptive parents succeed in maintaining the outward
forms of unconditional affection, that the child's tendencies in
this direction are minimized.
Parents feel that their natural children will take care of them in
their old age and carry on the family line through stewardship of
family lands. The factors which motivated parents to give up a
child in adoption wane in importance as the parents grow older. It
is a rare parent who does not make some effort to 'get the child
To put the whole matter very simply, the Nukuoro have certain
ideas about the character of the relationship that is attributed to
biological parenthood, and these ideas cannot be reconciled with
practices, such as adoption, which are related to ideas about the
proper sort of relationship which should exist between close kin.
It is important for Nukuoro to have good relations with close
kin-to be without kin is to be alone in a hostile world-and the
form of good relations is more easily maintained by periodic mag-
nanimous acts, such as giving up one's child in adoption, than by
innumerable daily acts of lesser symbolic value. The full explica
tion of this principle would take us far afield, but it explains, for
example, why most Nukuoro provide a great deal more help than
is asked for on those few occasions when they accede fully to the
demands of kin. If for example a woman has been asked to nurse a
sick relative for a few hours, she may stay for a week; if a man has
been asked to provide a few drinking coconuts, he may send the
fruit of an entire tree.
An adoption does not rearrange consanguineal relations. An
adopted child acquires no additional relationships-apart from the
relationship to his adopter-merely by virtue of his adoption. The
adopter may undertake to create links of relationship between the
adoptee and the adopter's close kin-indeed the adopter is
expected to do so for those living in the same household: the
adoptee should be taught to regard the adopter's spouse as 'parent'
and the adopter's other children as 'siblings'. Sexual relations be-
tween the ad op tee and any of these persons would, in theory, be
considered as bad as though they were with blood kin. The
adopter may also undertake to introduce an adoptee into relation-
ships with other close relatives, but there is no compelling reason
for him to do so; the child is usually consanguineally related to
many of these people in any case.
Unless a relationship is specifically created, then none is pre-
sumed to exist merely by virtue of adoptive links. For example,
the brothers of a girl adopted into a large family regularly have
sexual relations with the girl's adoptive sisters, even though their
mother was also adopted into the same elementary family. If an
adoption had any implications at all for preexisting consanguineal
relationships, then these boys would be 'brothers' to the girls with
whom they dally; but as they do not choose to treat the girls as
'sisters' and as the actual consanguineal relationship is not very
close, no one feels that the relationship is improper. Similarly, sex-
ual relations between non-kin who have been adopted by the same
person are not disapproved unless the parties grew up together in
the same household (i.e., unless they "became kin" through years
of treating each other as close kin).
When a child is adopted, as we have said, the natural parents
transfer certain of their rights in the child which are distinctive to
natural parents, but there are certain rights which cannot be trans-
ferred because they are inherent in the character of natural parent-
hood. One of these is the right to determine who shall have their
child. If an adopter gives a child to another person without first
asking permission of the natural parents, then the latter will surely
go and claim the child back. It is the natural parents who are ulti-
mately responsible for their child's welfare, and they may claim
their child from any adopter who is mistreating him. If an adopter
dies, it is expected that an adoptee will return to his natural
Thus it would appear that, notwithstanding the ideals of adop-
tion which are based on the fiction of social identity within the
kin group, an adopter can never achieve full "parental" rights,
since an adoptive parent does not have the type of consanguineal
connection to the child which establishes these rights. The re-
sponse to this implicit contradiction in Nukuoro notions of kin-
ship appears to be the maintenance of situations which are
Since 'adoption' on Nukuoro may involve as little as a single re-
quest for a child-a request which may not be followed by any
other concrete evidence of the adoption-most people do not feel
entirely certain about who has been adopted by whom.
interviews contain many statements such as "I don't know
whether I have been adopted; some people tell me that I have been
adopted by so-and-so, but I have never seen any real indication
that I have been adopted." Often a child living with his natural
parents first learns that he has been 'adopted' when older children
tease him by calling him by the name of a 'parent' whom he
does not know (a form of joking called hai maadua). Parents appear
equally vague on this point: A woman whose child was living with
the sister of her first husband and the sister's husband (who was
the son of the genitor's mother's sister) was asked whether the
child had been 'adopted'. "I don't know," she said, "they just
came one day and took the child 'overnight' and never brought
him back. I have been meaning to talk to them about the matter
[the boy is now twenty] but haven't yet gotten around to it." The
child's genitor asserted that he and his wife had taken the child as
an infant [a fact not mentioned by the child's mother] and that
the child had later gone of his own accord to live with his uncle.
The woman (who is childless) with whom the child is living
claimed that she 'adopted' the child and plans to leave him a major
portion of her land, but the woman's sister claimed that the child
was in fact adopted by the husband. The child spends considerable
time in the households of all of these people.
Even when a prospective adopter fails to take the child in adop-
tion, it is possible for him to maintain that the child has been
'adopted', especially if the child has not been adopted by someone
else. After all, the adopter asked for the child and the request was
granted. Such an adopter may give an occasional present to the
ad op tee, or take the child for an occasional brief visit. An adopter
of this sort may not manifest any interest in the child for many
years and then suddenly begin to show an interest. If such an
adopter does not have many natural children, the adoptee may
even inherit a bit of land from the adoptive parent. Thus it occa-
sionally happens that a child becomes 'adopted' in the eyes of the
community only after the death of the adopter and the division of
the adopter's estate.
It might seem strange that the Nukuoro are often vague about
matters such as adoption, about which we insist upon being spe-
cific, but there are many reasons for maintaining an ambiguous sit-
uation. If a child is adopted and later returns home, "face" can be
maintained by claiming that the child was never "really" adopted.
Similarly, although all 'children' of an adult should inherit a share
of his or her land, adoptees can be disinherited by claiming that
they are not "really" the children of their adopters. As in so many
areas of Nukuoro life, inexplicitness appears contrived by both
sides to allow for room to maneuver later on. An excellent
example of this is the following situation: The young son of A is
living with A's sister and her husband, B. B and his wife, who took
the child temporarily when A's family was sick, had no young chil-
dren and wanted a young child around the house. A and his wife,
who have had nine children in fifteen years of marriage were
happy enough not to have the daily responsibility for the child.
The child continually wants to visit his natal household. A claims
that the child is only allowed to come to his natal household to
"visit with his siblings since there are no small children in his adop-
tive household and he gets bored." On the other hand B takes the
child abroad with him, and otherwise acts as though the child were
truly his. None of the four persons most intimately connected
with this situation will say whether or not the child has been
'adopted'; all are maintaining a wait-and-see attitude. In this con-
rwction it is interesting to note that although the Trust Territory
Code has provided since 1953 a simple procedure for registering
legal adoptions with the local Community Court Judge, and most
Nukuoro are aware of this procedure, no one has yet availed him-
self of it.
To summarize the data, there is not only a lack of concensus on
Nukuoro about why a particular adoption has been made, there is
also little general agreement on who has been adopted, and by
whom, or even whether any 'adoption' has in fact taken place.
As Levy points out (chapter 4), the repeated experience of adop-
tion must communicate some sort of information to the individual
about the premises of his culture. My view is that the most impor-
tant information which is encoded in this sort of life experience is
that relatives should share with each other. Any specific act of
adoption encodes a number-probably a very large number-of
messages about the relationship of the principals, but it is here
maintained that the cultural message which all adoption acts com-
municate to all of the parties concerned is that relatives are inter-
dependent and that the maintenance of this network of inter-
dependency must take priority over the wishes of individuals, even
such strong wishes as attach to one's natural children.
Adoption, as it is practiced on Nukuoro, is an especially appro-
priate vehicle for the expression of cultural norms of kin-group
solidarity in that, by obliging parents to gi1e up their children, the
supposition that children belong exclusively to their natural par-
ents is modified in the direction of recognizing a multiplicity of
claims. The claims of particular parents and particular children on
each other must give way in the face of the authority of all elders
and the requirement that siblings should cooperate. To put the
matter another way, "adoption" reiterates not only the principle
of "group solidarity" but emphasizes the particular dimensions of
this solidarity.
It is surely the efficiency of adoption in encoding of this sort of
information, while avoiding the encoding of "wrong" information
(i.e., information running counter to important cultural axioms),
that accounts for the presence of adoption rather than some other
cultural form in this society. One bit of wrong information that
adoption comes perilously close to encoding is a denial of the as-
sumption that natural parenthood, in and of itself, establishes af-
fective ties which cannot exist between individuals who are not
related in this fashion. The encoding of this "false" information is
avoided by the outcomes of specific adoptions: while ostensibly
trying to make adoption work out, the Nukuoro are reiterating by
a reductio ad absurdum type of demonstration that it cannot pos-
sibly work well in the usual case.
Adoption reinforces the solidarity of the kinship group by
symbolizing the disinterested social sharing on which solidarity de-
pends, but the practice of adoption reiterates for everyone the im-
portance of "biological" (i.e., consanguineal) kinship.
If the Nukuoro did not live in a compact village on a remote
atoll, then it would of course be impossible for them to "give up"
their children in adoption with the arriere pensee of not giving
them up at all. It might be hypothesized that one of the reasons
why the Nukuoro must play a social game of this magnitude of
complexity is that most individuals absorb a great deal of informa-
tion from the thoughtless acts of their kin from which it is pos-
sible to conclude that no one much cares about them. They are, in
a word, "insecure."
Unlike the situation on many other Oceanic atolls, land on
Nukuoro is owned by individuals. Thus the system of land tenure
does not allow for the expression of norms of kin-group solidarity
in day-to-day cooperative use of jointly-owned property. Expres-
sion of the unity of the kin-group must find another outlet.
The frequency, extent, and persistence of adoption are ac-
counted for by the extremely economical way in which this insti-
tution encodes information of crucial relevance to the participants
in Nukuoro culture.
Since American adoption conveys an en-
tirely different meaning and is substantially different in form it
may not be appropriate to think of both institutions as "adop-
tion." If the American custom is labelled "adoption" then
the Nukuoro custom might be appropriately labelled "adoptive
fosterage" (or some such). As an institution it depends on the sup-
position that a child and his natural parents will never be separated
Nukuoro is a Polynesian outlier located in the Trust Territory of the
Pacific Islands, 164 miles north of Kapingamarangi and about 270 miles
southwest of Ponape. Contact with Europeans has been continuous for
almost one hundred years and has resulted in radical changes in technol-
ogy, the establishment of the Christian religion on the island, and the use
of money derived from copra production. Subsistence is, however, little
changed: taro and fish are the principal food items. In 1965 there were
276 individuals on the atoll, 20 (7 percent) of whom were ethnically not
Nukuoro. Nearly everyone was living m one compact village located on
the largest islet of the atoll. The data on which this paper is based was
collected on Nukuoro Atoll from September 1963 through August 1966
and July 1967 through September 1967, during which periods a total of
thirty months were spent in field research. Much of the data in this paper
appeared first in my dissertation on Nukuoro kinship (Carroll 1966) to
which reference may be made for further general ethnographic informa-
tion. It is a pleasure to acknowledge the support of the National Institutes
of Health (U.S. Public Health Service) for support of the field research.
My wife, Raymonde, collected much of the survey data on which this
paper is based, and she compiled all of the numerical data which I have
included here. Our many arguments about appropriate interpretations of
the data have clarified for me many issues which might otherwise have
remained obscure or misunderstood. My Nukuoro assistant, Topias Soulik,
wishes to record that he agrees with the ethnographic generalizations con-
tained in this paper. I am indebted to Michael Lieber, Harold W. Scheffler,
David M. Schneider, Albert J. Schtitz, and Martin G. Silverman for de-
tailed comments on earlier drafts of this chapter.
2 Numerical data in this chapter refer to one of two distinct populations:
the "total resident population" (hereafter abbreviated TRP) includes all
of those native Nukuoro who were resident on the atoll in March 1965.
Wherever appropriate, the population from which data are drawn is the
"total known population" (TKP) which includes all of those native
Nukuoro, living or dead, irrespective of present residence, about whom
the relevant information is available. This includes all native Nukuoro
born after 1891 and a few dozen others about whom something is known.
Numerical totals listed below are relevant sub-sets (all determinate cases)
for the population indicated.
Excluding those less than five years of age, 61 percent (I 22/201 TRP)
of the resident population had been adopted at least once. This figure is
almost exactly the same as that for the entire population: 57 percent
(247/468 TKP) one year of age or older (or who had reached this age at
death) had been adopted. 9 percent (23/247 TKP) of those adopted had
been adopted more than once: seventeen individuals had been adopted
twice, five had been adopted three times, and one individual had been
adopted four times.
3 67 percent (32/48 TRP) of the households contained one or more
adopted children.
4 49 percent (59/120 TRP) of the married adults over twenty-nine years of
age have adopted a child; another 43 percent (52/120 TRP) of these
adults have given up a child in adoption. 6 percent (7 I l 20 TRP) of the
adults have neither given nor taken a child in adoption but have married
someone who had previously adopted a child.
5 There were five unmarried men over twenty-nine living on Nukuoro at the
time of the census.
6 All Nukuoro women (TKP) who are now 25-34 years of age have demon-
strated fertility by giving birth to at least one child. 52 percent of the
women 35 years of age or older (TK.P) were either sterile or "relatively
infecund" (for a definition of this term see Carroll 1966, p. 63f).
A declining rate of adoption on Nukuoro is demonstrated by the fol-
lowing data: 13 percent (7/52 TKP) of the Nukuoro 40 years of age or
over were definitely not adopted; 38 percent (20/52 TKP) of the Nukuoro
20-39 years of age were definitely not adopted; 54 percent (52/97 TKP)
of the Nukuoro 5-19 years of age were definitely not adopted.
7 The relationship of contemporary adopters to adoptees who were actually
taken to their adopters' homes is as follows (the number of cases is indi-
cated in parentheses: quotes indicate that the relationship is classifi-
catory): F'F' (2); F'M' ( 4); M'M' (2); M'F' (3); MM ( 12); MF (7); Mgp (O);
FF (4); FM (9); Fgp (l); MZ (14); M'Z' (4); MB (5); M'B' (7); FZ (l2);
F'Z' (8); FB (5); F'B' (4); 'sibling' (5); 'child' (O); more distant relation-
ship (34). Analysis of these figures indicates that 49 percent (69/142) of
the adopters were grandparents or parent's siblings; 27 percent (39/142) of
the adopters were classificatory 'grandparents', 'parents' or 'siblings'; while
24 percent (34/142) of the adopters were more distantly related. These
data demonstrate the tendency toward adoption by close relatives.
8 Of those adopted 49 percent ( 113/232 TKP) were male and 51 percent
(119/232 TKP) were female. These percentages are not significantly dif-
ferent from the sex ratio in the society as a whole: females account for 45
percent (236/523) and males account for 55 percent (287/523) of all
those Nukuoro born between 1891 and 1964.
9 Relatively few children are in fact adopted after they are one year old: 90
percent (121/134 TKP) of the adopted children actually taken from the
household of their natural parents by an adopter were taken at or near
their first birthday; 4 percent ( 6/ 134 TK.P) were taken considerably after
their first birthday but before their fifth birthday; 6 percent (7/134 TKP)
were taken after their fifth birthday.
I 0 Analysis of the data enumerated in note 7 indicates that adoption by
'grandparents' ( 43/ 142 cases) is significantly more frequent than adoption
by 'parents' ( 59/ 142), when we consider that the incidence of the latter in
the total population is much larger. There was only one case of adoption
by a 'great-grandparent', and five cases of adoption by someone in the
same natural generation as the adoptee. There were no cases of adoption
by someone in a natural generation lower than that of the adoptee.
Of the children adopted by 'grandparents' 37 percent (16/43) were
adopted from real or classificatory 'sons' while 63 percent (27 /43) were
adopted from real or classificatory 'daughters'. The imbalance in these
figures is attributable to the adoption of illegitimate children and children
of divorced couples from their mothers, and the apparently greater re-
sponsiveness of females to parental demands of every sort.
11 There were no apparent laterality preferences in these adoptions. In adop-
tions arranged between real or classificatory siblings, 30 were of a
'sister's' child and 29 of a 'brother's' child; 49 children were adopted by
patrilateral relatives and 54 by matrilateral relatives (see note 7 for de-
12 We label this sort of child care "fosterage." It is extremely frequent on
Nukuoro. Literally every Nukuoro has been fostered by several adults for
extended periods. This reflects the degree to which "parental" responsibi-
lities are shared among a wide range of kin. Although fosterage is impor-
tant from many points of view-most notably its effects on socialization
processes-it is not germane to the central issues of this chapter and is not
discussed further. It should be noted however that some fosterage goes on
for so long that it becomes tantamount to adoption, especially when in-
heritance is concerned. A child who has grown up in a foster parent's
household and is dutiful in discharging filial obligations may be treated as
a natural child in the will of his foster parents if they die before the child
has gone to live elsewhere and if the child does not have living natural
parents who might acquire control over such land. Fosterage is an
obligation of kinship which everyone accepts willingly.
13 68 percent (93/136 TK.P) of the adopters were female. (Cf. the sex ratio
in the population as a whole, note 8).
14 The Nukuoro have noted that sometimes a couple who have remained
childless for several years will have a child of their own not long after the
adoption. The therapeutic effect of adoption on psychogenic infertility in
the United States does not appear to be established conclusively (see
Hanson and Rock 1950; Weinstein 1962).
15 Thus Nukuoro children do not exhibit the "genealogical bewilderment"
which is common among American adoptees (see chapter 1).
16 35 percent of those 15 years of age or older, 23 percent of those 10-14,
12 percent of those 5-9, and none of those under 5 (all figures TK.P) had
returned to their natal home from the homes of their adopters for reasons
other than the death of the latter.
17 Only 7 percent of those who stayed with their adopter until the latter's
death received no land from their adopters; 36 percent received less land
from the adopter than from their natural parents; 56 percent received
more land from the adopter than from their natural parents (all figures
TK.P). Many of the adopted children who failed to receive a fair share of
their adopter's estate were not jural adults at the time of their adopter's
death, and thus their land was placed in the hands of a guardian.
18 25 percent of the adoptees who had voluntarily returned to their natural
parents received no land from their adoptive parents at the latter's death;
21 percent received less land from their adopters than from their natural
parents; only 4 percent received more land from their adopters than from
their natural parents. Half of the cases of returnees were indeterminate
since the adoptive p ~ e n t s were still alive and had not yet willed their land
to anyone (all figures TK.P).
19 This figure is arrived at as follows: if 57 percent of a large enough sample
of Nukuoro will be adopted (see note 2), and 65 percent of this total will
be adoptions that are more than 'merely verbal' (see note 20), and if we
calculate further that 65 percent of the adoptees who have actually gone
to live with their adopters will not return to their natal home (see note
16), and 57 percent of these will receive more land from their adopters
than from their natural parents (see note 17), we conclude that 13. 73 per
cent of the adoptions will involve (a) transfer of a child to the household
of his adopters; (b) permanency of residence therein; and (c) the inheri
tance of a considerable portion of the adopter's land. This figure for "suc-
cessful" adoptions compares with that for other Oceanic societies where
adoption is practiced in a clear-cut fashion.
20 On Nukuoro 35 percent (72/206 TK.P) of the adoptions were 'merely
verbal' i muna donuhuu, i.e., the child was neither taken from his natal
home, willed land, nor otherwise trea4:d as a natural child.
21 The notion of "communicational efficiency" might usefully be applied to
the "incest taboo," rules of prescribed marriage, principles of incorpora
tion into descent groups, and the like.
Michael D. Lieber
Kapingamarangi is the southernmost atoll in the Eastern Caroline
It is included in the Ponape district of the U.S. Trust Ter-
ritory of the Pacific Islands. One of the two Polynesian outliers in
Micronesia, its inhabitants are Polynesian in language, culture, and
physical type (Emory 1965:76-78). The land mass of 276.4 acres
(Wiens 1956: 10) supports a population of 360, which is rapidly
growing. There are also 390 Kapinga presently living on the island
of Ponape.
Although Kapingamarangi was among the first of the Polynesian
atolls to be discovered by Europeans, it was one of the last to have
significant contact with them; the first English ships began to visit
the atoll in the late 1870s. Despite the many changes and innova-
tions brought about by contacts with the English, Germans,
Japanese, Americans, and Micronesians (see Lieber 1968) much
that is traditional in the society still remains, including the institu-
tion of adoption.
The approach to the description of adoption and fosterage em-
ployed here is integrative, combining analyses on the cultural,
social, and individual levels. The description is focused not only on
how children are adopted and who adopts them, but also on why
children are adopted. The inquiry into motivations for adoption
necessarily involves all three analytical levels.
Adoption and fosterage are behaviors exclusive to relations
among kinsmen; they involve the transference of parental status
and role from a parent to one of his consanguineal relatives. Adop-
tion also involves the inheritance of land, which also is exclusive to
relations among kinsmen. Thus, an understanding of adoption and
fosterage necessitates some prior understanding of Kapinga family
organization and land tenure.
Kapingamarangi descent is cognatic, and kin terminology is gen-
erational. Emory has presented the kin terms in his monograph
( 1965: l 13-115 ), and rather than reiterate them here, we will
discuss only those which are relevant to our present purposes. All
Kapinga kin terms are used in reference exclusively. If it is neces-
sary to get a relative's attention, he is addressed by name. While all
kin terms may be extended widely to relatives in appropriate gen-
erations, the terms for 'child' and 'sibling' are those extended most
often to relatives outside the nuclear family. The terms for 'father'
or 'mother' are less frequently extended-mainly to an adoptive
parent, to a relative with whom one wishes to emphasize his rela-
tionship for a particular purpose, or metaphorically, to an ances-
The term tama 'child' is widely extended in normal usage. A
Kapinga may refer to his own children and to children of relatives
in his own generation as his tama, and he is especially likely to do
so if any of these are co-resident with him. A person may also
refer metaphorically to a younger sibling or cousin whom he is
taking care of as his tama. People refer to their own and their
siblings' grandchildren as their tama. The possible extensions of
this term are important when we consider that those to whom one
may refer as his tama-aside from his own children-are those
whom he may adopt.
The 'sibling' term also may be extended widely. It is generally
used to refer to any two related people in the same generation.
There are actually two different terms involved in the sibling rela-
tionship, and they are used in a complementary manner. The term
tuahina is used to ref er to a kinsman. It can be modified by a term
to indicate sex, for example, tuahina taane 'brother' and tuahina
ahina 'sister'. A person's tuahina is usually the one from whom or
through whom a child is adopted.
The term hakahanau is also a sibling term, but it is used to
denote the relationship between two or more people rather than
to refer to a specific person. The difference in the use of the two
terms is the difference between saying, "He is 'my brother'
tuahina," and, "We are 'brothers' hakahanau." Hakahanau can
designate not only the relationship between two or more people
but also the sibling set as a whole.
The sibling set is the basic social structural unit of the society.
It is corporate with respect to economic activities, land tenure,
and ceremonial-religious activities. The sibling set is the funda-
mental land holding unit on the atoll. Since food production, con-
sumption, and cash income depend on land ownership, the atoll
economy is largely controlled by sibling sets. The sibling set func-
tions as a unit in life cycle ceremonies, major feasts, and work
projects involving any of its members. The sibling set rarely resides
as a unit after its members are adopted or married.
The sibling set ideally consists of both males and females. If a
couple has only sons or only daughters, then at least one child of
the opposite sex will be adopted. There are both long-term and
short-term reasons for doing this. Children are expected to do cer-
tain kinds of household work. These tasks require both male and
female labor. Boys, for example, climb trees for coconuts, bread-
fruit, and pandanus fruit, while girls are expected to help with
cooking and the care of children. Child labor is of enough impor-
tance that children are sent from the atoll to help relatives living
on Nukuoro and Ponape.
Balance of the sexes is also important to the sibling set. Each
sibling set has both a male and a female leader or taki. They are
usually the eldest male and female among the siblings, although
the parents can appoint a younger sibling to the position if this is
considered necessary or desirable. The position of taki is conceptu-
alized in terms of parental status and role; thus the male taki has
the status and role of 'father' of his younger siblings, while the
female taki assumes those of 'mother' of her siblings. In their roles
as stewards, the taki exercise control over property (land and
houses) owned corporately by the sibling set. They are also re-
sponsible for organizing major feasts and work projects involving
members of the sibling set.
In general, the two taki maintain continual communication and
cooperation among the members of the sibling set. At his demise,
the male taki is succeeded by a younger brother. When all the male
siblings of that generation have died, then the position of taki may
pass to the eldest son of the original taki or to a son of the last
taki. The female taki is usually succeeded by her own daughter or
by a daughter of the male taki.
Other than the sibling set, the most important kind of kin group
is the matahanau, which consists of all the descendants of a person
through both males and females over two or more generations.
Used in this general sense, the term denotes simply a cognatic
stock(Radcliffe-Brown 1956:22;Freeman 1961:199). Used in ref-
erence to a specific group of people, however, matahanau may
refer to groups of varying size and degrees of genealogical inclu-
siveness which are corporate for various purposes. A matahanau,
for example, is often a land-holding unit, but the group which is
corporate for this purpose rarely exceeds three generations (below
the apical ancestor) in depth. The preparations for a wedding feast
are carried out by several matahanau, each of which includes a per-
son's descendants for five or six generations. The matahanau in
this case includes a greater number of kinsmen than does the land-
holding group, and there is considerably more variation in gene-
alogical distance among its members. Members of the matahanau
and sibling set work together on the various kinds of projects
described above, but they do not reside together. The family units
comprising the residence group, on the other hand, reside together
but ordinarily work separately.
Most of the population of Kapingamarangi lives on two adjacent
islets, each of which is divided into a number of homestead sites
called kuongo. Each kuongo has definite boundaries and several
different kinds of houses included within it; some serve as resi-
dences and others are used for work purposes. There may be one
to four married couples (or widowed persons) living with their
dependents on a kuongo. Each couple has its own residence where
they and their pre-adolescent children sleep. Adolescent girls
usually sleep separately from their parents in another house on the
kuongo, while adolescent boys sleep either in the men's house or
in a canoe house on their own or a friend's kuongo.
Postmarital residence is ideally uxorilocal. A married couple
actually has a number of potential residences, however, and the
couple may reside in several residence sites or kuongo throughout
their married years. While the length of the stay on any one
kuongo varies, the couple usually resides on one of them most of
the time. The uxorilocal rule is generally followed, a married
couple residing with the wife's mother's or, in some cases, the
wife's father's family, with alternatives available if a move is felt to
be necessary.
The residence group, then, is composed basically of several
closely related women with their in-married spouses and children.
In the case of a couple residing virilocally, there may be several
women on the kuongo, one (or more) of whom is affinally related
to the others. The women who reside together often work to-
gether on such tasks as cooking and mat making, but each woman
usually prepares her own food for her husband and children. One
rarely sees a man on the kuongo during the day, and the men who
reside together rarely work together.
The residence group is of sociological importance in several
ways, but for the present purpose only one of these will be d . .s-
cussed. Residence determines an individual's primary family loyal-
ties. While an individual is related equally to his father's e:..1d
mother's families by descent, his primary allegiance is to those rel-
atives with whom he resided as a child. This allegiance is ~ x p r e s s e
by his participation in major feasts in which several matahanau are
involved. While an individual may be related to two or more
matahanau participating in a feast, he will work with only one; his
choice depends upon his feeling of primary allegiance. Thus, if a
man has resided with his mother's family, he will usually work
with relatives in his mother's matahanau.
Since adoption tends to shift the residence of a child from his
natal kuongo to another one, it may shift his primary affiliation
from one matahanau to another. Such affiliation is expressed not
only in feasts but also in the utilization of land rights.
Kinship and land tenure on the atoll are so closely interrelated
conceptually as to be symbolically isomorphic. With respect to
kinship and land, rights in personam and rights in rem are merged
rather than distinct. Thus, people who are and consider themselves
to be related by descent will also share an interest in one or more
pieces of land. Conversely, those people who share an interest in a
piece of land are, by definition, kinsmen. The norms of behavior
regarding utilization and disposal of land are at the same time
norms of kinship behavior. The details of land tenure and inheri-
tance simply elaborate these principles.
Land which is used for plantations and residence sites is distin-
guished from that used for taro plots. The Kapinga term for 'land'
refers to plantations and residence sites, while 'taro plots' are re-
ferred to by a separate term. This distinction involves not only a
differentiation in productive uses ,
but also a distinction between
the kinds of rights associated with 'land' and 'taro plots'. 'Land'
may be owned by individuals or by corporate groups; 'taro plots'
are owned by individuals. While there are different kinds of rights
over 'land', there is only one kind of right over a 'taro plot', that
is, unencumbered individual ownership. The two terms, 'land' and
'taro plot', will be used in their Kapinga sense throughout this
With respect to rights over land, a further distinction must be
made between ownership rights and use rights. Rights of owner-
ship over a plot of land include regular use of its plant life for
food; copra; leaves and wood for construction; rights of cultiva-
tion; and rights of occupation.
The rights over the eventual disposal of the land rests solely
with its owners.
A given plot of land may be owned at a given
time by an individual, by a sibling set (each member holding equal
rights to its use), or by several related sibling sets (i.e., by a mata-
lianau, each member of which has equal rights to its use). Over a
given period of years tnat same plot will usually have been held by
all three categories of owners in a kind of developmental cycle.
The larger the number of people regularly using a given plot the
greater is the likelihood that the land will be divided among thein,
tlwreby initiating a new cycle. An individual owner of a plot of
land exercises his rights to it at will. Members of a corporate land-
owning group, a sibling set or a matahanau, exercise their rights
through a taki, either obtaining his permission prior to using the
land (especially for copra making) or notifying him after using it.
In the case of corporate ownership, rights of disposal, as in the
division of the land, are exercised by consensus of all members of
the owning group. An example will clarify these points.
Prior to his death, a man willed to his six children a plot of land
of which he was sole owner. He named his eldest male child taki of
the sibling set (see first descending generation in Fig. 4). In the
second descending generation, there are seven sibling sets on the
same land formerly used by one. There are almost five times as
many people using that land, and all except young children have
equal rights to its use. The taki functions with respect not only to
his own siblings, but also to a whole matahanau. With the in-
creased pressure on the land, he must exercise a tighter control of
its use. Anyone wishing to use the land for any purpose must first
get his permission. With more people needing to use the land for
copra making, the taki must constantly rotate the portions of
land to be used and the people using them. Should the taki be
stingy about allowing his siblings and their children use of the
land, they may demand a land division. In that event, the plot will
be divided into six pieces, one piece for each sibling and his chil-
dren. In the event that one or more of the original siblings has
died, his children will receive his share in his name.
There is a
tendency now to try to avoid land divisions because of the increas-
ing population and the diminishing size of the land portions. The
trend is toward an increasing proportion of atoll land being owned
corporately, with a decrease in the proportion held by individuals.
Still the possibility of division exists as a threat to insure that the
taki will be fair in his role of steward to all the coparceners of the
In general, usufruct involves the right of occasional use of a plot
of land. The frequency with which a person exercises his use right
over a plot of land and the purposes for which he may do so vary
in an inverse proportion to the genealogical and social distance
between him and the owner of the land.
The less the distance,
the greater will be the frequency and extent of use; the greater the
distance, the less will be the frequency and extent of use. For ex-


ample, one of a pair of siblings, both of whom have inherited a
plot of land from their parents, might not hesitate to ask the other
for the use of his land to make copra or for a breadfruit tree with
which to make a canoe. His request would not be denied unless his
sibling was using the land or had reserved the breadfruit tree for
someone else. The grandchildren or great-grandchildren of each
sibling would not feel that they had the right to ask one another
for permission to use the land for the same purpose, although they
might ask for coconuts or breadfruit for a feast in which both sets
of decendants were involved.
The frequency and extent of exploitation of land over which an
individual has use rights tends to be greater if he and the owner
have a warm personal relationship. The degree to which one exer-
cises one's use right also depends, to some extent, upon the degree
to which the parent from whom the right was inherited exercised
it. The logical culmination of increasing genealogical and social dis-
tance between the owner of a plot of land and the heir of a use
right to that plot is the attenuation of the use right. Attenuation
of the use right by the holder may come about by his allowing the
right to lapse through lack of use and by not informing his chil-
dren of its existence. The owner of the land may attenuate the use
right by denying the use of the land to the holder of the usufruct;
this is tantamount to a denial of kinship ties between the two.
Adoption, as will be demonstrated, can serve to counteract the
trend toward attenuation of usufruct and its concomitant denial of
kin ties between the owners of land and those with rights to its
Ownership and use rights over a kuongo involve the rights of
regular or occasional occupation of houses and the use of trees
growing on it. As with plantations, only the owners of kuongo
have the rights of disposal. Occupation of houses is more impor-
tant than the use of trees, since there are so few trees that the
value of a kuongo as a source of food is negligible. The main house
of a kuongo is occupied by the taki of the sibling set or matahanau
which owns the kuongo, while other houses are more or less the
property of those who built them. Ownership and use rights over a
kuongo are inherited in the same way as rights over a plantation.
Since they are rarely divided in inheritance, kuongo are owned
corporately, usually by matahanau. For the same reason, one
would expect that many more people would hold hereditary use
rights to a kuongo than to a plantation at any given time. Since
people tend to live with close kinsmen or affines, however, use
rights to a kuongo tend to be attenuated more rapidly than those
to a plantation.
Adopted children may inherit taro plots, and rights of owner-
ship and use of land from both natural parents and adopters.
Through their adopters, adoptees may inherit rights to land as
individuals, as adoptive members of a sibling set, and as members
of matahanau. Such inheritance is a defining characteristic of
adoption and distinguishes it from fosterage and other relation-
ships that a child may have with kinsmen other than his parents.
The Kapinga practice both adoption and fosterage, distinguishing
them terminologically. The term e taahi means 'to adopt', while e
penepene means 'to foster' or 'to take care of. Adoption is the
transfer from the natural parent to another person of parental
responsibilities for obligations to and rights over a child. Adoption
involves the transfer of taro plots from the natural parent to the
adopter and the inheritance of taro plots and land rights of the
adopter by the adoptee. Fosterage is the temporary transfer, from
the natural parent to another person, of parental responsibility
for, obligations to, and rights over a child. Fosterage does not
involve the transference or inheritance of taro plots and land.
Children are usually adopted in infancy; in no case does adop-
tion occur after puberty. Children may be fostered at any age
before adulthood on the volition of either the natural parent or
the foster parent. A child may be left, for example, with a relative
while the parent goes to another island. A relationship of adoption
may become one of fosterage and vice versa. If, for example, a
supposed adopter dies without willing land rights or taro plots to
his supposed adoptee, the relationship which was understood to
have been one of adoption before his death has become one of
fosterage post facto. On the other hand, a person might foster a
child and later decide to adopt him. Both adoption and fosterage
occur almost always among people who consider themselves to be
consanguineal kinsmen.
People wishing to adopt a child make their intention known to
the natural parents before or immediately after the birth of the
child. The child stays with his mother until he has been weaned.
Up to this time, the adopter has been giving the natural parents
gifts of clothing and food, and taking the child to his home for
brief periods during the day. Once the child has been weaned, the
adopter takes the child to his home more or less permanently. If
the adopter is very busy and there is no one on his kuongo with
whom he can leave the child, he may take the child to the natural
parents for temporary care. Once the child has been taken to live
with the adopter, the child spends most of his time on his
adopter's kuongo especially at night. He may play with his natural
siblings and eat on his natural parent's kuongo, but he will usually
sleep with his adopter. The bulk of his support comes from his
adopter (food, clothing, school supplies, living expenses if he
should go to Ponape for schooling). His early training in fishing or
carpentry (if a boy) or in food making and mat weaving (if a girl)
will come from the adopter. The adopter also sponsors all feasts
on the adoptee's behalf.
The adopter not only assumes the parental obligations to the
child, but also acquires parental rights over the child. The most
important of these parental rights are those over the child's labor
until and, in the case of girls, after the child's marriage. When the
adoptee is a girl, the adopter also acquires rights to her potential
spouse's help in the support of the household. The adopter has full
authority to make all important decisions affecting the child, in-
cluding the approval and sometimes the choice of the spouse.
There is no attempt by the adopter to withold a child's identity
from him. Adoptees learn at a relatively early age who their nat-
ural parents and siblings are. The natural parents do not resign all
authority or all responsibility for the child. They can call on the
child to come and stay with them for short periods; they can peti-
tion the adopter when they need the child's help. They must also
be consulted in any major decisions affecting the child, and they
take a major part in life cycle ceremonies on his behalf. The natural
parents, more often than not, also provide the major portion of
the child's inheritance in land and taro patches.
Should an adoptee begin to spend a great deal of time on his
natural parents' kuongo, they are expected to force him to return
to his adopter. If the natural parents feel that their child is being
mistreated, they can demand the child's return; this can be done at
any time before the child reaches maturity. A child who is con-
stantly running home to his natural parents because of maltreat-
ment or discontent with his adopter will be allowed to remain
with his real parents; the adopter will simply cease coming to the
natural parent's kuongo to get the child.
There are three kinds of land and taro plot transfer involved in
any adoption relationship: from the natural parent to the adopter;
from the adopter to the adoptee; and from the natural parent to
the adoptee. Transfer from the natural parent to the adopter oc-
curs when the adoptee reaches the age of five or six years. At that
time the natural parent gives a taro plot or two to the adopter.
The adopter keeps it for the adoptee, giving it to him after the
adoptee's marriage. It is supposedly given for the purpose of sup-
porting the child, but the adopter actually uses it as he pleases for
as long as he pleases. The age at which the adoptee receives the
taro plot is left to the discretion of the adopter.
Transfer from the adopter to the adoptee, is one of inheritance
and occurs when the adopter makes his last will and testament.
The adopter may will land, or taro plots, or both to the adoptee,
depending upon how much land the adopter owns and uses, how
many taro plots he owns, how many natural children or other
adoptees he has, and the personal relationship between the
adopter and adoptee. For example, if the adopter has no natural
children but has three adoptees, he may apportion his land to one
and taro plots to the others. If one of the adoptees had ignored him
in his old age, the adopter may disinherit that adoptee altogether.
If the adopter has natural children, he may (as often happens)
make the adoptee a full coparcener with them; the adoptee be-
comes, to all intents and purposes, a full member of the sibling set.
In any subsequent division of the adopter's land by the sibling set
or their descendants, the adoptee or his children would be entitled
to an equal share with the others. The adoptee's position in this
case is actually rather tenuous, since there are several ways by
which the natural heirs can withhold his inheritance; they can
grant him a use right only or deny that he was ever really adopted,
i.e., assert that he was only fostered. On the other hand, the
adopter might will the adoptee a use right or no right to his land
but give him taro plots instead. The adoptee's claims to the latter
are much less tenuous.
Transfer from the natural parent to the adoptee, also a transfer
of inheritance, occurs at the time the adoptee's natural parent
makes his last will and testament. Although it sometimes happens
that a person seeking to adopt a child assures the natural parent
that he will provide the child's entire inheritance, an adoptee usu-
ally inherits land rights and taro patches from his natural parents,
often on an equal basis with his natural siblings. If he has already
inherited land and taro plots from his adopter and is one of several
natural siblings, his natural parents may will the bulk of their hold-
ings to his siblings, leaving him a more or less token inheritance.
There are, as the foregoing description indicates, many possible
variations in land transfers, but there is only one principle under-
lying them: a child is entitled to inherit land rights and taro plots
from his parents.
Adoption is a social replacement of one parent by another, but
it is rarely a complete replacement. Thus, the natural parent's obli-
gation to provide a taro plot to the adopter can be seen as a partial
fulfillment of his obligations to his child as well as to the child's
adopter. Since social replacement is not complete, the child is
therefore entitled to inherit from his natural parent and his
adopter, both of whom are consanguineally related to him and to
each other.
The gifts given by the adopter to the natural parents before the
child is taken and the transfer of the taro plot from the natural
parent to the adopter are but two formal exchanges in a long series
of less formal ones. The parties to these transactions are related by
blood and thus share mutual obligations for support. The closest
cooperation and the most stringent obligations of support
[ Sahlins' "generalized reciprocity" ( 1965: 14 7)] , exist among those
most closely related. The natural parent and adopter, regardless of
their genealogical distance, share a close relationship because they,
in effect, share a child. In any feast given by one, the other would
have to be invited to participate, whereas their genealogical rela-
tionship might not necessitate such an invitation had the adoption
not occurred. The natural parent, too, is expected to make occa-
sional gifts to the adopter on behalf of his child. Such gifts might
include fish, cloth, tinned food, etc. Gifts are not obligatory after
the child has reached adulthood.
A measure of the importance of adoption among the Kapinga is
the frequency with which it occurs. Reckoning the incidence of
adoption among the Kapinga is, at best, difficult. Ideally, the
adoptee is taken by the adopter to live with him after the child has
been weaned. Then until he reaches maturity the child stays with
the adopter; a female adoptee continues to live with the adopter
after her marriage. If this were true in all cases, then a household
census would be sufficient to locate the majority of adoptees on
the atoll. The actual situation, however, shows many variations.
Many adoptees do live with their adopters. Others have been
adopted only by declared intention but have not gone to live with
their self-proclaimed adopters. Some adoptees have returned to
their natural parents because of mistreatment by the adopter. In
other cases, adoptees are being cared for by their natural parents
or by other relatives while the adopter is elsewhere. Some
adoptees who have reached maturity have moved away from the
adopter's household to live with their spouses or with other rela-
tives, and in a few of these cases the adoption relationship has
been or is being attenuated. Thus, to list simply the number of
adopted children living in each household gives us little real indica-
tion of the incidence of adoption. The situation is further compli-
cated by the fact that the actual validation of an adoption is the
willing of land or taro plots to the adoptee prior to the death of
the adopter. There is, then, no technical way of knowing how
many of the present adoptees have "actually" been adopted.
Since we are concerned with the motives for adoption and the
goals which adoption is purported to implement, we will calculate
the incidence of adoption by considering children as "adopted"
whenever there has been a declared intention to do so. Of 720
people canvassed on the atoll and on Ponape
372, or 51.7 per-
cent, have been adopted.
Children are adopted by individuals, not by couples. The adopter
need not be married to adopt a child, nor need he be an adult;
there is at least one case of an eleven-year-old girl adopting a child,
although her parents are actually supporting the adoptee. Adop-
tion almost always takes place between cognatic kin. While the
adopter may be related to the child through both his mother and
father, he will press his claim through only one of these relation-
ships. Further, the adopter may be related to the relevant parent
in two or more ways, but he will utilize only one relationship,
usually the closest, to adopt the child. The relationships between
adopters, parents, and adoptees for the 372 cases are presented in
Table 12. While there are at least thirty different kinds of genea-
logical or other relationships between them, the relationships tend
to pattern into relatively few categories, and each different kind of
relationship is listed under the relevent category in the table. Since
our major interest is the motivations for adoption, the relation-
ships are listed according to their importance for the adoption.
For example, in at least fifteen of the cases, the adopter was re-
lated either to both parents or to one parent in two ways. In each
such case, only the relationship used by the adopter to press his
claim is listed. Again, in category 5 of the table, that in which the
relationship between the adopter and the adoptee derives from a
previous adoption, there is a genealogical relationship between the
parties in twenty-five of the twenty-nine cases, but it is less rele-
vant to the cases than the prior adoption.
Among the Kapinga, the closest cooperation and mutual sup-
port exist among siblings. This being the case, we could expect
that the importance of sibling solidarity would be reflected in
adoption practices and it is. The data presented in Table 12 indi-
TABLE 12 Relationship of Adoptees to Adopters
1. Adoptee is a child of the adopter's real or
classificatory sibling.
a. child of a sibling
b. child of a first cousin
c. child of a second cousin
d. child of a third cousin
2. Adoptee is a classificatory sibling of adopter.
a. first cousin
b. second cousin
No. Total
183 49.1
125 33.6
41 11.0
15 4.0
2 0.5
39 10.4
22 5.9
17 4.5
Table 12-continued
Relationship No. Total
3. Adoptee is a real or classificatory grandchild
of adopter. 99 26.5
a. sibling's grandchild 37 9.9
b. grandchild 43 11.6
c. first cousin's grandchild 6 1.6
d. second cousin's grandchild 2 0.5
e. great-grandchild 4 l.l
f. sibling's great-grandchild 7 1.8
4. Adoptee is an ascendant (genealogically) of adopter. 3 0.9
a. parent's first cousin 3 0.9
5. Adoptee is the child of a person whose primary
relationship to the adopter is through that of a
previous adoption.
29 8.0
a. the adopter's adoptee's child 16 4.3
b. the adopter's adopter's child 3 0.8
c. the adopter's parent's adoptee's child 4 1.1
d. the adopter's parent's adoptee 0.3
e. the adopter's grandparent's adoptee 0.3
f. the adopter's parent's adoptive sibling's child 0.3
g. the adopter's grandparent's adoptee's child 0.3
h. the adopter's fostered child's child 0.3
i. the adopter's child's adoptee's child 0.3
6. Adoptee is a child of a previous union of adopter's
own or ascendant's spouse. 3 0.9
a. the adopter's parent's spouse's grandchild 0.3
b. the adopter's grandparent's spouse's great-
grandchild 0.3
c. the adopter's spouse's grandchild 0.3
7. Genealogical relationship vague or non-existent.* 7 1.8
K. Adopter is a non-Kapinga. 8 2.1
9. Adoptee is a non-Kapinga. 0.3
TOTAL 372 100.0
See note 8.
cate immediately that the predominant relationship between
adopters and natural parents is that of real or classificatory sib-
lings. If we include the cases in category 2 and those adoptees
whose adopters are their grandparents' 'siblings', the sibling rela-
tionship is seen to be the central one in 71.5 percent of the total
cases. lf we exclude from consideration category 7
and those
adoptions involving non-Kapinga,
we can group all the relation-
ships between adopters and natural parents into two major classes:
that in which the adopter and natural parent are real or classifi-
catory siblings, and that in which they share a real or classificatory
parent-child relationship. All adoptees, then, are either classifica-
tory siblings, classificatory children, or real or classificatory grand-
children of their adopters.
Adoption among the Kapinga is a pattern of social behavior
operating within a universe of people who are, or who consider
themselves to be, related by blood.
It involves roles which are
appropriate to kinsmen, especially to siblings and to parent and
child. The individual who adopts is utilizing his kin status with
respect to the natural parent and the child. His reasons for doing
so may focus primarily on his relationship with the child or on
that with the parent, but rarely on both equally. In other words,
the adopter may activate his kin tie to the real parent for the pur-
pose of establishing a solidary relationship between himself and
the parent and, possibly, the latter's siblings as well. Here the
adoptive relationship focuses on the adopter and the natural par-
ent, and the acquisition of the child is secondary to it. As the par-
ent acquiesces to the would-be adopter's request for his child, he
gives recognition to the ties of kinship between himself and the
The motives for adoption on Kapingamarangi are numerous, and
any given case of adoption may be motivated by several considera-
tions. The long- and short-range reasons for wishing to adopt may
or may not be explicitly recognized at the time one adopts. The
basis of any given adoptive relationship also tends to change as the
adopter and adoptee grow older and their individual needs change.
The various kinds of motives, explicit and implicit, which per-
tain to any given case of adoption, may or may not be closely cor-
related with the categories of kin involved. In adoptions occurring
among close kin, i.e., primary and secondary relatives,
one or
several motives may be operating at the initiation of an adoption.
When adoption occurs among distant kin, however, the motive for
its initiation is usually to establish or restore close kin ties between
the adopter and the natural parent. Adoption or fosterage of a
child by his great-grandparent, on the other hand, involves a single
motive: that of providing a child to perform necessary labor for
the adopter.
Emory discusses five motives for adoption (1965: 158). We shall
expand these somewhat and add several others.
l. "To provide children for childless couples." As in so many
other societies, children help to solidify marital ties on Kapinga-
marangi; they are also the emblem of adult status for both parents.
There is a good deal of both male and female ego involvement
implicit in the conception, bearing, and rearing of children. A
Kapinga male is constantly under pressure to reaffirm his mascu-
linity among other men, and his capacity for having children is an
important factor in demonstrating or reaffirming his masculinity
to himself and others. Although women do not appear to have as
great a status problem among their own sex, incapacity to bear
children is, nevertheless, a problem.
Until recently, it was common practice for a childless couple to
"look for a child." The woman would ask her husband to go to
another woman and have a child by her. The married couple
would then take this child and raise it as their own. More often
than not, the wife would tell her husband which woman he should
choose, usually the wife of the man's first cousin. The issue of
such a union is called a ta ma harahara 'the child looked for'. Since
it is thought that a child receives his blood wholly from his father,
the tama harahara was preferable to adopting a child, as it gave the
couple a 'real child' tama tonu of their own and enhanced the hus-
band's manliness. With the influence of Christianity, this pattern is
no longer practiced; adoption has replaced it as a means of obtain-
ing children.
One also adopts children to provide heirs for land and taro
patches. The number of children one may adopt depends upon
rnw's preferences, relationships with close kin, and the amount of
land one has. If, for example, a person has no children and has
received land from both his father and mother, he has three ave-
nues of relationships through which children may be adopted: from
his father's sibling set or extended family; from his mother's sib-
ling set or extended family; or from his own siblings. Any child
whom he might adopt from his father's family would inherit lands
which the adopter holds from his father. A child adopted from his
mother's family would likewise inherit lands which the adopter
holds from his mother. A child adopted from one of his own sib-
lings would tend to inherit land from either or both sides of the
adopter's family. Unless the adopter's parents were close cousins
(rarely the case), an adoptee from his father's side of the family
would be considered a 'non-relative' tangata hua kee by his
mother's kin and, therefore, would not be entitled to a share in
what the latter consider to be their land. The child of an adopter's
real sibling, however, would be genealogically related to both sides
and so would have use rights to land of both.
2. "It provides parents for children who have lost a mother or
both parents." Such loss may be through death or through long-
term separation while either or both parents are on Ponape. There
seems to have been not only a higher infant mortality rate in the
past than at present, but also a higher mortality among women in
childbirth. This was the case through the early Japanese period
until it became common for women to go to Ponape to have their
children at the hospital. A trained health aide and a dispensary on
the atoll, as well as the periodic visits of a medical doctor, have
further lowered the death rate in recent years.
A common situation in which parental surrogates are provided
by adoption or fosterage occurs when a long-term separation of a
child from one or both parents is imminent. People who are going
to Ponape, no matter how long they expect to stay, are loathe to
take small children, especially infants. The families of the parents,
especially the children's grandparents, try to keep the children on
the atoll. There is a feeling that Ponape is an unhealthy place for
adults and children. The old people remember only too well the
epidemic of dysentery in 1918 which took the lives of over fifty
of the first immigrants to Ponape. Often there is an unstated fear
that one or both parents might not return to the atoll for a long
A child fostered for the purpose of keeping him on the atoll is
sometimes adopted by the relative who has been taking care of
him (another example of a fosterage relationship shifting to one of
adoption). There are both pragmatic and emotional factors in-
volved here. The man (or woman) who has assumed parental re-
sponsibility over a period of, possibly, years usually feels that he
(or she) should be rewarded with parental rights over the child's
labor. It is considered unfair for the child's natural parents to take
him when they return to the island or summon the child to
Ponape to help them. There is also the possibility that the emo-
tional bond between the child and the parent surrogate may be
strong enough in itself to motivate the adoption.
3. "It provides a father, or both parents, for children of un-
married women." The Kapinga recognize illegitimacy as such and
have terms for it. An illegitimate child is termed lama kita 'child
found' or tama hai tiri 'child made up'. The former term is the
polite one. The latter term is equivalent to the English "bastard,"
with all of its connotations, and is used to taunt or shame a person
or to provoke a fight.
When it is apparent that an unmarried girl is pregnant, her
mother will question her about the identity of the father. The
girl's statement is usually accepted as fact, but if it is known that
there may be two or more possible fathers, the infant is examined
carefully by the woman present at its birth to see which of the
possible fathers it resembles. If there is agreement among the
women as to the probable father, their report is accepted as fact.
Once the father's identity is known, he has several possible
alternatives of action. If he is not married, then his family will
want to adopt the child, since the child is 'their blood'. The initia-
tive must be taken by the father; he asks his mother or a sister to
go to the mother of the girl who bore his child and ask for it. This
is done at the feast held for the child at its mother's kuongo the
day after it is born. The man's mother takes food to the feast and
literally begs for the child. If the father does not ask one of his
family to take his child, he may be questioned by his parents
about the child. If he admits it is his, then the procedure is that
described above. If he adamantly denies that the child is his, there
will be no attempt to adopt it.
There are 63 illegitimate children (16.9 percent) in the 372 re-
corded cases. Of these 63 children, 35 (55.6 percent) have been
adopted by members of the father's family, while 28 of them
(44.4 percent) were taken by the mother's family. The specific re-
lationships involved in these adoptions are summarized in Table
If the father of an illegitimate child is married, he may ask his
wife to take the child. If the wife learns that her husband is the
father of such a child, she may question him about it and then
decide to take the child on her own initiative. In any event, her
consent is necessary before the child can be recognized by the hus-
band or his family. This is not considered adoption, however. It is
more or less analogous to the procedure appropriate to the 'child
looked for'.
It does relate to adoption in a negative sense; if the
wife refuses to take her husband's illegitimate child, his family
may not do so either. If one of the husband's close kin does adopt
the child after the wife has refused to take it, the wife will divorce
her husband. Such an adoption would be interpreted by the wife
as follows: her husband's family does not like her; and/or her hus-
band's family does not respect her; and/or her husband's family
prefers the girl who bore his child to his lawful wife. A further
problem is that of land inheritance and the potential competition
between the wife's own children and her husband's child. Legiti-
TABLE 13 Relationship of Adopters to Ad op tees in Adoption
of Illegitimate Children
No. Related on No. Related on
Relationship Father's Side Mother's Side
Parent's sibling 13 14
Parent's classificatory sibling l
Parent's adoptive sibling 1 0
Parent's adopter 4 1
Classificatory sibling l 0
Grandparent 7 5
Grandparent's real or classificatory
sibling 8 4
Great-grandparent 0 l
Great-grandparent's sibling 0 2
TOTAL 35 28
mizing the child, whether done by the wife or through adoption
by the husband's relative, makes the child eligible to inherit from
his father.
The illegitimate child who is not taken by a member of his
father's family will more often than not be adopted by a member
of his mother's family. This does not legitimize the child, but it
does, as Emory points out, provide the child with a set of parents
and at least some inheritance in addition to what he might get
from his mother.
4. "It serves to lighten the burden of parents with many chil-
dren." This statement is especially true since World War II and the
drastic reduction in infant mortality. The altruistic motive of
adopting a child who is one of many siblings in order to help the
parents is rarely sufficient in itself to stimulate an adoption; nor
are the natural parents relieved of all responsibility for the child.
People who are looking for a child to adopt, whatever their rea-
sons may be, first look for available children. The adopter will
look for a relative with many children, usually his sibling or first
cousin, feeling that here lies his best chance of getting the child he
needs. When he takes a child from a relative with many children,
he can say that he has helped his relative. The relative can claim
the same thing in turn, since he has provided that adopter with a
needed child. This kind of adoption is termed e tahi ki hakamama
sangata 'adopt to help the person'. Children are adopted for this
purpose from among a wide variety of categories of relatives; one
need not adopt from his close kin if there is a more distant relative
with many children. Such a relative is happy to have his child
adopted. For example, the wife of a Kapinga store owner living in
Ponape needed help with housework and the care of several chil-
dren. She wanted a girl about ten or eleven years of age. Her hus-
band's brother's son had thirteen children. The husband took the
ship to Kapingamarangi to talk with his nephew and came back
with the nephew's eleven-year-old daughter. He originally wanted
lo foster the girl, but as he was relatively wealthy his brother's son
subtly but firmly pressured him into adopting the child.
5. "In former days it was a means of having in the family of the
tauihara or 'non-sacerdotal class' a member of the tautonu or
'sacerdotal class'." The presence of such a child with the tauihara
family meant that the family could participate in activities
reserved for tautonu people. In feasts or ceremonies having to do
with the cult house, the tauihara family could then prepare food
to be used by the participants. The food was taken to the cult
house and presented there by the adoptee on the family's behalf.
On the days when only tautonu men were allowed to fish on the
open sea, the tauihara man could put his adoptee into his canoe
and go out to fish. If the adoptee were a girl, then she would
provide tautonu children on the kuongo after her marriage.
To Emory's list of motives may be added others which are oper-
ative in initiating adoption.
6. Child labor and balance of sex. Adoption provides parents
with children of the desired sex where these may be lacking. For-
merly this is said to have been especially the case for the adoption
of female children, who would eventually marry and bring their
husbands to their parents' or adopted parents' kuongo to provide
fish. At present there seems to be a slight preference for the adop-
tion of males over females; 220 of the 372 adoptees (or 58.9 per-
cent) are males while 151 (or 41. l percent) are females.
The balance of the sexes among a couple's children is important
in at least two ways: the kinds of labor needed and expected from
the children by the parents; and sibling group composition and
function vis-a-vis the siblings. One does not necessarily have to
adopt a child to acquire access to his labor. One might foster a
relative's child for a specified or for an indefinite period of time. It
may not always be possible to get such a child, however, especially
for an indefinite period. Once a child has reached an age at which
he can help his parents in daily chores, they are anxious to keep
him on their own kuongo.
When a child is adopted as an infant, then the adopter's even
tual rights over the child's labor may not necessarily be the major
reason for the adoption, but it is still an important motivating
factor. The earlier example of an eleven-year-old girl adopting a
small child is a case in point. The adopter asked her mother if she
could adopt the child; the mother gave her permission and went to
the mother of the adoptee to get her consent. The major burden
of support fell to the adopter's mother. When this woman was
asked why she allowed her child to adopt, her answer was: "That
little girl will grow up and help us some day." Other examples will
serve to illustrate the variations in kin relationships pertaining to
this aspect of adoption.
A male child was adopted by a woman from a male relative
when the child was an infant. This woman is married but has no
children of her own. She has another adopted child, a girl living
with her and her husband, both of whom are of late middle age,
probably in their fifties. The boy is now sixteen year of age and
has lived all his life with his adopter. He helps them by climbing
trees. going to the outer islets by himself to bring coconuts, taking
his adopted mother to the outer islets and helping her in the taro
plot, going fishing on occasion, etc. A short time ago the boy's
natural mother decided to send him to Nukuoro to help her
brother, who is married to a Nukuoro. She requested the adopter's
permission to let the boy go, promising that her husband would
bring the adopter coconuts whenever she needed them. The
adopter reluctantly gave her permission; she did not wish to pro-
voke hostility toward herself or between the parents of the boy,
but she was furious about it. The adopter felt that since she had
taken care of the boy since infancy, she should have full rights to
his help now. Furthermore, she reasoned, the boy had helped his
parents, as well as her, while the burden of his support fell solely
to her and her husband. Thus she felt that the boy's mother had
no right to deprive her of his help.
Most cases of adoption by a grandparent or grandparent's sib-
ling involve labor as a motivating factor, but this is by no means
the only factor in such adoptions. Older people, however, much
more than younger ones are aware of the child's potential help to
themselves and their eventual need for it.
All cases of adoption of a child by his great-grandparent involve
labor as the chief motivation. This sort of adoption often begins as
a fosterage arrangement. An old person who can no longer walk
any great distance will take one of his great-grandchildren, usually
a child about five or six years of age, and keep that child with him
constantly. The child does sim pie tasks such as bringing water to
the old person, taking food or utensils away when the old person
is finished with them, and cleaning the house area. This pattern
has its own term, huaiaka taku ipu 'to bring up my cup' (of sea
water to wash my hands). The child becomes the person's hands
and feet. As the child grows, he takes on greater responsibilities
for the old person's care; the latter, in turn, wills a taro plot or
two to the child.
Another mode of adoption is the result of the use of a child's
labor rather than the need for labor motivating the adoption. This
kind of adoption pertains solely to female adopters. A girl, who
has either been adopted or fostered and has been taking care of
her adopter's child for a long period of time often adopts that
child. While there are many examples of this kind of adoption, tho
case shown in Figure 5 in particular dramatically illustrates tho
A married couple had several children. After the wife died, the
husband remarried and had other children by his second wife.
Before his death, he willed the children of his second marriage to
the children of his first marriage. One such half sibling pair con
sisted of a son of the first marriage who had inherited a daughter
of the second as an "adoptee." When the girl married, her half
brother told his son to adopt her first child, a girl whom we shall
call Eve. Eve lived with her adopter, and, when he had children,
Eve took care of them, finally adopting one of them, a girl. When
Eve married, her adopted daughter stayed with her and as Eve had
children, the adopted daughter took care of them, subsequently
adopting Eve's eldest daughter. Later Eve moved to Ponape with
1 2
Figure 5. A series of adoptions
her husband and children. Her adopted daughter went to Ponape
after she had married in order to have her child at the hospital.
When the child was born, Eve immediately adopted it. This final
adoption is the result of and an expression of the very strong emo-
tional ties which grew out of long years of mutual help, depen-
dency, and intimacy between Eve and her adopted daughter.
The balance of sex in relation to the composition and function-
ing of the sibling set itself has already been discussed. It need only
be pointed out that the presence of both a male and female taki are
so important to the sibling set that if one of either sex is lacking, it
will be provided by the parents through adoption. Such is the im-
portance of this adoptee, both to the parents and to their children,
that if the parents show preferential treatment to any child, it
would be to the adoptee. For example, a man with four sons was
pessimistic about being able to have a daughter. He needed not
only to provide a female taki for his sons, but also a girl to help his
wife with the care of his children and household work. He re-
quested his half brother's fifteen-year-old daughter, stating explic-
itly that he wanted to adopt, not to foster, the girl in order to
provide a taki for his children. It should be noted that since the
girl was old enough to be of help to her own parents, a request to
foster the child might have been denied. That the request was one
of adoption and was buttressed by designating the girl taki of the
adopter's children made his claim a strong one. The half brother
had six daughters and so was in no danger of a permanent loss of
female help. Moreover, his daughter, as taki of her adoptive sib-
lings, would occupy a strong position with respect to her adopter's
land, of which she would be a coparcener and which she would
control during her lifetime, an_d his taro plots, which she would
hold for her adopter's children.
The adopter's request was granted and the adoptee lived with
lhc adopter and his spouse until after the birth of their daughter
and the adoptee's marriage, whereupon she moved back to her
father's household. The adopter subsequently had three more
daughters, one of whom was adopted by the man's adopted daugh-
ter and taken to Ponape by her and her husband. Her taki status
has been tacitly dropped, but by adopting one of her adopter's
children, she has assured herself of inheritance or use of at least
011c or two of her adopter's taro patches. It would hardly be of
advantage to the adopter's own child for the man to disinherit his
adoptee, with whom his natural child lives.
7. Decreasing social distance among kinsmen. Adoption can
serve to establish or reactivate social relationships between distant
kin. It reminds one of potential or dormant membership in an ex
tended family, residence group, or land-owning group. This kind
of motive for adoption differs from those previously discussed in
that it centers on the relationship between the adopter and the
natural parent. The child is the means through which the relation
ship is established and maintained. Adoption for this purpose is a
means of decreasing social distance between related people and is
conceived of in spatial terms-e tahi ki hakahohomai sangata
'adopting to bring the person (natural parent) closer (to the
adopter)'. Such adoption serves to initiate close kin behavior pat
terns between two people who are genealogically distant kin. It ii
not wholly restricted to distant kin; it may be used by genealogi
cally close kin whose social relationships have become distant
through lack of contact or other reasons.
The phrase e tahi ki hakahohomai sangata covers various goa11
which the adoption is purported to fulfill. Just as motives for
wishing to hakahohomai (literally, 'cause to become closer') a rela
tive tend to be highly individualized, so are the particular kinds of
genealogical relationships between the people concerned. There
are a few general patterns, however, and these will be discussed
Of the sixteen cases of a person adopting the child of hi1
adopted child, fourteen of these involve hakahohomai as the pri
mary motive for adoption. In each of these fourteen cases the
adoption, which we will call "secondary adoption," has occurred
after the original adoptee has married and is living away from the
adopter's kuongo. Male adoptees usually are living with their
wives; female adoptees are living either with their natural parents
or with their husbands' families. In each case, lack of contact
between the adopter and adoptee is responsible for the former's
initiation of secondary adoption.
The kinds of contact and mutual expectations between adopter
and adoptee after the latter has married differ depending on the
sex of the adoptee. A male adoptee is expected to live with hi1
wife after he marries, but his obligations to his adopter do not end
with his marriage. He is expected to send occasional gifts of food,
especially fish, to help with any important or heavy work that his
adopter needs done, and to visit him periodically.
As a male adoptee matures there is a tendency to avoid frequent
contact with his adopter, especially if the adopter is a male. This
tendency intensifies with the marriage of the adoptee. This is not
peculiar to the adopter-adoptee relationship but is rather a general
pattern between parents (especially fathers) and their children
(especially sons), who are te maru 'not free to talk'. The child is
'ashamed' to talk to the parent unless there is an important prob-
lem to be discussed. Conversing freely with a parent (including
joking) is considered appropriate only to young children.
Expectations of a female adoptee are somewhat different. Since
females usually live uxorilocally after marriage, the adopter has a
reasonable expectation of continued contact between himself and
his adoptee after her marriage. He expects her to live with him, to
help with the women's work on his kuongo, and to bring her hus-
band to his kuongo to live and work. These expectations are
greater if a woman has lived with her adopter since infancy. While
it is perfectly permissable for the female adoptee to live with her
natural parents or on her husband's kuongo, her obligations to her
adopter will involve more than the token visits expected of the
The adopter whose child stays away from him for a long time,
i.e., who becomes mokoamai 'farther away from him', has alter-
native patterns of behavior open to him. He may either let the
adoption relationship attenuate or he may attempt to re-establish
contact between himself and the adoptee. His choice between
these alternatives depends on his emotional involvement with
and/or his potential investment of land in the adoptee. The usual
means of re-establishing contact between the two is for the
adopter to make a secondary adoption, taking the child of his
adoptee. The rationale for this is that since the adopter has the
child of his adoptee, the latter will have to go to the kuongo of the
adopter in order to see his child.
Another general pattern involving the hakahohomai motive is
adoption between descendants of half siblings and stepsiblings. This
is especially so when a man marries a woman with children and
subsequently has children with her. The man must consider the
offspring of her previous union as his own; he is said to hakapuni
ana tama 'put his children together' (into one sibling set). This is
more than a stated ideal of kinship. Since his wife's child or ch::
dren by the previous union are older than his own children, they
will be taki of the latter with respect to any land owned corpo-
rately by them.
Whatever the immediate reasons may be for adopting the chil-
dren of half siblings, there is a very strong and verbalized interest
in demonstrating solidarity among them. In situations where land
has been divided among them or by them, as is most often the
case, the adoption serves to remind each of those involved of his
rights to the others' plots, and to minimize the possibility of land
disputes among them.
There is always a potential problem with land disposition where
half siblings are concerned, particularly the lands of the man who
is the father of one set of siblings and stepfather of the other. It is
possible that the descendants of his own children might at some
time try to exclude the descendants of his stepchildren from his
land, claiming that since the latter are not really descended from
him, they have no rights to the land. As long as a feeling of close
kinship is maintained among these descendants, this problem will
not arise. Adoption is the major means of maintaining closeness
and is probably the most practical solution to the problem. A
parent would hardly try to exclude the person who has his own
child from a piece of land to which both have a right; rather, the
parent has a personal stake in the maintenance of the adopter's
land right.
A striking example of half sibling group solidarity which has
been reinforced by adoption is the following. A woman with two
children married the former secular chief of Kapingamarangi. He
accepted these children as his own and subsequently had other
children with his wife. When she died, he married another woman
who also already had two children by a Danish ship captain. Again,
the chief accepted these children as his own and had other chil
dren with the woman. When she died, he remarried and had two
more children.
According to traditional Kapinga custom, the chief had to get
permission from the children of his first wife before he could re
marry, because these children had exclusive rights to their father's
land. He could not will his land to any children of his second mar-
riage but could only ask the children of the first to make sure that
the others got a share of his lands. The possible strains between
the half siblings, only one of which involved the inheritance of his
land, were great. To solve the problem, the chief first divided his
lands among each of the children of his first marriage (including
his wife's two children by her first marriage); then he willed each
child of his second marriage as an adoptee to one of the children
of his first. Instead of leaving the children of his second marriage
with vague use right to his lands, he created several half sibling
pairs, each in an adoption relationship (see Fig. 6), thus insuring
support and inheritance for each child. This was only the beginning
of a long series of adoptions among the chief's descendants (shown
in Fig. 7).
Social relations among the chiefs descendants have remained
dose, and are reinforced by the adoptions continually practiced
among them. It is also significant in this connection that there has
not been a single land dispute between any members of this mata-
llanau, while there have been cases of land dispute within other
matahanau whose members share a similar kind of genealogical
relationship but have not reinforced their social relationships
through adoption.
Adoption is an effect of the desire of kinsmen to maintain close
relations among one another, rather than a cause of it. Adoption is
a way of implementing the desire for closeness, but it does not by
itst!lf insure that closeness will be maintained in the face of social
strains or hostilities which might, and sometimes do, arise. An-
other case will demonstrate this point; it reveals both structural
variation of stepsibling adoption and the failure of adoption to
bring about the long-range goal which motivated it.
A man and a woman, each of whom had a daughter by a pre-
vious union, married. The man, as required by Kapinga custom,
rnnsidered his wife's child as his own. Since his stepdaughter was
okkr than his natural daughter, she was made taki of his natural
daughter and her children and all of the father's lands. The couple
had no other children. Both these girls married and had children
while their father was still alive.




I I --.
.-,-, I
... I : I I I
--- I
- I I
- - - I
\ I I
\ I I
\ 11_
\ \ , ,..
The man's natural daughter, whom we shall call Ann, went to
live with her husband and children on another islet, while the
man's wife and stepdaughter and her husband and children lived
on his kuongo with him. Seeing that there was already physical
and social distance between Ann and her stepsister and fearing
that this would increase over time, with possible tension concern-
ing the status of his lands, he told Ann's daughter to adopt a
grandson (A in Fig. 8) of his stepdaughter.
The father felt that the adoption of A would bring the children
of the women closer together. At the time, the maneuver did have
the desired effect. When Ann's husband died, she and her children
came back to her father's kuongo to live; the descendants of the
two stepsisters still live there together.
If the adoption did have a conciliating or consolidating effect at
first, that effect was short-lived. Ann's children were never happy
about living under her stepsister's stewardship, and they were less
happy when the stepsister designated her son (number 2 in Fig.
8) to succeed her as taki. Her son eventually divided the lands of
her stepfather among the children of the stepsisters leaving only
the kuongo as the undivided property of all of them. When he
died, taki 2 was succeeded by his brother's son (number 3 in Fig.
8), and taki 3 holds that position at present. There is still friction
between the two families concerning stewardship of the kuongo.
Ann's children and grandchildren feel that, since they are the
lineal descendants of the former owner, stewardship of the kuongo
Figure 8. Ann's adoption of her stepsister's grandson
should belong to them, not to the present taki, who is genealogi-
cally unrelated to them.
The animosity which exists between the two families has been
reflected in the adoption relationship between Ann's daughter and
A, her adoptee. When A was in his twenties, he went to Ponape
and stayed there several years. When he came back to Kapinga-
marangi, he married and moved from his adopter's household to
that of his wife; he has had little contact with his adopter since
then. His adoptive mother and his real mother, both of whom live
on the same kuongo, are bitter enemies. Rather than adopting one
of A's children as a means of re-establishing communication and
kin links between adopter and adoptee, the adopter has chosen to
let the relationship drop. Both parties are well aware of the pres-
ent status of the relationship, but neither has made any move to
change it.
The descendants of the two stepsisters in the above example are
not consanguineal kin, and since there was no issue from the mar-
riage of their parents, there was no intervening consanguineal link
between the two women's descendants. Thus, there are, in effect,
two unrelated families living on a kuongo, and the lineal descen-
dants of the original owner live under a taki who is not a descen-
dant of the owner. It is not surprising, then, that the adoption
failed to maintain close relations between the families. Adoption is
a pattern of kinship behavior, and the adopter and adoptee were
not consanguineal kin. Had there been issue from the marriage of
the parents of the stepsisters, they and their descendants would
have shared a common blood tie. With this common blood tie
(since half siblings are not differentiated from full siblings by the
Kapinga) all the children would have shared membership in a sib-
ling set and would have had an equal stake in maintaining the rela-
tionships and rights in the land they shared. Without the common
blood tie, there remained only the interests of two unrelated
families in a common plot of land, interests which could only con-
flict. The adoption, initiated not by the adopter but by her grand-
father, was but an empty gesture that scarcely survived the latter's
Adoption to restore close relations among kinsmen is also prac-
ticed by individuals explicitly for the purpose of maintaining the
solidarity of the matahanau in the face of factors which may
divide it. One of the major factors promoting divisiveness is com-
petition over land rights. Since rights to land are shared only by
kinsmen, to deny an individual his use rights to one's land is to
deny kinship with him and vice versa.
The denial of an indivi-
dual's use rights very often leads to dispute and litigation, especi-
ally if the land is owned corporately. Land disputes, which always
engender bitter passions among the Kapinga, may be motivated by
personal aggrandizement or may arise through the denial of rights
to kinsmen by stewards who are unaware of the genealogical rela-
tionship. Any such dispute between members of a matahanau will
almost certainly lead to fission of the matahanau.
Ignorance of genealogical relationships, especially among
younger people, may also lead to the commission of incest (as
defined by the Kapinga) and marriage between people within a
proscribed degree of kinship.
This would be detrimental to the
integrity of the matahanau for two reasons. First, marriage and
sexual relations are considered to be appropriate to people who
are non-kin or at least very distant kin; thus, people who marry
are, for social purposes, non-kin. Second, obligations to one's
affines conflict in many ways with those to one's consanguineal
Older people are generally more knowledgeable than younger
ones are about the genealogical connections which link them and
their descendants with their collateral relatives and the latter's
descendants. Probably because of this knowledge, older people
tend to be more attentive to and more concerned about the status
of relations among members of their matahanau. If an older per-
son sees, for example, that friction is arising over land or that a
marriage is being contemplated between his own descendant and
that of a classificatory sibling, neither of whom realize that they
are in fact related to him and to each other, he is in a position to
quell the friction or prevent the marriage by demonstrating to
them their relationships. The most dramatic means of accomplish-
ing this goal is by adoption. By adopting a child of his classifi-
catory sibling's descendant, the older person demonstrates not
only his own relationship to the natural parent, but his descen-
dants' relationship to the parent and the parent's siblings as well.
This single strategic adoption functions as a "message" to many
more people than the three people directly concerned. The "mes-
age" states to the adopter's and his sibling's descendants that they
are all kinsmen, i.e., one matahanau, and should therefore act
The strategem becomes even more effective when the older per-
son orders one of his children or grandchildren to make the adop-
tion, because the older person may be unable to care for the
adoptee himself, he may not live long enough to establish a strong
rapport with the adoptee and the natural parent, and, since friction
and fission are more likely to occur among his younger relatives, it
is of greater advantage to the promotion of solidarity among them
that the adopter be one of those most likely to be affected by the
potential friction. It was this kind of strategy which prompted the
adoption in the example in Figure 8, involving the stepsisters. That
the strategem was not effective in the long run owed not to the
strategem itself, but to an absence of consanguineal kinship be-
tween adopter and adoptee. The eight cases involving adoption of
the adopter's classificatory sibling's grandchild (category 3c and d
of Table 12) are examples of adoptions motivated by a desire for
solidarity within the adopter's matahanau.
Adoption can however, act to weaken the solidarity of a mata-
hanau as well as to promote it. Defined as it is by cognatic de-
scent, the membership of a given matahanau overlaps with that of
others. An individual, in fact, belongs to as many matahanau as he
has known-ancestors. He exercises his membership in any of them
insofar as he participates in activities organized by them (e.g.,
wedding feasts) and insofar as he exercises his use rights to lands
owned by his kinsmen. An individual's primary loyalties lie, how-
ever, with those relatives with whom he was co-resident as a child.
If he spent his childhood on his mother's father's kuongo, for
example, he would tend to identify himself most closely, although
not exclusively, with the latter's matahanau. Through adoption,
fosterage, divorce, or any combination of these three, members of
the same sibling set may have resided separately and may therefore
be affiliated with different matahanau. Moreover, there is often
competition between the families of the mother and father for the
child's loyalty, and adoption by a member of either side is a means
of securing it. Such competition is one reason why some adoption
relationships attenuate; for example, a mother may entice her
child to leave an adopter who is the child's father's relative to
return to the mother's household. Competition is especially keen
among the kinsmen of high status individuals, such as the chief,
the court judge, the minister, and in precontact times, members of
the sacerdotal class. Since adoption is both a way of securing the
loyalty of a child and of maintaining close relations with the
parent, the children of such persons are much sought after as
adoptees, with people usually competing to adopt them, as may be
seen in the following example.
A former chief was approached separately by a man and his
wife, each of whom was related to him in a different way and each
of whom wanted to adopt his son. The chief sought to resolve the
potential conflict in a way which is seldom utilized; he gave his
son, whom we shall call Ben, to both petitioners as a couple. This
proved to be no solution at all, since the couple continually fought
over exclusive possession of the child. Moreover, the family of
each spouse continually attempted to wean the child away from
the other. The adoptive father emerged victorious in the conflict
when his wife died. Ben stayed with his adoptive father's mother
and her husband, and when the deceased adopter's mother con-
tinued to entice Ben to stay with her, his adoptive father forbade
him to visit her. Her efforts to maintain a close relationship with
Ben did not cease. Several years later when she and Ben were re-
turning to the atoll on the same ship from Ponape, she was able to
persuade him to spend his summer vacation from school on her
kuongo. Ben married several years later and went to live on his
wife's kuongo, which is adjacent to that of his deceased adopter's
mother, so that the two have maintained more or less continual
Related to the motive of adopting to establish close social rela-
tionships among kinsmen is another which applies mainly to
people occupying the status of taki. This status involves not only
responsibility to one's kinsmen, but also a certain amount of pres-
tige as measured by performance of the role of taki, especially as
regards his reputation for fairness and generosity to the kinsmen
for whom he is responsible. Part of every taki's inheritance is the
history of the land over which he is steward and the genealogical
relationships pertinent to it. Thus, he is aware of most or all of the
people who have use rights to the estate. These people are meta-
phorically his 'children'; the extent to which they recognize him as
their 'father' by coming to him for permission to exercise their
rights affects his prestige. That certain relatives are not exercising
their rights may be interpreted by him to mean that they do not
recognize him or respect him as their taki. Such lack of recogni-
tion may threaten his image. In the face of such a threat, the taki
may seek to remind the relative in question of their mutual ties
and of his good intentions by adopting one of the relative's chil-
dren, or by having one of his children make the adoption. The
Kapinga characterize this kind of motive by the phrase ti taki ku
tahiaka tono ara 'the taki has shown (his relatives) the path to
himself. Adoption for this purpose takes place among distant kin
only, since close kin usually have little reservation about exercising
their mutual rights.
Contrasting with previously described adoption motives to es-
tablish close relations among kinsmen is that in which the adop-
tion is said to be motivated solely by the greed of the adopter. The
Kapinga characterize this kind of adoption by the expression e
taahi ki korokoromanga 'adopt to beg for things'.
In this case,
the adopter is said to adopt a child for the purpose of taking
advantage of his close relationship with the natural parent, especi-
ally when obligations owed to him by the parent affect his ma-
terial profit. The benefits which the adopter seeks to reap from
the relationship vary with the economic and social status of the
natural parent. An individual who has many taro plots, for ex-
ample, may have several children adopted by relatives who seek to
acquire the use of one or more of the taro plots. The children of
people who own retail stores or who are salaried government em-
ployees are consistently adopted by people of various degrees of
kinship who stand to profit by the liberal extension of credit at
the local store or by gifts of food or clothing. This is especially
apparent in the Ponape community where the Kapinga must de-
pend to a great extent on a money economy for their subsistence,
<md where people are usually competing to adopt the children of
those with a regular source of income. The adopters tend to be
unmarried women (with children) who Jive in households in which
there are no adult males to support them. The 1mphcations are
obvious. The children of high status individuals, such as the chief,
court judge, and minister, are usually objects of competition by
would-be adopters. The adopters may desire such benefits as favor-
itism of the chief or judge in litigation involving land as well as the
use of taro plots, the prestige of close association with the parent,
and the prestige of having a child who will eventually be an adult
of high status. In some cases, the adopter is unrelated to the per-
son he seeks to exploit but is related to the latter's spouse. He will
adopt the child from the spouse, and whatever benefits accrue
from the adoption will be effected through the spouse. Adoption
for this purpose is not restricted to distant kin, and where the per-
sonal profit motive is involved in an adoption, it is not necessarily
the only one.
8. Adoption as social replacement. Adoption can be used as a
means of social replacement whereby the adoptee replaces another
individual in a particular kinship status. The Kapinga refer to this
sort of adoption as e taahi ki pono ti rohongo sangata, 'adopt to
take (or fill) a person's place (or position)'. An adoptee might re-
place one of his adopter's natural children, one of his adopter's
siblings, a person in his adopter's land-owning group, or another
adoptee of his adopter. As implied by these various alternatives,
adoption for the purpose of social replacement encompasses sev-
eral distinct kinds of phenomena. Several factors are common to
all of them: the adoptee is usually a primary relative
of the per-
son he replaces; the adoption is a means by which the adopter
adjusts to prior or contemplated changes in his social relationships
with certain kinsmen; and the inheritance of land and taro plots is
involved, at least implicity. as a major motive for the adoption.
One generally recurring pattern is the adoption of one's de-
ceased child's child. This is considered to be an expression of one's
love for the deceased and, in part, an emotional compensation for
one's loss. The adoptee, if he is the only child of the deceased,
replaces his parent as an heir to the adopter's estate. If he is one of
several children, he might receive one or more taro plots that
would have been inherited by his parent.
There are variations of
this pattern, such as the adoption of one's deceased child's
adoptee or of one's deceased child's adoptee's child. An example
of this last variation has already been described, i.e., the case of the
chiefs son, Ben, for whose loyalty his adoptive father and mother
and their families competed. When Ben's adoptive mother died,
the latter's mother sought to keep Ben with her as a social replace-
ment for her daughter. When she was finally able to adopt his
child, she was not only able to maintain close relations with Ben
but to attain her original goal as well. The adoptee will receive
from her taro plots and possibly land rights which would have
been willed to her daughter.
The principle of social replacement can be used by the adopter
to decrease the number of people in a land-owning group or to
withhold land from a designated heir. Secondary adoption can be
used by a natural child of an adopter to ease the adoptee out of
his position as a member of his land-owning group, with rights
equal to those of the natural children of the adopter. Before look-
ing at actual cases, let us take a hypothetical case (see Fig. 9) to
make this rather complex pattern clear.
A man, X, has three natural chtldren, A, B, and C. He adopts a
child, D, and wills that the adoptee will have equal rights to his
land along with his natural children. Should the children decide to
divide the land, which X has willed to them undivided, then D
must get a share along with A, B, and C.
If A, B, C, and D are all married and have children, there will be
many people subsisting on the land. If A, the taki, does not want
to give D a share in the eventual division of land, and if he wants
to cut down the potential number of people using the land before
Figure 9. Adoption to eliminate heirs to a father's estate
it is divided, he may adopt one of D's children, E. Rather than
willing E a share in the lands which A hqlds, what A may do is to
will him a few taro plots and possibly a use right to his father's
lands. If the lands are divided later, A can claim that E has already
gotten D's share of the inheritance and that the rest of D's chil-
dren have no further claims on the land. Thus by treating E as his
adoptee and willing him taro plots, A has eliminated E and his sib-
lings as heirs to the original adopter's estate.
One case is important because it involves rights to a kuongo. A
woman with one daughter adopted a child whom we shall call
Kay, the daughter of a first cousin's son. The two girls lived on
this woman's kuongo until the adoptee married. When the adopter
died, her natural daughter became taki of the kuongo (see Fig.
The present taki (number 2 in Figs. 10 and 11) had a son
who married. His wife adopted a child from her second cousin.
This adoptee happened to be Kay's grandson also, but the adop-
tion was not made through Kay's descent line. This adoptee is now
living with the taki while the taki's son and his wife are in Ponape.
The relationship is diagrammed in Figure 11.
Figure l 0. Adoption of Kay
Figure 11. Adoption of Kay's grandson
Recently, Kay went to her adoptive sister and asked if she could
build a house on the kuongo. The taki had promised the vacant
spot to a non-relative for the purpose of building a canoe house.
When Kay complained to her relatives about the taki's refusal to
let her build on the kuongo, the taki, her adoptive sister, answered
through relatives that since her family had one of Kay's children
(the grandchild) living on the kuongo, Kay had no further right to
occupation or other use of it. Although the relationship between
Kay and her adoptive sister and that between Kay and her grand-
son were irrelevant to the adoption of the latter, Kay's adoptive
sister was able to use both relationships as a rationale for denying
Kay's use right to her kuongo. The adoptee has, according to
Kay's adoptive sister, replaced Kay.
The adoption of one's adoptee's child, i.e., "secondary adop-
tion," is another recurring pattern. Although the major motive in
most instances of secondary adoption is to restore close relations
between the adopter and the original adoptee, the secondary adop-
tion may result in the social replacement of the latter. The secon-
dary adoption thus functions as a "message" to the original adoptee
that he has not been fulfilling his obligations to his adopter and
that the adopter is waiting for him to do so. Also implicit in the
message is the possibility that if the original adoptee does not
respond with the appropriate behavior he may be replaced by his
child as heir to the adopter's estate.
Depending upon how large an estate is involved and how many
children the adopter and adoptee have, this threat may be of
greater or lesser importance to the adoptee. If the adopter has no
children but has one or two plots of land and several taro plots,
then the adoptee stands to lose a good deal by the secondary
adoption should it become one of social replacement. Since the
original adoptee would, in all likelihood, have willed his inheri-
tance from his adopter to all of his children, his replacement by
one of the children would be tantamount to the disinheritance of
himself and the rest of his children. The secondary adoptee would
then replace not only his parent but also his siblings as heir to the
adopter's estate.
If the adopter has little land and has children of his own, and
the adoptee's children will get their major inheritance from him or
his spouse, then such a secondary adoption would cause the
adoptee little concern. For instance, a man who has no children of
his own adopted his brother's daughter. This man had land and
taro plots from his adoptive father, his natural father, and his
mother. His adopted daughter lived with him until she married;
she and her husband then moved to her parents' home. The
adoptee's mother and the adopter had never gotten along well, and
she had embarrassed him more than once in public. It was the
mother who insisted that the adoptee, after her marriage, move
back to the mother's homestead. The adopter now has no one on
his kuongo to help his wife. His adopted daughter has two chil-
dren, and the adopter has adopted the elder of the two, a girl, who
lives with him on his kuongo most of the time. The adopter, when
asked whether or not he had made a decision about willing his
lands to the original adoptee and her daughter, replied, "I don't
know yet what I will do; I'm waiting to see what she (the original
adoptee) is going to do before I make any decision about dividing
my land."
Adoption then, can be used as a lever to compel the fulfillment
of obligations to the adopter. This is possible because it is the
adopter, and the people in and above his generation, who hold and
control the most important single resource on the atoll-land. This
applies not only to adopters and adoptees, but to parents and their
children as well. A parent's ability to compel the kinds of behavior
which he expects from his children depends not only on his own
personality, but also on the kinds and amounts of land that he
holds and how he holds them. This becomes increasingly apparent
as the parent and children grow older, as illustrated by the follow-
ing example.
A man's marriage had produced several children. When his wife
was dying, her first cousin came to take care of her. The cousin
already had children. When the man's wife died, he decided to
marry her first cousin. It is customary for a widower to obtain his
children's consent before he can remarry. All but one of the man's
children refused to give consent, but he married the woman in
spite of his children's objections. Then he adopted a child of the
son who had supported him previously. When he divided his lands
among his children, he willed a large piece of land and several taro
plots to his adoptee. This was land that would have gone to the
man's own children, so that their refusal to support him cost them
The implication is, then, that adoption may be used also as a
means of social control within the family, but only by those who
are in a position to use it in this manner. This latter qualification is
necessary for the reasons already outlined. While it is not a major
feature of adoption, the social control motive is clearly latent
within the system.
9. Finally, a kind of motivation only alluded to previously is a
more personal, emotional one. The Kapinga often refer to adop-
tion by saying that it is an expression of ti aroha 'love' among
kinsmen. Certainly adoption is a means of expressing and reaffirm-
ing close emotional ties between kinsmen. The case of the series of
adoptions between Eve and her adoptive daughter (Fig. 5) is
one example of this. One adoption listed in Table 12 as involving
no kinship or only a vague kinship relationship between the
adopter and adoptee affords another clear example. The adopter, a
woman, had been rescued by a young man from what she con-
sidered to be certain death during a famine on the atoll. At that
time, she told the man that when he had children, she would
adopt one; this she did several years later. Although the two had
an ancestor in common, both the natural parent and the adoptee
felt that this kin tie was irrelevant to the adoption.
Adoption may be motivated also solely by the emotional needs
of the adopter. Women who have not been able to bear children,
whether they are married or single, are particularly anxious to
adopt children; many adopt several children. This need to validate
one's feminine status and to fulfill the need for emotional inter-
action between women and 'their children' is not restricted to
childless women, however. Older women who have borne several
children tend to adopt and foster infants and very young children.
In such cases, the natural children have married and moved to
their spouses' households. The older women usually adopt or
foster their grandchildren or their siblings' grandchildren. Al-
though the reasons for adoption and fosterage by older women
may involve the need for help in their households, this is not al-
ways necessarily the case. Women past the childbearing age tend to
be quite concerned with their image as functioning and fully pro-
ductive members of their society. Related to this is a positive need
for the emotional interaction with children and the feeling of
being needed by them, which is not unfamiliar to us in our own
Viewed in ideal cultural terms, adoption and fosterage comprise a
category of behavior which forms part of a larger class of be-
haviors, the boundaries of which are defined by kinship. The rela-
tionships among the individuals involved in adoption and fosterage
are those among kinsmen, and the behavioral norms which regu-
late the relationships are those of kinship. The relationship
between the adopter and the natural parent is modeled on that
between siblings or that between parent and child. The relation-
ship between the adopter and the adoptee is modeled on that
between natural parent and natural child. While all who share a
common ancestor also share a common interest (ownership or
usufruct) in land, inheritance of land and taro plots is usually by
natural children from their natural parents. Involving as it does the
transference of parental status and role from the natural parent to
another (related) person, the adoption relationship also involves
the inheritance of land.
Viewed from the perspective of the emotional, social, and
economic situations which motivate people to adopt children,
adoption is a multipurpose social mechanism with which people
may implement their needs and designs through relationships with
their kinsmen. Adoption may be used to acquire children and to
relieve parents of the burden of their children's support. Adoption
can be used to provide children with parents and siblings. Adop-
tion may be used to acquire land or taro plots and to withhold
them from another, to acquire heirs and to disinherit heirs. It can
be utilized to establish, maintain, or sever social relationships with
one's kinsmen. Adoption can be used to facilitate adjustment to
changing social relationships with kinsmen and to promote change.
It may be a means of acquiring prestige, political support, financial
credit, labor, or emotional succor. The very variety of emotional,
social, and economic ends for which adoption is employed leads us
to the following conclusion. The cultural models of kinship rela-
tions and behavior in general and of adoption in particular afford
an ideal framework within which, and by means of which, people
can manipulate their social relationships to their immediate or
future advantage.
This study is based on field research conducted on Ponape and on Kapin-
gamarangi from July, 1965 to August, 1966. The research was supported
by the University of Oregon on a grant from the National Science Founda-
tion. My special thanks to H.G. Barnett, director of the field research, to
Kenneth Emory, and Vern Carroll, Martin Silverman and Harold Scheffler
for their help in the preparation of this manuscript.
2 Virilocal residence on a long-term basis does occur occasionally. If the
husband is the only child of a couple, or if all his siblings have moved to
other kuongo or to Ponape, the couple may reside with his family. If the
wife's kuongo is crowded and she is one of the younger of several sisters,
or if her relationships with her family are inimical or strained, she may
choose to live virilocally.
3 The distinction between 'land' and 'taro plots' is also terminologically and
conceptually based on the distinction between dry ground and wet
ground, respectively. The Kapinga term for 'taro plot' is the same as that
for 'water' wai. Considering the continual and tedious care which must be
lavished on a taro patch, it is to the advantage of the Kapinga that taro
patches be owned by individuals rather than groups, to ensure the maxium
amount of care with a minimum amount of coordination of individuals.
One has only himself to blame for the neglect of a taro plot.
4 Under some circumstances land can be acquired by gift (e.g., from a
spouse), but this is a special situation which is actually only a variation of
the usual mode of inheritance and need not concern us here.
5 With the exception of the gift of land to one's spouse (see note 4), there
are a few rare cases of land being given to non-relatives. In all of these
cases the donor either gave land which did not belong to him in the first
place or gave land to a woman on behalf of a child she bore him. Neither
case alters the principle described here, and the latter is but a variation of
6 With both categories of persons, kinship is stressed. The owner or steward
of the land to which a person holds a use right is, metaphorically, the
latter's 'father'. Those who have use rights to an owner's land are likewise
his 'children', and he is obligated to "take care of' them.
7 The Ponape population is relevant to our study for the following reasons:
(a) most of the adoptees there were adopted while living on the atoll; (b)
people in Ponape maintain their land rights on the atoll; ( c) there is a con-
tinual movement of population between Ponape and the atoll; and (d)
adoption practices are identical among Kapinga in both places.
8 In three of the seven cases, the informants had been told by a parent or
grandparent to adopt the child. They asserted that a genealogical relation-
ship existed, but could not recount it. In one case, the adoptee knew that
there was a very distant relationship, but stated it was irrelevant to the
adoption (see p. 202). In one case a child was adopted by a non-relative
who found the child alone and starving during a severe famine in 1917. In
two other cases, the adoption resulted from a curer-patient relationship. In
both cases, the adopter and ad op tee had a common remote ancestor, who
was irrelevant to the adoption. There is good evidence to support the con-
clusion that this latter pattern was introduced to the atoll by Micronesian
castaways just prior to European contact. It has been integrated into kin-
ship and adoption behavior in a somewhat traditional manner, but limita-
tion of space precludes its explication here.
9 Even with out-group adopters, the natural parent and the non-Kapinga
posit a sibling relationship. Three of the adoptees were children of
Kapinga women by Micronesian fathers whose families adopted them.
10 In the two cases in which the adoptees were genealogical ascendants of
their adopters, the adoptees were infants and were considered by the
adopters to be classificatory children.
11 There are, however, at least two cases in which a person effected the adop-
tion of his spouse's relative. The adoption was made through the spouse.
In one of the cases, the present adopter simply continued caring for the
adoptee after the adopter's death.
12 Murdock's definition of "primary" and "secondary relatives" is used here
13 There are four such cases at present, but they are not included in the 372
cases cited here, since they are not considered cases of adoption.
14 This may be due to the increasing importance of males as wage earners on
the atoll and on Ponape. One informant stated, for example: "We used to
prefer female children, but now a male child is just as good. He can get a
job in Ponape and help his parents with money."
15 It should also be pointed out in this connection that the only way an
adopter can guard his adoptee's inheritance from possible encroachment
of his natural children is by making the adoptee taki of the latter. Even
this stratagem would be effective only during the adoptee's own lifetime.
After the latter's death, the taki status would revert to the adopter's nat-
ural children and with it control of their father's land.
16 This is made quite explicit in litigation at present. The steward or owner
(usually the defendant in Kapinga cases) almost always denies any knowl-
edge of kinship with the individual who is pressing his claim in court.
17 Such ignorance does in fact exist, more so among males than among fe-
males and more among younger than among older people. This may stem
from selective memory, or from the particular interests of given age
groups. 'Incestuous' relations are not always an accident. It seems that in
precontact days individuals of the priestly class did practice first-cousin
18 The term, korokoromanga, is a derogatory epithet applied to sycophants.
19 See note 12.
20 When several children survive a parent, the eldest son or daughter, in the
usual manner, holds the parent's land and most or all of his taro plots for
the sibling set.
Martin Gary Silverman
In this chapter I will present an analysis of adoption among the
people of Rambi Island, Fiji, focusing more on the cultural than
on the social aspects. I am most interested in the area that Hooper,
in this volume, calls "ideational," concentrating upon what Geertz
refers to as "the fabric of meaning in terms of which human beings
interpret their experience and guide their action" (1957:33).
The object of a cultural analysis of Banahan adoption is not to
account for the location of particular people in particular places,
but to explicate how adoption as culturally formulated relates to
"the fabric of meaning in terms of which" Banabans give sense to,
and make sense from, social relationships. I introduce certain more
sociologically oriented materials to facilitate comparisons with
other studies.
Rambi Island was settled in 1945 by the natives of Ocean Island,
or Banaba, a phosphate island in the Gilbert and Ellice Islands
Colony, and by some other Gilbertese.
(The Banabans speak
Gilbertese, and regard themselves as related to but distinct from
the Gilbertese people.) The population of Rambi in 1964 to 1965
was almost two thousand. As a preface to the description of the
system of adoption, we must consider some aspects of 'Banahan'
as a social identity, and the meaning of the key kinship term, utu.
By legislation of the Colony of Fiji, a status of "member of the
Banahan Community" has been created. This status is interpreted
to include not only Banabans by birth and adoption, but also Gil-
bertese who migrated with the Banabans in 1945, and people cur-
rently married to Banabans. It is being a Banahan, however, that is
regarded as conferring upon a person the basic right to reside on
Rambi. Gilbertese, even if born on Banaba, are still considered
'outsiders', 'foreigners'. The link that an in-marrying Gilbertese has
with the Banahan community is defined by his Banahan spouse or
Banahan children.
The underlying logic is as follows: Rambi was purchased with
invested Banahan phosphate royalties, these royalties represent
Ocean Island lands, and only Banabans own Ocean Island lands.
Therefore, the Banabans own Rambi and the Gilbertese are their
guests. The only way for a person not born a Banahan (i.e., a
Banahan by filiation) to become one is to be adopted as a Banahan.
We now turn to the second essential background matter, the
term utu, which has several referents. When people are speaking
about adoption and it is stated that someone has 'become utu' to
another, this means that they have become kinsmen. In a construc-
tion such as 'that utu', the term labels the bilateral kindred and
the bilateral descent unit. The founder of a descent unit resided at
an ancestral hamlet on Banaba, and his descendants are referred tn
as 'members of that utu' or 'members of that hamlet' (see Maude
1932; Silverman 1966, I 967a). They are identified with the ham-
let as an area of land.
The central ideas associated with the ideal construct of utu are
diffuse solidarity, continuity, and equivalence. These ideas are
symbolized in common blood and land. Members of an utu are
referred to as being 'just the same' as one another. Because they
are 'just the same', a person can represent his cousin or his mother
at a family gathering (see Silverman l 967b). This sameness of con-
sanguines is conceptualized in terms of common substance (blood)
and/or common interest in land (in the case of Banabans, land on
Ocean Island). By "common interest" I mean the common hamlet
identification referred to above and/or the fact that a person can
envision circumstances in which he might inherit the land cur-
rently belonging to even a remote kinsman, when that kinsman's
line dies out. Some people did say, when asked by the ethnog-
rapher, that one of the reasons for maintaining certain kinship
relations is that people have an eye to the future, to the possible
inheritance of land. But since the nearest kin have prior rights, the
chance that a person could come into possession of a distant kins-
man's land is very slim indeed, and the amount of land might be
very small.
The term kaititi is used as an expression of genealogical close-
ness. When X and Y are kaititi, if X's line dies out, Y will inherit
X's land, and vice versa. People sometimes say: "They are close,
they are kaititi," or simply, "They are kaititi."
The point here is that kinship is symbolized in terms of both
blood and rights in land. Land functions symbolically in some of
the same ways as blood; an assertion of common interest in Ocean
Island land is similar to an assertion of common blood. Both land
and blood are "natural," essential, and divisible,
and it is the
passing of title to at least one Ocean Island land from adopter to
adoptee with the approval of the 'close utu' that finalizes an adop-
Although one could imagine various symbolic devices for alter-
ing the blood (as in blood brotherhood), Banabans do not make
use of them. But the parent can transfer different proportions of
his land to each of his children. One of the reasons for this is the
children's relative fulfillment of the filial role. Thus, whereas
blood is one way of conceptually grounding the utu relationship,
land is different in that it both grounds and is a consequence of it;
its parceling out expresses the quality of relationships. As symbols,
land and blood both function as synecdoches, although in different
Land expresses the double meaning of utu as a social category
and as a behavioral category (Schneider 1968). Utu itself is used
verbally: utuna means 'to treat as kinsman'. There are situations
where two people who are not utu do in fact treat one another as
kin; there are also situations where two people who are utu do not
so treat one another.
In giving land to another, one gives him part of one's social per-
sonality, and self-perpetuation is important to Banabans. Cases
will be described later where someone is adopted 'to replace'
someone who has died. The notion of replacement implies that the
adoptee receive some kind of title to at least part of the deceased's
share of land.
The relationship between land and social personality is partially
expressed in the term mwi 'consequence', 'remains', 'trace'. The
lands inherited by a person from his father contrast with those
inherited from his mother as 'the male mwi' and 'the female mwi',
respectively. "There is also his mwi in the Gilberts" is a common
phrase meaning that a Gilbertese adopted by a Banaban still main-
tains his property rights and relationships in the Gilberts. When
some Banabans wanted to make it clear to the ethnographer that
certain children (not their own) were just 'staying with them' and
were not adopted, they would say, of a particular child: "His mwi
is with his parents."
To indicate that a person belongs by adoption rather than blood
to the utu of some other people, one might point out: "He has
some land of theirs." Banabans speak generally of land as both
'the property of the person' and 'the property of the utu', but
they are not referring to two different kinds of tenure. Title to
specific pieces of Ocean Island land is a feature of an individual's
distinct social personality and of his social personality as a mem-
ber of various utu.
The population to which these figures apply is those residents of
Rambi present during the 1965 census conducted by the Banahan
Trust Fund Board under my general supervision. The figures in
Table 14 may be revised slightly after the computer analysis of the
data now in progress. The statements made in the body of the
paper are derived from data which include people who are not
currently resident on Rambi. Data about the adoptions of persons
now deceased are included also. For example, there were three
people adopted as 'parent', who died before the census. The divi-
sion into under twenty-one and over twenty-one age-groups is
made to isolate roughly the first "Ram bi generation." Detailed
data on age and other matters are available, but their meaningful
interpretation would require a mass of material as yet incom-
TABLE14 Population of Ram bi
Under 21 Over 21 Total
Banabans by birth 1101 482 1583
Banabans by adoption 4 37 41
Not sure 5 6
Others 57 234 291
TOTAL 1163 758 1921
pletely analyzed and rather tangential to the concerns of this chap-
The meaningful interpretation of the figures presented would
also require a discussion of certain aspects of the shift from Ban-
aba to Rambi, which will be dealt with elsewhere. Some of the
directions that this discussion could take will be obvious from the
figures themselves. The question of possible bias in the statistics
will also be taken up elsewhere; I suspect that the figure of one for
grandchild adoptions of Banabans now over twenty-one is a result
of inadequacies in data collection.
Within the population as defined above, the cases of adoption
which have been counted are limited to all known registered and
unregistered adoptions by Banabans, and by non-Banabans who
TABLE 15 Banahan Adoptions (Registered and Unregistered)
Banabans Non-Banabans
Adoption Relationship Under 21 Over 21 Under 21 Over 21 Total
Child 41 28 4 23 96
Grandchild 10 0 4 15
Child or grandchild* 0 I 0 2
Parent 0 0 0 4 4
Sibling 0 3 10 14
TOTAL 51 33 5 42 131
tnformation incomplete.
lived on Ocean Island or Rambi (see Table 15). This excludes a
few cases of Banabans adopted by Gilbertese in the Gilberts, and
adoptions of Gilbertese by Gilbertese in the Gilberts.
Where a Banahan or a Banahan sibling set adopted a non-
Banaban as sibling, I find it most useful for sociological purposes
to make a single entry, that for the non-Banahan. The rates of
adoption are summarized below.
Total population I 921
Number ofadoptees 131
Percentage of population adopted 6.8
Total households 291
Households containing at least one adoptee I 01
Percentage of households containing at least one adoptee 34. 7
'Adopted' in Gilbertese is tabekaki, which also means 'lifted up',
'raised up'. This is quite appropriate since an adopted person is at
least partially 'lifted up' from one set of family relationships into
another. The Banabans distinguish forms of adoption by the kin-
ship relationship of the adoptee to the adopter. The forms are as
follows: 'child adoption', 'grandchild adoption', 'sibling of same
sex adoption', 'sibling of opposite sex adoption', 'father adoption',
and 'mother adoption'. The latter two types are also referred to as
'the replacement'.
'Child adoption' will be discussed first, and
then the other forms compared to it.
Two forms of te natinati 'child adoption' are recognized: te
natinati proper and te nati-ni-kauatabo 'child of two places' or
'child of two ends'.
The latter name refers to the equality in the
relationships of the adoptee to the family of his adopter and to his
natal family. A person adopted in this way can receive substantial
amounts of Ocean Island land from both families.
In 'Banahan custom' te katei-ni-Banaba, whether or not a per-
son properly maintains a kinship relationship with his natal family
after 'adoption as child' depends on whether or not he has re-
ceived Ocean Island land from them; it need not be more than a
token amount. One or both of his real parents may give him land,
with different consequences, as we will see. In te natinati proper,
this land from one's real parent or parents is called 'the accom-
paniment' and 'the wanting-to-return'. As the receipt of land from
an adoptive utu symbolizes incorporation in it, the receipt of land
from a natal utu symbolizes a continuing link with it.
A person can be 'adopted as child' (when an infant or adult) by
either a married or unmarried adult, or by a married couple, gener-
ally upon the initiative of the potential adopter. One of the rea-
sons why only one of a married couple participates in the adop-
tion is the objection of the near kinsmen of the non-adopting
spouse when the latter is childless. Gossip usually ascribes the
refusal of near kinsmen to greed for the land of their relative: if
the latter dies childless, his lands will be divided up among them.
Although a man may adopt someone as child while his spouse does
not (and the child thus becomes a member of his kindred and the
descent units of which he is a member), the woman and her hus-
band's child properly behave in consanguineal roles, as is also nor-
mative for steprelatives.
The situation where only one of a married couple adopts a per-
son as child can be extended, as it were, one generation backward.
An individual can adopt someone to his 'mother's (or father's) side
only'. The land that the adoptee receives is land inherited by the
adopter from that parent, and he becomes a cognatic kinsman of
the adopter's kinsmen 'on that side'.
It is said of a child adopted away that 'he has gone'. His adopter
is responsible for him, and he is responsible to his adopter. If he
has received land from his natural parents, the phrase 'but he has
his accompaniment' is added to 'he has gone'. The child is un-
equivocally a member of his adopter's utu, but as long as an adop-
tion (of whatever form) from outside the utu is remembered,
other members of the adopting utu point out that the adoptee 'has
our land'. This means that the adoptee is a kinsman, while at the
same time asserting that the other utu members would have more
land if the adoption had not taken place. The adopted person is
responsible to these other members of his adoptive utu for his
social position. It is by their grace that he occupies the social posi-
tion that he occupies.
It is recognized that the person adopted away may still love
members of his real family and may want to express these feelings.
As a Banahan remarked, "the blood loves its own." The adopters,
if they want to, can permit the maintenance of some relationships.
If, for example, an adopted child's real kinsman is getting married,
the former can present a gift known not as 'an utu contribution',
as the ordinary gifts of kinsmen are known, but as 'the wanting-to-
eat'. Members of the utu come to the wedding 'not just to eat at
the feast, but to help with the work'. Their contributions go to-
ward defraying the expenses of the feast (which are high because
of the utu members very presence). The term 'the wanting-to-eat'
appears to suggest that the giver is directly exchanging money for
food, which is contrary to the norms of kinship. People point out
however that the returning adopted child does not present 'the
wanting-to-eat' merely because he wants to eat at the feast; he
comes also to help. The phrase used to describe the gift denotes
the expression of a solidarity that is on the one hand broken, but
on the other, unbreakable.
Some Banabans indicate that in the 'truest' (i.e., most ancient,
uncontaminated) Banahan custom, the child receives no land at all
from his real parents; the emphasis is on severing relations with
them. Today this applies particularly to adopted Gilbertese. H. C.
and H. E. Maude wrote:
On Banaba (Ocean Island) adoption from outside the kindred and if possible
from outside the island was actually preferred, as it was considered that the
son of a fellow-islander would tend, after his adopter's death, to carry on the
name and fame of his true parents, whereas a total stranger, removed from his
home-island, would rely for his local prestige upon the name of his adopter
and thus perpetuate his memory. Such an adopted child could inherit all the
adopter's land, even to the exclusion of the adopter's real children. This
forms an instance of the vast difference, also noted in other phases of social
organization, between the culture of the Banabans and that of the Gilbertese.
Gilbertese adults adopted by Banabans point out that the latter
dislike the Gilbertese' receiving a share of their real parents' land;
they should be 'finished with their Gilbertese side', it is said. These
are remarks underscoring the way the Banabans guard the main-
tenance of their identity and their rights to Ocean Island land.
These rights are a focus of controversy with the British administra-
tion. On the same theme, it is said that the wise Gilbertese
adopted as a Banahan will marry a Banahan rather than another
The elected Island Council has forbidden the adoption of non-
Banabans (although not retroactively), but some recent adoptions
of Gilbertese have occurred. The prohibition is explained as being
for the good of both Banabans and Gilbertese. There are Banabans
who later cast out their adopted Gilbertese relatives, 'and then
where will they go'? In addition, there are adopted Gilbertese who
later ignore their adoptive Banahan kin. Some Banabans are also
afraid that Ocean Island land might get into the hands of other
Gilbertese. An adopted person, as any other person, should not
dispose of his land without the consent of his close kindred, in this
case the close Banahan kindred. Yet an adopted Gilbertese, as the
Banabans see it, might be able to dispose of his newly acquired
Ocean Island land in an improper way, by some clever means giv-
ing it to his Gilbertese relatives.
With the possibility in mind that an adopted child later may pay
insufficient attention to his adopters, it is considered wise to give
him initially only a little land. At the same time, the adopted child
is often referred to as becoming 'the first-born' vis-a-vis the natural
children of the adopter, perhaps ultimately receiving more land
than they do (note the Maudes' remark quoted above). I once
asked a young man whether he considered as 'brother' a man
adopted as brother by his sister. He said that he did, and indicated
that with adopted kinsmen one takes even more special care than
one does with real kinsmen to fulfill the expectations of the role.
This is perhaps one of the paradoxes of adoption. Relationships by
blood are natural and durable. Relationships by adoption may not
be natural and are visualized as more terminable, but a person is
under a special obligation since the arrangement was entered into
voluntarily (at least by the adults involved). You cannot say of a
person, "If she didn't want to care for him why did she bear
him?" in the same way that you can say, "If she didn't want to
care for him, why did she adopt him?" Adoptive relationships are
the outcome of choice, not chance.
There is thus a sense in which the parties to an adoption are "on
trial," and twice, to my knowledge, the matter of canceling an
adoption has recently come before the Rambi Island Council.
(Other cancellations have occurred informally; the land transfers
had not yet been officially made.) Both cases involved the adop-
tion of a non-Banahan by a Banahan. In one case, a woman com-
plained that her adopted Gilbertese daughter (and the latter's
husband) went away to the Gilberts, resumed relations with the
adoptee's natal family, and ignored instructions to return to
Rambi. In the other case, a young man asked for the cancellation
of his own adoption, and the adopters agreed. He said privately
that they had turned him out of their house for no reason.
The simplest adoptions to effectuate are those within the utu.
Among close kinsmen, especially siblings, it is difficult to refuse
the request for the adoption of one's child. They are, after all,
'just the same' as oneself. The initial situation of adoption within
the utu is a dramatic statement of the strength and diffuseness of
the kinship tie. Yet the central idea, or the outcome, need not be a
strengthening of the solidarity of the kindred. As in many aspects
of Banahan kinship, and Banahan social structure as a whole, one
finds the concomitant expression of two themes: the unity of the
social group, and the marking off, by the individual, of a domain
distinctive to himself. For example, in adopting my sibling's child,
I may indicate my preference that the child inherit only through
me and not through my sibling. I am the one who 'governs' the
child; he is my child, and I am emphasizing this point. At the same
time, however, my sibling and I are 'just the same', and it is be-
cause of this fact that I could, in the first place, adopt his child
with such ease.
This kind of ambiguity was manifested in the episode illustrated
cQ == D E ===6F

_ ....
Figure 12. Characters in an adoptive joke
in Figure 12. G, a young boy, was adopted by E and F of Buakoni-
kai village. He was playing on the Buakonikai wharf when B, who
had been visiting the village, was about to depart in a small boat to
return to her own village of Uma, where C and D lived. B is a
woman with an ironic sense of humor, and as the boat was prepar-
ing to pull away she called out to G: "Don't you want to come
with me to Uma? Isn't D your father and C your mother?" The
"answer" is both yes and no. C and D are 'father' and 'mother' to
G as biological parents, and also as the brother and brother's wife
of his adoptive mother; but E and F are the ones who have
adopted him. D and E are 'just the same', but D has relinquished
the parental role to E.
We will now consider acts of adoption which are interpreted by
the people as involving, in a more explicit manner, the strengthen-
ing or maintenance of some relationship already in existence.
These cases can be treated under four headings: the solidarity of
the spouse-pair; the assertion of claims over children in special cir-
cumstances; the strengthening of a previous adoptive relationship;
the realization of an adoption planned but interfered with.
1. The solidarity of the spouse-pair is relevant where both
spouses adopt a child already related to the husband, and another
child already related to the wife. The situation is thus equalized.
Each spouse cares for and gives property to a child initially related
to himself, and also to a child initially related to his spouse. The
cases in which the couple adopts a child related to both spouses
are similar in this respect.
The question of spouse-pair unity is also significant in under-
standing a kind of adoption regarded on Rambi as new: the adop-
tion as child of a deceased child's spouse. In one of two such cases,
the adopted daughter of a childless couple told them before her
death not to grieve because there was someone, her (Gilbertese)
husband, who would replace her.
In the other case, a deceased Banahan woman's non-Banaban
husband was adopted after his wife died, "to look after the chil-
dren." He replaced her. The man later was remarried, to a Gilbert-
cse, and he is reported to be giving Ocean Island land to the chil-
dren of his second marriage. Some of his first wife's relatives
commented that this was improper. According to them, the
purpose of the adoption was to ensure his position as guardian of
the children from the first marriage, and he was made custodian of
their lands for this reason. The adoption, in this view, was not an
unequivocal one, and he was transgressing its terms.11
2. The assertion of claims over children was involved in the
second spouse-of-deceased-child adoption discussed above. One
cannot be sure that children will be well-cared for by the surviving
spouse; if he is a foreigner he might take them away from the is-
land. Where the utu of a deceased or divorced parent is fearful for
the fate of the children, an utu member may adopt them. From
the point of view of the position of children, the adoption of a
deceased kinsman's offspring can serve some of the same functions
as leviritic and sororitic marriages.
3. A person's identification with the utu into which he has
been adopted can be dramatized through an adoption by him of a
member of that utu. In one case, a man adopted the son of one of
his adopted siblings "to strengthen the adoption with that utu."
Another case is as follows: A Gilbertese boy, X, was adopted by
a childless Banahan, Y. X went to Y and said that he wanted a
sister from Y's utu. Y indicated that he had thought about such an
adoption himself. The boy's request demonstrated his identifica-
tion with the utu of his adopter. The adopter showed by his own
action that although he had adopted from outside the utu, he was
still concerned with it, and wanted some of his land to remain in
the hands of its blood members.
4. In some cases an adoption is planned with a certain person,
but does not eventuate. Another member of the person's family is
then adopted to realize the adoption, by replacing his kinsman
who was not adopted. The simplest instances are those in which
the expressed motivation is the rewarding of a kindness. For ex-
ample, a man is particularly well-cared for by a certain family on
another island and wants to adopt one of their children. The pros-
pective adoptee dies and is replaced by someone else from that
The strengthening or maintenance of a relationship can thus be
quite explicitly recognized by the people as motivations for child
adoptions. There are other explanations which are more general,
but it should be noted that in many cases the participants ascribe
their acts to a number of motivations. A childless couple may state
that they simply wanted a child; Banabans delight in young chil-
dren, and parenthood is considered a normal feature of adult
A childless couple may also explicitly refer to the need
for a child to inherit their land and to help them.
A childless person who adopts a kinsman's child, and thus has
a specific person committed to support him in the future, will be
less likely to call upon his kinsmen for major economic support
when he is old. Thus it can be argued that in the long run child
adoption by the childless functions to maintain equable relations
among kinsmen by minimizing the likelihood of situations arising
where the more onerous obligations entailed by the kinship rela-
tionship would be put to a test.
Where a couple's children are all of one sex, it is considered
eminently reasonable to adopt a child of the other sex. There is a
general idea of the need for both brothers and sisters, performing
complementary roles, to complete a sibling set. The question of
sex-role definitions is an important one that cannot be pursued
here, but I will note anecdotally that when people first learned
that I had no sisters, they asked in surprise why my family had not
adopted one for me.
People also explain some adoptions as arising out of the desire
for a particular child or to help the parents of a particular child.
The adopter may have helped the mother during the child's in-
fancy and come to love the child, to feel about it as one feels to-
ward one's own child, and thus want it to become his child. In one
case, a little girl became attached to a couple (the man was her
father's brother, and the woman her mother's sister) when they
were visiting her parents; she wanted to leave with them and was
allowed to do so; an adoption resulted.
My data on the expressed attitudes of the natural parents are
not as complete as those on the attitudes of the adopting parents.
When one has many children, there is an advantage seen in farming
some of them out through adoption; one can still intend to give
them land, and thus not 'send them away' entirely. When it is a
childless couple that wants to adopt one's child, or a couple with
few children but a substantial amount of Ocean Island land, the
child will be seen as fortunate because of its prospective inheri-
tance and because of the care that will be lavished upon it. In cases
of intra-utu adoptions there is also the element of acting as a good
kinsman, especially within the actual sibling group. I asked one
man, X, whether his younger brother, Y, could have refused X's
request to adopt Y's child. A man often given to hyperbole, X
asserted that first of all a younger sibling must obey his elder, and
second, their mother had been in favor of the idea, and "when the
boss [i.e., their mother] speaks, the matter is settled."
Let us by taking a hypothetical case inquire a little more closely
into some of the consequences involved in child adoption. Since in
most child adoptions by Banabans of Banabans there is already a
consanguineal or affinal link between the adopter and a parent of
the adoptee, we can take the hypothetical genealogical situation as
representative. To simplify matters, assume that all the people rep-
resented in Figure 13 are Banabans. M is adopted by F and G. He
receives no land from K and L; it was agreed that he would 'go
away' from them. Although he has 'gone' from K, K is still his utu
'kinsman' both in the genealogical and behavioral senses through
the new relationship since she is G's 'sister'. M and his relatives
through L and E are defined as no longer having binding claims
upon one another, although the incest prohibition remains.
loses his descent unit statuses which were transmitted through L
and E. He acquires through B and F new statuses and new kins-
In certain cases the adopted child might have to be taught
where his primary obligations reside. I was present during one inci
dent when M walked into the house of his real parents, K and L. L
said to M, angrily, "This is not your house. Get out!" The child,
about six years old, seemed uncomprehending and reluctant, but
finally left. K and G in this example were 'sisters', and F and L
were 'brothers'. If M had been a 'true' child of F and G, this would
have been a gross violation of the norms of kinship on L's part.
There are many occasions on which both families are brought into
close interaction and go into and out of one another's houses. This
was a lesson, at an age when lessons have to be learned.
Let us change the nature of the hypothetical case, to bring up
another important question: the contexts which call for a state-
11--<J :I:

:::C I
CJ :
-- I I
11 '

ment as to whether a child is or is not adopted. Here, K and L
have a large number of children, and K has just given birth to an-
other. G comes to take the new child, M, to relieve K's burden; K
agrees. M continues living in the household of F and G.
In many practical respects, others do not find it necessary to
know whether an adoption has taken place in such a case. G's rela-
tives through B, and F's relatives, do not consider M as their kins-
man, but in day-to-day interaction they will not treat him indif-
ferently because they know he is being cared for by their kinsman.
M, as he gets older, will extend special courtesies to those relatives
since, by doing so, he is paying proper respect to the people who
are rearing him. Even if he had been living with K and L and there
was no question of an adoption, he would call F 'father' and B
'grandmother', and be courteous and respectful toward them.
As M grows toward maturity, the possibilities multiply of situa-
tions arising which call for a more explicit statement of his status,
and by this time understandings about that status will probably
have been reached. At the time of his marriage the question will
become important. The complex set of rituals associated with mar-
riage cannot be discussed here, but the problems that could arise
include who will sponsor the marriage and which people will be
enjoined to attend as kinsmen. There are similar potential prob-
lems with respect to M's participation in the ceremonies held by
the utu of B, F, E, and L. If the land transfers have already been
made or the intentions already stated the situation is clarified. If
not, people can be in a quandary.
The granting of title to Rambi land also calls for some state-
ment of the 'placing' of the child. The Rambi lands subdivision, in
progress during my period of fieldwork, is another matter that will
be treated in detail elsewhere. In brief, it was decided that each
member of the community should receive individual title to one
acre of garden land. For reasons of agricultural efficiency, it was
considered desirable to arrange the holdings so that the plots of a
number of people who could be expected to cooperate on a lon1
term basis would be contiguous. The members of each village were
asked to form themselves into utu for land allocation purposes. Of
Banahan children whose adoptions were unregistered (i.e., merely
the intention had been expressed), 80 percent were counted as
part of their adoptive rather than their natal group.
The forms of adoption differ in their 'strength'.
Adoption as
child has the greatest finality; the other varieties are recognized as
more easily breakable and more easily interfered with after the
death of the adopter by a kinsman greedy for his land. For ex-
ample, a childless woman who lived on Ocean Island before the
resettlement wanted to adopt as grandchild her Gilbertese hus-
band's adopted daughter's son, who was 'grandchild' to her
through her husband. But she reconsidered because "grandchild
adoption is not strong, it might be washed out"; so she adopted
him as child.
The nature of child adoption as the securest form relates to two
things we have already discussed. First, child adoption is the only
form in which the severance of relations with original kinsmen is
stressed. Second, it is only the person adopted as child who is gen-
erally expected to receive a share of land at least equal to the share
he would have received had he been a member of the utu by birth.
Adopted grandchildren are almost always classificatory grand-
children to the adopter. These adoptions are explained mainly in
terms of simply wanting the child or seeing the child as a replace-
ment for a person whose adoption was previously contemplated.
The adopted grandchild can be in an especially favored position,
since he may receive more land than he would otherwise be heir
to-land from both his parents and from his adoptive grandparent
or grandparents.
The child raised by a 'grandparent' of some kind is often ex-
pected to be less accustomed to hard work than if he were raised
by his parents. Yet he should want to help his 'grandparent'. These
points are dramatized in a Gilbertese song entitled, "Lament of an
Adopted Grandchild":
All I do is wake up to work;
I get the firewood, I light the fire, I boil the rice.
I boil my father's and mother's food.
They have eaten and I, I myself eat.
(verse repeated)
All my grandparents say of me:
Sulky; wants to eat; an habitual thief!
(verse repeated)
Exactly who the "father" and "mother" are was not specified,
but they were some relatives of the adoptive grandparents living in
the household. The irony of the song lies in the presentation of
the adopted grandchild's life as one of drudgery, with only unkind
words from his adopters. But in the singing and drinking group
discussing this song at my request, one Gilbertese man, who often
gained great satisfaction from offering minority interpretations on
various matters, said that his comrades were perverting the true
meaning. The child, he pointed out, is in fact proclaiming his joy
in serving the members of his adoptive family, and the hard words
of his grandparents were spoken in jest, and actually showed love.
Sibling and parental adoptions relate Banabans and Gilbertese
primarily. The ideal-typical parental adoption is of a Gilbertese
stepparent by his Banaban stepchildren. The people adopted as
parents are Gilbertese who do not have Banahan children, and thus
lack one of the links which would make their position in the com-
munity more secure.
A stepparent is supposed to treat his stepchildren well, but at
the same time trouble is expected. The stepparent who is in fact
very kind to his spouse's children can as a reward be adopted by
them. He 'replaces' the deceased true parent and is given a share of
the latter's land.
The relationship that has grown between step-
parent and stepchild, one of parental and filial affection, is the
basis for conferring upon the stepparent utu status, and thus Bana-
han citizenship.
Adopting a parent's second spouse (when the parent is still
alive) is also indirectly a statement of solidarity with the surviving
parent, who is liable to pay more attention to his new spouse than
to his children; adoption states that affection reigns where it really
ought to.
In one case of parental adoption, the Banahan wife, A, of a
Gilbertese, B, asked him, before she died, to continue caring for
her adopted child, C. B is known for his kindness toward C, who
later registered him in a father adoption as having replaced A. This
is an instance of a man succeeding to the position of a woman,
rather like the cases of a couple adopting the spouse of their de-
ceased child. In both situations the care of children can be a major
The fact that the typical parental adoption is of the Gilbertese
stepparent presents us with an interesting analytic problem. Since
there is an explicitly affective element grounding the adoption,
why not adopt Banahan stepparents, to whom one could also be-
come emotionally attached?
The answer lies, I suggest, in the dependency of adoptee on
adopter mentioned earlier. When one person gives land to another,
the recipient is under an obligation to the giver. The general point,
which we have encountered before, is exemplified in the following
case, where the recipient did not manifest awareness of the obliga-
tion. A man, X, commented privately that a certain eider's an-
cestor was adopted as a Banahan and received land from X's
family, indicating that the elder was thereby in a special position
of obligation to that family. The elder did not, however, extend
special courtesies to the family. X commented that this was not the
kind of thing that one should advertise (it is actually the kind of
remark that might issue forth in an angry outburst), and if the
elder chose to forget it, "it was all right" with him. X was expres-
sing his generosity and the principle that people do not like it to
be known that they are under such obligations to others. Beyond
the domains of kinship, age- and sex-roles, the people have a
strongly egalitarian ethic, and even within those domains, obliga-
tion is often felt as oppressive. When an adopted child or grand-
child receives land from his adopter, the ultimate direction of
obligation complements the direction of authority: the elder gen-
eration is superior, the junior generation inferior. In the case of
adopted parents, however, there is a reversal; the adopted parent
becomes obligated to his adopting children. For this obligation,
the Gilbertese receives something valuable in return: Banahan
status. For the Banahan it would make little sense.
Parental adoptions differ from the other forms in the nature of
the property arrangements that are made.
When Ocean Island
land is given to an adopted parent, it is assumed that the land will
be 'returned to its true place'-to the adopting children-after the
adoptee's death. The latter is 'living' on it or 'eating' from it until
he dies,
and from the point of view of the larger utu his position
is marginal. He is not implicated in the continuity of the utu, and
continuity is one of the ideas central to utu. In parental adoptions,
it is the idea of mutual affection and solidarity that is stressed.
If one asks people other than the adopters whether the adoptee
is a member of the utu, and whether he can participate in descent
unit affairs, their answers are often of the "Yes, but ... "type. It is
conceivable for such a person to participate, but he is not ex-
pected to do so. In a limited way, the adopted parent 'replaces', or
assumes the kinship identity of, a deceased parent; but the replace-
ment is limited in the sense that members of the utu other than
the actual adopters recognize him as kinsman only as a matter of
courtesy. When the adoptee dies, it is as if he had never lived at all.
His social personality is not perpetuated through the children who
adopted him. (This contrasts, of course, with the case of a person
adopting, say, a stepchild.) 1
I mentioned the assumption of the return of the land to the
adopting children. One of the most ambiguous set of responses to
any question I raised about adoption concerned this matter. If the
adopted parent married again and had children, could he give the
lands to those children? It was a hypothetical question since there
was no concrete case which could be cited. Some said that the
land really belonged to the adopting children and that the
adopted parent could not dispose of it at will. Others said that it
was his land, and if he wanted to give it to his new children he
could. These different interpretations reflect the "double mean-
ing" of land and, indirectly, the double meaning of utu. On the
one hand, land 'belongs to the person'; on the other, the land once
occupied by a common ancestor is 'our land' (i.e., land in which
there is a common interest). The primary relationship is between
adopter and adoptee, but where land has passed, this relationship
is set within the larger context of the utu.
In sibling adoption the idea of equality, which is the essential
feature of siblingship (as contrasted with other consanguineal rela-
tionships), is shown in the emphasis on a mutual exchange of land.
The relationship created does not end with death, as in parental
adoption. The terms for sibling adoption (te i-taritari 'sibling of
same sex adoption' and te i-mamane 'sibling of opposite sex adop-
tion') also mean 'brotherhood', and are frequently so used in
oratory. If one states that a particular person is i-taritari or
i-mamane with another, however, there are two specific meanings.
The first is that they are in the relationship of te bo 'the meeting',
which is strong and enduring friendship thought to be like utu. Te
i-taritari and i-mamane are adoptions when land has been trans-
ferred. As one would expect, in 'the meeting' alone a non-Banahan
does not become a Banahan; in a sibling adoption he does. He
owns Ocean Island land, and in local theory Banahan status and
the ownership of Ocean Island land are implicit in each other.
As in any kind of adoption, if a person wants to adopt another
as sibling the members of the 'near kindred' must approve. The
adopter's real siblings will consider him their sibling, but it is the
adopter's share of land in which he shares. He becomes part of the
sibling set, but it is the adopter who brought him into it. A sibling
set may also want to adopt someone to replace a deceased sibling,
in which case, the person assumes the position and share of land of
that deceased sibling.
One case will illustrate the operation of these principles in com-
bination. Before World War II on the island of Beru in the Gilberts,
a woman cared for and was very friendly with a Banahan man, and
the latter indicated that he wanted her to become his sister.
On returning to Banaba, he informed his family of his wish. He
died on Banaba, and, after the war, the woman and her family
came to Fiji. On Rambi she was welcomed by the man's elder sis-
ter. At a meeting of the siblings and some other relatives a formal
i-mamane was proclaimed, and she was announced as replacing her
deceased friend. The latter had a daughter on Rambi, and his
Ocean Island land was divided between her and the adopted
woman. The arrangements for the inclusion of the daughter in the
woman's own property subdivision in the Gilberts were not yet
made, but such arrangements were contemplated.
Sibling adoptions are generally the 'returning of a kindness' of a
Gilbertese to a Banahan during a time of difficulty, especially
illness. Several cases originated during World War II. They can also
be grounded simply in 'the peoples' feelings'. The important point
to note here is that it is one's kinsmen who are supposed to be
supportive in times of trouble. The person who tenders this kind
of support and who is not in fact a kinsman is behaving as kinsmen
are supposed to behave. One does not really expect non-kinsmen
to behave in this way.
One particular kind of situation in which sibling adoption may
occur is when a Banahan sibling set adopts their non-Banahan half
sibling, who thereby becomes a Banahan and 'equal' to them. (The
situation arises, for example, when a Banahan man marries a
Gilbertese woman and has children with her; he dies, and the
woman marries a Gilbertese and has a child with him.)
There were two cases where a sibling adoption was not between
a Banahan and a Gilbertese, but between two Banabans. In the
first there was no transfer of Ocean Island land and none is en-
visioned. The man, however, is part-Gilbertese, and he said he
would give his adoptive sister a share of his land in the Gilberts. By
getting a piece of land on another island one has a 'place' there;
this is one of the gifts one bestows upon an adopted sibling.
A factor which may contribute tote i-mamane both in the sense
of 'the meeting' and 'brother-sister adoption', as in the case just
considered, is this: the normal pattern of 'friendship' is between
people of the same sex. Calling a non-kinsman of the opposite sex
a 'friend' rao can have overtones of a sexual relationship. The
word rao is used for 'friend', 'kinsman', 'spouse', and 'lover'.
Familiarity between two non-kinsmen of the opposite sex in the
same generation is an invitation to salacious gossip; te i-mamane
circumvents this.
The second case of sibling adoption between two Banabans was
different. Before death, a woman adopted a young man as brother
to replace her deceased brother and to care for her only daughter.
The woman was afraid, it is said, that her family might harm the
young and defenseless girl in order to get her land. The adoptee's
role was to protect her. He received the deceased brother's land,
but none from his natal utu; his social position derives from his
adoptive sister.
This adoption was rather like a retroactive child adoption. The
woman had no sons and no brother to look after her daughter; she
adopted a brother herself. One can speculate that it seemed to her
more reasonable to adopt a brother than a son since the brother's
generational position would give his word more weight with the
members of the family, who were suspected of being avaricious.
Actual bodily harm toward one's child after one's own death is
not generally expected from collateral kinsmen. Collateral kins-
men are supposed to care for orphans, but when they cannot be
counted on to do so, a solution can be sought in adoption.
One final question that must be attended to is the extent to
which adoption relates groups. The kinsmen of someone adopted
by one of my utu are not 'my utu', but kinsmen of someone who
is a kinsman of mine. It is still appropriate for me to extend cour-
tesies to them if we should be together; I do this as an expression
of my relationship with my kinsman. I may help my child prepare
a gift for the life-crisis rite of one of his adoptive utu, if the former
is living with me rather than with them at the time. I am doing it
'for him', as in helping a child who belongs to another church pre-
pare his contribution to the church, it is my relationship with him
rather than with his church that is concerned.
In contrast with the fosterage pattern described by Lambert for
the northern Gilberts, adoption is not a structural principle relating
groups in the society, nor is it the regular expression of a continu-
ing relationship between groups, although an adoption does have
implications for people other than the adoptee and the kindred of
the adopter. The idea of a heritable relationship involving two
units is rather found in 'the meeting', where neither utu actually
incorporates something which originally belonged to the other.
In the simplest cases of adoption certain values are obviously
fulfilled by "culture" stepping in where "nature" fails (as in the
case of a childless couple) or has not, as it were, had an oppor-
tunity to fail (as in the case of an unmarried person). Adoption
can also constitute the recognition of a state of affairs where ex-
pectations of the behavior of kinsmen as persons fall short of the
ideals of kinship behavior (e.g., to ensure the care of one's child,
or oneself); where the ideals of kinship behavior are enacted by
people enjoined to so act but who are considered particularly
liable not to (as in the case of stepparents); or where they are
enacted by people not even enjoined to do so (as in some adop-
tions of non-kinsmen).
An interesting feature of parental and sibling adoptions in par-
ticular, and true to a lesser extent of child and grandchild adop-
tions, is that they often represent the addition of kinship in the
categorical sense to kinship in the behavioral sense, or the conver-
sion of a 'feeling', as the Banabans put it, into a social position,
instituted through the creation of a land tie. Blood and land ties
arc both utu ties, and the ultimate statement of solidarity and
identity is the status of utu.
Adoptions in every form involve the bestowing of a kindness
upon the adoptee, and there are many cases in which the act of
adoption is formulated as an act of reciprocity for a kindness per-
formed by the adoptee or some of his kin. The giving of land can
perhaps best be interpreted as a giving of self, a giving of some-
thing that is personal, precious, and permanent. It is something
personal, but at the same time one's title to it is expressive of
one's utu membership and of one's identity as a Banahan. The
formulation of the relationships of the adoptee to the land, to the
adopter, to the adopter's and his own utu, and to the Banahan
collectivity if he is a foreigner are all of one piece.
To repeat a point made earlier, although land and blood operate
symbolically in some of the same ways, they are not equivalent. A
person cannot give his child blood on his father's side only, or
exclude it from his child's child. Land symbolizes kinship as apply-
ing both to a category of person and a category of behavior. Just
as a person can give land to someone who acts in the role of kins-
man but is not his utu-such a person then becoming his utu-he
can also give more land to the child who acts as a child should,
than to the child who does not.
This chapter is based on eighteen months' fieldwork on Rambi Island, Fiji,
during 1961, and 1964 to 1965. The support of the Social Science Re-
search Council and the American Educational Foundation in Australia is
gratefully acknowledged. David M. Schneider of the University of Chicago
and H. E. Maude of the Australian National University gave extremely
helpful advice at many points in the research. I am obliged to the Rambi
Island Council and the Government of Fiji for facilitating the research in
the field. Vern Carroll, Nicholas S. Hopkins, Bernd Lambert, Henry Lunds-
gaarde, and David Schneider made many useful comments on a previous
draft of the chapter. Much of the interviewing on individual cases of adop-
tion was conducted by my Banahan assistant, Kauongo Bio, while I was
engaged in other tasks.
2 Some Banabans are living in the Gilberts and other parts of Fiji. A Bana-
han Representative resides on Ocean Island.
3 See Schneider 1968 regarding the symbolism of blood.
4 The Gilbertese terms are: te natinati 'child adoption'; te tibutibu 'grand-
child adoption'; te i-taritari 'sibling of same sex adoption'; te i-mamane
'sibling of opposite sex adoption'; te i-tamatama 'father adoption'; te i
tinatina 'mother adoption'; and te onea-mwi 'the replacement'. The first
six are partially formed by the reduplication of kinship terms, e.g., nati
'child'. I- 'one another' is a reflexive prefix; its usage is consistent with the
nature of the adoptive forms in the terms for which it occurs. Questions of
terminology will be fully discussed in a future publication.
5 In his Gilbertese-French dictionary, Sabatier (1954:912) translates kau-
atabo as follows: "mettre, faire tenir ll chaque bout, chacun ll un bout."
Ka- is a causative prefix, and ua a morpheme meaning both 'carry' and
'two'; tabo means'end', 'extremity',and 'place'. 'Child of two places' was
how the word was explained to me (in Gilbertese) when I first heard it.
6 There is one case which represents a partial transfer to Rambi land of ideas
associated with Ocean Island land. A young man, X, for several years
helped a Banahan man, Y, and the latter's Gilbertese wife, Z, while the
couple's children were still young. The man is Z's sister's son. Y and Z
brought X with them from Banaba to Rambi. Y explained that he wanted
to adopt X, but the latter's family objected. They approved of X helping
his aunt and her husband, but wanted him ultimately to return to the
Gilberts. Y indicated that a compromise was reached by adopting X in the
nati-ni-kauatabo form, granting him Banahan status. But the only land he
is to receive is Rambi land. The understanding is still that X will return to
the Gilberts.
7 In the schedule of "customary land conveyances" drawn up by H. E.
Maude during the Ocean Island Lands Commission (copy in Files of Rambi
Island Council), it is stated that 'the accompaniment' passes to the family
of the adopter if the adoptee dies childless. Most local legal experts I con-
sulted on Rambi said that it would revert to the original utu, it would
'return to its proper place', although the matter is recognized as somewhat
8 At this point I am outlining the general conceptions of adoption and their
relationship to land. The actual legal transfer of title, especially when
Banabans are adopted, need not take place until some time after the adop-
tion has been informally arranged.
lJ From the ways in which some recent genealogies were recited, I suspect it
likely in many cases that the fact that a person is an adopted rather than a
"biological" child will be forgotten. It is also worthy of note that some
people adopted away without 'the accompaniment' are not included in the
genealogies of their natal groups.
I U With one exception, in all cases in which a Banaban was adopted as a child
by a consanguine, and their terminological relationship was known, the
adoptee was already terminologically 'child' (i.e., cognate, G-
) to the
adopter. In the one exception the adoptee was a classificatory 'grandchild'.
An Ellice woman adopted a close 'sister' as 'child', and my Banaban assist
ant, who had interviewed her, commented upon it as being rather bizarre.
11 The idea of a limited child adoption is not confined to the 'spouse-of
deceased-child' form. There is a controversial case which illustrates the
same kind of stipulation asserted for the instance just discussed. Before the
resettlement, G, a Gilbertese man, cured an old childless Banahan, B, of
illness. B adopted G. G had a child in the Gilberts who recently came to
Rambi, apparently to settle there. Some commented that this was wrong;
B had given G his land 'to eat from until he died.' It was an arrangement
involving only himself; he had violated its terms by "bringing his child" to
Rambi as if the child were a true Banahan. The decisive factor in cases of
this nature should be the original terms of the adoption, but those terms
may be under dispute.
12 During a discussion about a childless couple who had adopted, and later
had a child of their own, a Banahan suggested one motivation for adop-
tion: the continued presence of an infant could inspire an issueless couple
to fertility. This arose at a time when there was no further opportunity to
pursue it, and I do not know how general the concept is.
13 The problem of the definition of the incest boundary in Banahan kinship
is a complex one which will be treated elsewhere. There is general disap-
proval of sexual relations and marriage between people whose grand-
parents were siblings. There are alternative views which extend the line
even further but recognize an attenuation of it for a person and the mem-
bers of the utu from which he was adopted away.
14 In the view of some older Banabans, the forms of adoption also differ in
their origin. They assert that the parental and sibling forms are the result
of contact with Gilbertese.
15 Parental adoption is only envisaged where a true parent is deceased, pre-
sumably because a person must be dead to be replaced, since land is in-
16 This assumes that the economic advantage in holding provisional title to a
few pieces of Ocean Island land-which might amount to earnings of only
a few pounds per year from the phosphate company-would be out-
weighed by the element of obligation in the relationship.
17 This applies except for the "limited" child adoptions discussed above.
18 These phrases are, of course, metaphorical, since the people are living on
Rambi, not Banaba, and the cash received by individual land-holders from
the British Phosphate Commission does not provide their livelihood. David
Schneider has suggested that the land symbolism here may be similar to
the food symbolism of other Oceanic adoptive systems.
19 I include all the different forms under the term "adoption" because that
word seems best suited for the cultural domain under analysis. The authors
of the chapters in this volume found it difficult to agree on a sociological
definition of adoption, but I recognize that in some usages 'parental adop-
tion' might be considered a kind of reverse fosterage.
20 This case of sibling adoption between Banabans is one of the few cases
where a certain ceremony, in some ways similar to that reported by
Grimble ( 1957), is remembered to have been performed. The spirits of the
four directions were called upon, and the young man's change of name
declared. The name is sometimes changed in adoption; taking a name from
the adopting utu obviously symbolizes the change in social personality.
One explanation offered for the calling of the spirits in this instance was
that they should help the adopted brother in case some members of the
family had bad intentions toward him.
21 Relationships of 'the meeting' have been known to last two or three gener-
Henry P. Lundsgaarde
The present-day generation of Gilbert Islanders treats adoption as
a contractual or legal agreement between two parties: the adopter
and his consanguineal kinsmen constitute one party; the adoptee
and his consanguineal kinsmen constitute the second. As other stu-
dents of Gilbertese culture have described the general features of
both adoption and fosterage (Lambert 1964; also chapter 10 this
volume; Maude and Maude 1931; Silverman, chapter 8), I have
chosen a different and more specific analytic approach. A discus-
sion of some of the legal aspects of adoption, as exemplified by
data from Tamana Island in the southern Gilbert Islands, brings
additional features of the institution to our attention.
The relevance of introducing a legal perspective to the analysis
of ethnographic data is, briefly, that since every society prescribes
ideal modes of conduct for its participants, who, in turn, continu-
ally challenge or reformulate these ideals, the legal transactions
between people naturally embody the cultural beliefs, values, and
behavioral patterns defined as requisite to cooperation and social
continuity. We can, therefore, conceptualize adoption as a par-
ticular configuration of abstract behavioral principles supported
by social sanctions. The existence of complex relationships be-
tween such principles and actual social behavior were empirically
validated by Llewellyn and Hoebel in their penetrating study of
Cheyenne law (1941) and by Pospisil's analysis of the Kapauku
legal system ( 1958). Eight court cases from Tamana Island have
been included in this chapter to further illustrate the nature of
these relationships in the context of Gilbertese culture.
"A contract is a promise or a set of promises for the breach of
which the law gives a remedy, or the performance of whicl:i the
law in some way recognizes as a duty" (Atiyah 1961: 20). The
Gilbertese adoption contract most visibly involves the creation of
specific rights and duties connected with the use or owner-
ship of real property. It also signifies the public recognition of a
well-defined series of reciprocal economic and ceremonial obliga-
tions between two kin groups or between members of the same
family. These reciprocal obligations usually endure until the
adopted child reaches maturity. At that time the adopter dis-
charges the final part of his contractual promises by transferring
legal title in part or all of his estate to the adoptee. When this title
transfer has been approved by the local island Lands Court the
adoption contract has been discharged. While this frees both
parties from all further obligation toward each other it does not
mean that they thereby dissolve emotional or social bonds. If, on
the other hand, these emotional and social bonds break or weaken
before the title transfer has been finalized, the adoption contract
may be discharged by a breach of contractual obligation.
A breach of contract may be induced by members of either
party. An adopted child may, for example, be returned to his
family if he fails to fully assume the economic responsibilities
expected of junior kinsmen of his age and sex. The natural parents
of an adopted child may, on the other hand, attempt to terminate
the adoption contract if they feel that the adopter neglects to dis-
charge his obligations toward their child or themselves.
Although breach of the adoption contract is rare, it is signifi-
cant to note that it is precisely when such a "trouble case," i.e., an
exception to the rules governing adoption, occurs that we learn
what the rules are. Many of these rules pertaining to the distribu-
tion and definition of rights and duties are made quite explicit
when two parties declare their intention to enter a contractual
relationship before the island Lands Court. At that time, before a
proposed adoption has been approved by the court, the adopter
learns that he may have to obtain the consent of his parents, sib-
lings, or natural children before he formally enters an adoptive
relationship. The social ramifications which are created by an act
of adoption can be better understood if we focus attention on the
place of adoption in the sub-cultural legal system involving indi-
vidual and corporate rights, privileges, powers, and immunities
connected with the control of property.
The Gilbertese recognize two distinct modes of adoption, which,
in linguistic usage, are distinguished by applying either a frequen-
tative term for 'child' natinati or by applying the analogous term
'grandchild' tibutibu to denote the relationship. The crucial dif-
ference in the usage of these terms lies in the differential rights
and privileges conferred on the adopted person by his adopter. In
the case of the person adopted as natinati, these legal rights and
privileges are identical to those ordinarily conferred on natural
children by direct descent. The person adopted as tibutibu expects
to receive the rights and privileges normally extended to any
stranger who enters a formal friendship relationship with a non-
Although these two modes of adoption are well defined, it is
exceedingly difficult to be specific about the frequency of adop-
tion in the Tamana Island population. In the first place, the
adopter does not fulfill his part of the adoption contract until he
transfers the title of some of his property to the adopted person.
In the second place, my data do not show the exact age at which
various individuals were adopted. I simply wish to acknowledge
this lacuna in my data and, in passing, to call attention to the sta-
tistical aspects of the problem, because the frequency of adoption,
obviously, is directly related to the importance of adoption as an
element of Gilbertese social organization. The data summarized in
Tables 16 and 17 show the distinct preference among the Tama-
nans for adopting consanguineal kinsmen. Figure 14 illustrates the
high proportion of children in the island populace who, theoreti-
TABLE 16 Sex and Status of Adoptees
Ethnographer's Island Court
Census Records*
Relationship between
Male Female Male Female
Adopter and Adoptee No. %
No. % No. % No. %
Kin 26 39.4 24 36.4 12 48.0 4 16.0
Not Related 9 13.6 7 10.6 6 24.0 3 12.0
SUBTOTAL 35 53.0 31 47.0 18 72.0 7 28.0
TOTAL MandF 66 (100%) 25 (100%)
*Tamana Island government records (data are for 1964).
cally, would be eligible for adoption on the basis of age alone.
It is desirable to call attention briefly to the distinction between
legal adoption (i.e., an adoption registered and approved by the
island Lands Court) and customary or de facto adoption (i.e., an
adoption sanctioned by custom or kin group consensus). Table 16
illustrates the present disparity on Tamana Island between legal
and de facto adoptions. Customary adoption may be recognized as
legal if, as in a case of disputed inheritance rights, acceptable wit-
nesses will testify before the Lands Court and support the claim of
the adopted person against claims of the natural heirs.
The data summarized in Table 18 show that 92.5 percent of all
legal adoptions on Tamana Island are of the natinati variety. This
mode of adoption, according to the Provisional Lands Code estab-
lished by the Tamana Island Council (see below), confers specific
inheritance rights on the adoptee with respect to his adopter's
estate. The person adopted as natinati receives the same inheri-
tance status as his adopter's natural children. Adoption of a person
as natinati therefore reduces the size of the inheritance share
which would normally be transferred to an adopter's natural chil-
dren. This, in part, explains why disputes often arise between
adopted and natural children following the death of their com-
mon benefactor. In addition, it is now readily apparent why a per-
son usually must obtain the consent of his children before
adopting someone as his natinati.
On the other hand, the person adopted as tibutibu can only
TABLE 17 Preferential Modes of Adoption*
Kinsmen Non-kinsmen Total
Male Female MandF Male Female MandF Male Female MandF
No. % No. % No. % No. % No. % No. %
No. % No. % No. %
26 39.4 24 36.4 50 76.0 9 13.6 7 I0.6 16 24.0 35 53.0 31 47.0 66 100.0
*Data from ethnographer's census, 1964.
20 -

' .
' / ~
\' I
Totlll males: 568 (- - - - - -- )
T otlll females: 673 ( )
Totlll population: 1241
Adepted from McArthur and
Craig 1964: 107 (data .,. for
\" -.
I ~ \
, I t
', ,' ' .
.... .,, \ "
' ,---,
' " \
\,,' \
. '
' .
... \/.\
,, .
.... __ .
o-)) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) )
.. . .... ... <t ,, .,. ,, .t- tit <I' .<At </' # ..... .... </' 41
a ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
Figure 14. A1e and sex of Tamana Island residents
TABLE 18 Frequency and Mode of Legal Adoptions on Tamana
Island, 1957-1964 a
Relationship of Adopter(s)
Mode of Adoption
Year M F
to Adoptee Natinati Tibutibu
1957 + Not Related +
+ MoSi-SiSo +
1958 + WiBrSo-F aSiHu
+ MoSiSoDa-FaMoSiDa
1959 + SiDa-MoBr
Not Relatedc +
Wi(ex husband's)So-MoHu
+ Not Related +
1960 + BrSo-FaBr +
1961 +
Not Related +
+ F aSiSoDa-F aMoBrSoc +
+ Not Related +
+ WiBrSo-FaSiHu +
+ SiSo-MoBr +
+ (Wi)FaBrSoSo-FaFaBrDa(Hu)
1962 +
Not Related
(Hu)SiDaSo-MoMoBr(Wi) +
(Wi)FaBrDaDa-MoFaBrDa(Hu) +
1963 + Not Related +
1964 + (Wi)BrSo-FaSi(Hu) +
+ (Wi)SiSo-MoSi (Hu) +
+ Not Related +
(Hu)MoSiSoSo-FaMoSiSo(Wi) +
+ Not Related
TOTAL13 2 12 25 2
NOTE: M: Male Adopter; F: Female Adopter; M+F: Both Husband and Wife Adopters.
aTamana Island Government Records.
beoth husband and wife adopters.
c Adoption by two or more persons.
expect to receive a small share of his adopter's estate. This share is
often transferred at the time of the adoption or, in other words,
well before the time when the adopter dies or settles his estate.
The Provisional Lands Code for Tamana Island (l 6 June 1950)
provides, in part, as follows:
A 'land-of-the-adopted' may only be given if the adoption has previously been
approved by the Lands Court. An adoption shall be approved by the Lands
Court if it is confirmed that the natural children of the adopter or his next of
kin, if he has no natural children, will be left with sufficient land for their
support after the gift of a 'land-of-the-adopted'. If, however, the natural chil-
dren are guilty of neglecting their parent or live elsewhere and cannot look
after him, the adoption may be approved regardless of their claims. An adop-
tion may be annulled by the Lands Court on the request of the adopter if it is
confirmed that the adopted child has not fulfilled his obligations. If an adop-
tion has not been annulled prior to the death of the adopter, the adopted
child may be granted a 'land-of-the-adopted' by the Lands Court notwith-
standing the failure of the adopter to make such a bequest.
Only a brother or sister of the natural parent may adopt a child as his or
her son or daughter. In such adoptions the child may inherit land from the
adoptive parent and his or her spouse and may be disinherited by the natural
parent. In all other adoptions the child shall inherit land from his or her nat-
ural parents. An only child may not be adopted as a son or daughter but only
us a grandson or granddaughter.
An adopted grandson or granddaughter may not be given more than one
land plot, one taro pit, one pond and five niba [taro pits capable of holding
one plant] by the adopter. The adoption of a stranger is not prohibited. If
the adopted person dies without issue, the 'gifts-of-the-adopted' shall revert
to the donor or his heirs. If the adopted person dies leaving issue, the 'gifts-
of-the-adopted' shall pass to such issue, and the reversionary interest of the
donor is lost.
The above regulations governing adoption on Tamana Island
exemplify the relationships between Gilbertese conceptualizations
of proprietorship and adoption. These relationships similarly allow
I he ethnographer to conceptualize adoption as a contractual ar-
rangement between two principals (the adopter and the adoptee)
logether with their respective consanguineal kinsmen. The connec-
lion between these abstract principles and concrete social behavior
will be further illustrated in the following analyses of court litiga-
1 ions involving adoption.
The Tamana Island Lands Court, presided over by a Gilbertese
magistrate appointed by the Resident Commissioner of the Gilbert
and Ellice Islands Colony, regularly adjudicates all matters related
to adoption. In particular, the Lands Court officially approves or
disapproves any proposed adoption, decides disputes between
adopted and natural heirs in cases of inheritance, and codifies new
island regulations formulated through a consensus of island 'elders'
uni mane.
The following cases, heard by the Tamana Island Lands Court
between 1955 and 1963, illustrate the nature and social context of
disputes involving adoption. These cases, while simplified by the
Lands Court's scribe who usually manages to jot down only the
essential points of an often prolonged discourse, show how the
court frequently disregards written regulations in favor of situa-
tional variables, such as the age of the child to be adopted, the
economic situation of the adoptive parents, or the probable hard-
ships resulting from the enforcement of a particularly exclusive or
inclusive testament. It should be noted that the prefix Te denotes
masculine and Nei feminine gender.
The first case from Tamana Island concerns a formal request
from an unmarried woman for the adoption of her sister's infant.
The Lands Court had to decide whether a parent can recant his
promise to allow another person to adopt his child. The court's
decision, which favored the parent, indicates that the parent of a
child who is promised in adoption may recant such a promise If
the would-be adopter fails to show that he can satisfy the custom-
ary economic and social duties toward the child. A verbal agree-
ment between two parties cannot be said to comprise an adoption
contract unless either party can meet the obligations required to
make the relationship legal.
CASE NO.: 3/1955 Tamana Island Lands Court
NEI KANERE: I wish to adopt Nei Riteti, daughter of Te Beniata
and Nei Terenga. It is right that the Lands Court ask Nei Terenga
NeiKane"' c5 NeiTerenga 0 == D TeBeniata
I 6
"' I - - - - - - - - - - - - - .... Nei Riteti
Figure 15. Case 3/1955
if she will accept the adoption. For some reason she has now de-
cided that it would not be right for me to adopt her child despite
the fact that she earlier agreed that the adoption would be accept-
NEI TERENGA: I claim that the proposed adoption would be un-
wise as the child is too young to be adopted and Nei Kanere has
no husband who can help in the care and support of the baby. I
propose that she cancel her wish as she will certainly face a num-
ber of difficulties if she adopts my child.
DECISION: The Lands Court favors what has been said by Nei
Terenga. The proposal for the adoption of Nei Terenga's child is
The second case further illustrates the primacy of obligation as
an element of the adoption contract. The formal change in status
from a de facto to a legal relationship is marked by the registration
of a property title in the adoptee's name. This change may occur
when the adopter feels that the adoptee has satisfactorily fulfilled
his duties or when the adoptee appeals to the Lands Court for
part of his adopter's estate. To claim all or part of an estate, how-
l'Ver, a person must be recognized by the court as being legally
adopted. These ambiguities in legal status can create problems
when adoptive parents die without having registered the adoption
of a person who becomes a claimant before the Lands Court. The
udoptee may possibly win his case against natural heirs if witnesses
will testify on his behalf. For these, and other reasons, it is com-
111011 to find that Would-be adopters first obtain the consent of
their consanguineal kinsmen before they initiate an adoptive rela-
Nei Ata
Te Kaie
Te Biremon
Figure 16. Case 25/1955
CASE NO.: 25/ 1955 Tamana Island Lands Court
NEI ATA: I have treated Te Biremon as my adopted son ever
since his childhood. He has been faithful and looked after me.
Because of this I wish to register him as my adopted 'son' natinati
and to transfer to his name the title of some of my properties.
TE KAIE: I sincerely support her proposal of adoption. Te Bire-
mon has been true and faithful in his care of Nei Ata.
NEI ATA: I would like Biremon to have the following lands [stipu-
lated by name] and also the following taro pits [stipulated by
DECISION: The Lands Court agrees to the proposed adoption
and to the distribution of the properties, but the adoption can
only be confirmed when Te Biremon's parents on Arorae Island
have given their consent.
Following the court's request a telegram was sent to Te Bire-
mon's parents on Arorae Island. The reply received expressed the
parents' consent to the adoption.
DECISION: Te Biremon is registered as the legal titleholder of
the properties donated by Nei Ata.
The relationships between the elements of consent and obliga-
tion are amplified in the third case, in which a brother attempted
to prevent his sister from legally adopting a child who did not
belong to their own kin group. The decision of the court indicates
that a person has the right to contest an adoption by one of his
siblings if he feels that it may threaten his livelihood or, more im-
portant, the future welfare of his own children. When the brother
stated that the appellant "hates my children" he was, in effect, say-
0 !
11---0 l
<:J ~ , , JI
1 1 ~ ~
ing that his sister was ignoring her duty to adopt his child, as is
prescribed by custom. Although his appeal for cancellation of his
sister's petition was overruled, his objection to the adoption was
given serious consideration by the court. From this we can infer
that an adoption contract will be recognized as legal if the prin-
cipal parties to the adoption (the adopter and the adoptee's nat-
ural parents) both give their consent to the adoption. However, it
is also clear that the court will not approve such an adoption un-
less these two parties can satisfactorily demonstrate that the adop-
tion will not jeopardize the welfare of the adopter's consanguineal
kinsmen by depriving them of adequate land plots.
CASE NO.: 9/ 1957 Tamana Island Lands Court
APPELLANT: Nei Tutaima
RESPONDENTS: Te Tiaban, Te Teina, Nei Beteri, Nei Mamao
NEI TUTAIMA: I appeal to the Lands Court to register Nei Mamao
as my adopted 'grandchild' tibutibu.
TE TEINA: I do not object to the proposal.
NEI BETERI: I agree to the adoption.
TE TIABAN: I oppose the adoption because Nei Tutaima hates
my children.
TE KAURING: This must be well considered because the adoption
is opposed by Te Tiaban, Nei Tutaima's brother.
DECISION: Te Tiaban is not responsible for the support of Nei
Tutaima, and if she wishes to adopt the child, Nei Mamao, it is
agreeable to the members of the Lands Court. The adoption is
officially recognized by the court.
The element of mutual consent is once again stressed in the fol-
lowing case. The court, at first, seemed to waver in its decision to
approve the proposed adoption because of the complications
which could arise if the adopter's consanguineal relatives were not
consulted before the adoption was registered. However, since the
adopter was not residing on his home island (which, in fact, means
that he was deriving a living from his wife's estate), the court
simply ruled that they cannot be concerned with whatever compli
cations may arise on the adopter's home island. Note that this is a
natinati adoption and, therefore, of much more consequence than
the tibutibu adoption contested in the previous case.
{Beru Islander)
- - {from Tam1<11 Island) Te Ta......_
1 -----
pn>posed -
Figure 18. Case 40/1957
CASE NO.: 40/ 1957 Tamana Island Lands Court
Tee ....
Nei Bitita
APPELLANT: Te Mamanne (Beru Islander married to Tamana Island
TE MAMANNE: It is my desire to adopt Erate, the son of Te
TABUATA: It is agreeable to me if Te Mamanne wishes to adopt
my son, Te Erate, as his natinati.
N El BITIT A: I also agree to the proposed adoption.
DECISION: The adoption can readily be approved by the court
since both parents of the child have agreed to the adoption. Te
Mamanne, however, has relatives who must also give their consent
before the adoption can be legal.
However, since Te Mamanne is a Beru Islander, the Tamana Is-
land Lands Court really is outside of its jurisdiction. However,
both of the child's parents have given their consent and the adop-
tion of the child as te natinati is granted.
In the next case also we see the importance of consent. First we
note that all the siblings of the natural parent appeared in court to
voice their approval of the proposed adoption but that the
1u.loptee's mother, and her relatives, did not appear in court. I do
not know whether the adoptee's maternal relatives deliberately
sluycd away from the court, or whether they simply did not know
lhat their presence was required. The former assumption is prob-
ubly the correct one since most Gilbertese, although seemingly
casual about executing official obligations with any degree of
promptness, profess both interest in and knowledge of formal pro-
cedure in all important areas of their culture.

CASE NO.: 27/1957 Tamana Island Lands Court
APPELLANT: Nei Rongite
RESPONDENTS: Te Karibanang, Te Barenaba
NEI RONGITE: I appeal to the Lands Court to register Te Iere-
rua as my adopted 'child' natinati.
TE BARENABA: I accept the adoption.
TE KARIBANANG: I accept the adoption.
DECISION: It appears that the adoption is accepted by the
father's side only and not by the mother's side. The adoption pro-
posal therefore cannot be approved by the Lands Court until such
consent is obtained. Te Iererua's mother and her relatives will be
summoned to the next meeting of the Lands Court.
The two following cases are brief and uncomplicated. In the
first of these, we observe how the adoption contract is finalized by
the registration of a land plot in the name of the adoptee. Al-
though the adopter's natural children did not appear in court, they
had apparently voiced their consent. This slight oversight of a pro-
cedural technicality by the court may result in future litigation
between the adoptee and the adopter's natural children. However,
it must be acknowledged that the jurors and the island magistrate
probably already know if such complications would be likely to
arise, since they, like all other adult members of the Tamana Is-
land populace, have detailed knowledge of the state of social rela-
tions between family members in their small society.
CASE NO.: 30/ 1963 Tamana Island Lands Court
NEI TEAKO: I have come to make an allotment of a plot of
Nai Taako
Figure 20. Case 30/1963
Nai Mera
land in the tract called Aontekeri. I request that Nei Mere should
be registered as owner of plot No. 35. My children have given their
DECISION: Plot No. 35 on Aontekeri will be registered in Nei
Mere's name.
In the next case to be considered, a man has inherited by de-
fault a plot of land promised to a cousin adopted by his parent as
tibutibu. He corrects this injustice by transferring title in one of
his plots to his cousin. In this particular case it would have been
easy for the natural heir to keep all of his inheritance since his
rights have been confirmed by an act of the Lands Commission.
We see, however, the pride which the Gilbertese take in being
known among their fellows as honorable and generous in their
dealings with others.
CASE NO.: 4/ 1955 Tamana Island Lands Court
TE ETIA: I have come to return the land plot named Tebun to Nei
Tenoe. This plot was given to Nei Tenoe by my mother before all
the lands on the island were registered by the government's Lands
Commissioner. At the time of the Lands Commission (l 950's), I
won the plot back.
NEI TENOE: It is correct that the land plot was given to me by
Nei Tina before the Lands Commission. I do not oppose Te Etia's
proposal to transfer the title of that plot to my name.
DECISION: It is agreed that the land should be returned to Nei
Tenoe in accordance with the former will.
The final case illustrates the complications which often arise
when natural and adopted heirs have to divide an estate following
the death of their common benefactor. The appellant based his
claim for a greater share of his father's estate on the legal distinc
tion between the natinati and tibutibu modes of adoption. The
European officer who adjudicated this appeal upheld the appel
lant's claim by redistributing the lands so that each of the de
ceased person's adopted 'grandchildren' received one share of land
only and the claimant received the remainder of the estate since he

I i
I -E
1 I
('o I

l!I .8
19 -


:E "
l !I


was adopted as a 'child' natinati. In view of the fact that all the
legatees, with the exception of one person, were siblings, the re-
allocation of a few land plots might appear to be a trivial matter.
However, the four brothers were adopted in the first place because
their natural parents possessed very few lands which they could
transfer to their children. One of these (Te Robati), for example,
had no lands other than those he received from his adopter
(Lundsgaarde l 966:Appendix A-1 and A-2). Since Te Temake was
the eldest sibling in the group, he was given the preferential in-
heritance status conferred to a person adopted as natinati.
APPEAL NO.: l /1960 Tamana Island Lands Court (adjudicated by
a European government officer)
APPEAL OF CASE NO.: 20/ 1959
RESPONDENTS: Te Maunga, Nei Mereka, Te Robati, and Nei Atata
(last two respondents not present)
The appellant claimed that Te Robati, Te Maunga, and Te Maemae
were adopted by Te Tekeke as 'grandchildren' and not as 'chil-
dren'; therefore, they should have received only one land for the
adoption and he, Te Temake, should have received the remainder
of the estate. Te Temake further claimed that Nei Mereka was not
adopted. Respondent Te Maunga agreed that Nei Mereka was not
adopted but that she was nevertheless included as a benefactor in
Te Tekeke's will. It was stated that Te Tekeke was nursed by Te
Robati. The Lands Court members initially agreed during the
course of the proceedings that since the appellant had not ne-
glected Te Tekeke, the adopted 'grandchildren' were only entitled
to one land plot apiece, as provided for in the Lands Code. In fact,
however, the final decision followed the terms of Te Tekeke's will,
even though the Lands Code specifies that only wills which com-
ply with the Lands Code should be allowed. The magistrate, gov-
ernment scribe, and court members in this first "trial," as well as
Tc Tekeke in making his will, had also overlooked the fact that
one of the plots which Te Tekeke had willed was held by him as a
'land-of-adoption'; this land, upon Te Tekeke's issueless death,
should have reverted to the donor's family.
The distribution of Te Tekeke's estate was adjusted by the mag-
istrate after the court had heard Te Temake's second appeal. Each
of the respondents was given one parcel of land as his 'land-of-
adoption', and Nei Mereka was allowed by the court to retain the
small plot which had been willed to her by Te Tekeke. The
appellant, Te Temake, was awarded title to the remainder of Te
Tekeke's estate because he was Te Tekeke's adopted 'child'.
We will now examine briefly the role of adoption in intra- and
inter-family relations. The court cases presented in the previous
section illustrate that Gilbertese adoption involves a large number
of people who are related to the two principal parties. These
people partly participate in the reciprocal obligations created
between the adopter and the adoptee; they may be directly con-
cerned with the transfer of property which has to take place if the
adopter honors his obligations toward the adoptee. This transfer
of property rights and privileges is a likely source of conflict be-
tween the benefactor, his natural or adopted heirs, and all other
potential beneficiaries who have any legal claim to the benefac-
tor's estate. If this transfer of property takes place within the
family it simply serves to consolidate an ancestoral estate and re-
affirm reciprocal obligations between kinsmen (see Fig. 23).
It is generally recognized that one quick way to discourage
reciprocal expectations, i.e., reduce the demands for economic
goods or services, is to refuse to adopt one's sibling's child or,
worse, to adopt someone who is not a relative. Such an act is inter-
preted as an insult against one's siblings, and it most effectively
serves as a warning signal to one's natural heirs that, unless they
prepare to take better care of their aging parents, their inheritance
share will be markedly reduced. We note, however, that most
adoptions on Tamana Island have involved members of the same
consanguineal family unit (see Tables 16, 17, and 18). In this way
adoption reinforces family solidarity by establishing additional ties
between kinsmen, and it is not to be regarded solely as a con-
venient way to re-allocate rights and privileges to scarce natural
resources among destitute members of the family. Conflicts involv-
ing equity (i.e., problems relating to the interpretation of cus-
1 1 ~ ~
~ ~
_, __ j
I \ ~
\ 1 - ~
\ \ I I
\ I I
\ I\ I
I I \
I /\ \
I \ \
1 1 ~
tomary rules, principles, or precedents to be followed by the two
parties who have entered a contractual relationship) could also be
avoided if only childless couples were permitted to adopt children
from destitute families. As the matter stands, however, the adop-
tion of a non-relative, particularly if the adoption is of the natinati
variety, is commonly regarded as a direct slap in the face to one's
consanguineal kinsmen and leads to fragmentation of an ancestral
estate and weakening of reciprocal obligations within the corpor-
ate kin group (see Fig. 24).
The contrast between the information derived from the eight
court cases concerning adoption and the various motives for adop-
tion outlined by Tamana Island informants in interviews illustrates
the complexity of adoption as a social institution. The most fre-
quent responses to questions about the motivations for adoption
l. The desire to transfer property if a couple is childless.
2. Provision for a destitute or orphaned child who in terms of
inheritance has been slighted by his own family members.
3. Insurance of a comfortable old age by taking into the house-
hold a younger person who will supply his adoptive parents with
food and generally perform the necessary tasks requiring the
strength and endurance of youth.
4. The reciprocation and solidification of a cordial social rela-
tionship with members of a different kin group, or the strengthen-
ing of the emotional and economic ties with one's sibling or more
distant consanguineal relatives.
These stated motivations for adoption, as well as the traditional
procedures followed by all parties involved in an adoption, have
not been significantly altered since the 1930s when the Maudes
first described and analyzed Gilbertese adoption (Maude and
Maude 1931 :225-232).
Although the data from the court cases illustrate the behavioral
reality of these motivations, they also suggest that the Gilbertese
conceptualize adoption as more than a simple procedure for creat-
ing families for childless couples or as a countermeasure against
the debilities of old age. If, for example, childlessness were a pri-
mary motive for adoption it would be difficult to understand why







































such a high proportion of Gilbertese with natural children should
choose to adopt additional children when they already have large
families to support. It would also be difficult to comprehend why
people living at a marginal subsistence level would choose to be-
come involved in long-term economic obligations when they al-
ready are burdened by obligations toward consanguineal kinsmen.
The discrepancy between the motives for adoption and actual
behavioral performance has partly been explained by noting how
the contractual relationship between two principal parties can lead
to conflicts among their consanguineal kinsmen. These secondary
parties may either gain or lose when the adopter transfers title to
par. of his estate to the adoptee and thereby concludes his final
legal obligations specified by the adoption contract.
In any situation in which a person acts upon a motivation lim-
ited to specific ends in such a way that his relationships with
others are modified, the modifications generate "feedback" into
the motivational system. The obtaining of this feedback may be-
come a primary source of motivation. In the case of adoption, act-
ing upon a motivation limited to obtaining an adopted child
almost certainly entails the antagonism of those whose legacy is
thereby diminished. In an occasional case, the provocation of this
antagonism appears to become an end in itself.
This is a revised version of a paper read at the annual meeting of the Ameri-
can Anthropological Association in Denver in November, 1965. I wish to
thank Vern Carroll, Ward Goodenough, Bernd Lambert, Michael Lieber,
Martin Silverman, and the other contributors to this volume for valuable
criticisms and suggestions. Data on adoption were collected in the year
1964-1965 during fieldwork in the southern Gilbert Islands with the finan-
cial support of a National Science Foundation grant administered by
Homer G. Barnett, and a Woodrow Wilson Dissertation Fellowship. Addi-
tional grants from the University of California and the National Institute
of Mental Health (# I R03 MH 13042-01) permitted me to revisit the
Gilbert Islands during the summer of 1966 and helped to defray the costs
of revision of the earlier version of this chapter ..
Bernd Lambert
In a number of Polynesian and Micronesian societies the relation
of a chiefly descent group to the members of descent groups of
lesser status resembles a kinship tie. It is commonly believed that
the founders of such related groups were actually consanguineal
kinsmen, and that their descendants are still bound by the obliga-
tions appropriate to such relationships. For example, on Truk, the
commoners of a district trace their genealogies back to men of the
chiefly matrilineage. Relationships between high- and low-ranking
lineages are patterned after those between a matrilineage and the
children of its male members (Goodenough 1951: 135-137).
The analogy between a political connection and a kinship tie
tends to be restricted to the allocation of property rights, the exer-
cise of authority, and the display of formal deference. Although
the idiom of kinship may adequately express political relation-
ships, members of an aristocratic descent group are not expected,
in everyday social life, to treat commoners as they would their
own brothers or sons.
The northern Gilbert Islands provide an x ~ m p l of a system of
social stratification patterned on a special form of nonconsan-
guineal kinship. On the atolls of Butaritari and Makin, political
relations were traditionally modeled on a contractual bond that I
will call "guardianship. "
A "guardian" offers to rear the child of
another family in return for certain land rights. He does not exer-
cise a true parent's authority over the child unless, as happens
occasionally, he later adopts his ward. However he must give his
young ward preferential treatment, and bring ceremonial gifts to
an adult ward. All of the members of a child-giving group are con-
sidered superior in rank to the guardian.
The political significance of guardianship relates to divided
possession of land. Title to an estate is divided between "provi-
sional" and "residual" owners, a practice which has been reported
for several other Micronesian societies. The general features of this
custom are the same wherever it occurs: "Provisional" titleholders
are given the right to use the property, sometimes in perpetuity. In
return they owe certain goods or services to the donors, who
retain a residual title (cf. Goodenough 1951: 38-42, for exemplifi-
cation of the Trukese case). In the northern Gilberts, provisional
title to property is given to a child's guardian by the child's par-
ents to compensate the guardian for his trouble. The guardian's
rights pass to his descendants, who are expected to continue the
relationship with the residual owners, usually by rearing one of the
ward's children. Since the ward and his children retain a residual
title to the land, the property is eventually subject to overlapping
claims. Were it not for the custom of guardianship, land rights
would be less debatable, since siblings generally partition their
hereditary estate so as to give each heir exclusive rights to a parti-
cular share.
In traditional northern Gilbertese society, divided title to land
was the basis of the relationship between aristocratic and com-
moner descent groups. Most estates on Makin and Butaritari be-
longed jointly to two descent groups. The commoners, as provi-
sional titleholders, had the primary responsibility for cultivating
the land. Some commoners believed that their founding ancestor,
possibly an immigrant, had begun the association by assuming the
guardianship of an aristocrat's child. In any case, all commoners
tended to behave as though in some sense they were the guardians
of their superiors. Even the terminology of guardianship was car-
ried over to the political relationship.
The first part of this chapter will place guardianship in the
framework of northern Gilbertese kinship by contrasting it with
sibling and adoptive relationships. It will be noted that a brother
and sister sometimes divided the title to a garden plot and that
they might thus found, respectively, a commoner descent group
and an associated aristocratic descent group. Adoption, as well as
guardianship, will be discussed in some detail in order to facilitate
comparison with the other societies described in this book. The
second part of this chapter will point out the correspondences
between guardianship and the commoner-aristocrat relationship.
To my knowledge, Butaritari-Makin is the only Oceanic society
that patterned its system of social stratification after a kind of
"fictive kinship."
The northern Gilbertese case should, therefore,
be of considerable interest.
In the final section of this chapter, Butaritari-Makin guardian-
ship will be compared with analogous institutions in other Micro-
nesian societies. It will be shown that guardianship can be con-
sidered as the northern Gilbertese equivalent of "complementary
filiation" -the relationship between an individual and the lineages
of the parent from whom he does not trace descent (Fortes
1953: 33). There are marked resemblances between the Gilbertese
guardian-ward relationship and the father-child relationship in
Truk. Since descent in Truk is matrilineal, father and child belong
to different corporate groups. Similarly, a northern Gilbertese
guardian, although he is often consanguineally related to his ward,
stands outside the highly solidary group of near cognatic kinsfolk.
In both cases the relationship is predicated on divided ownership
of land: in each case, the user of the land owes the residual title-
holder deference and gifts but is not really subject to the latter's
jural authority.
It is significant that aristocratic-commoner relationships are
expressed in an idiom of guardianship in Butaritari-Makin and in
an idiom of patrilateral filiation in Truk. If it turns out that extra-
lineage kinship serves as a model for political superiority elsewhere
in Micronesia, it will be possible to speak of an ancient pattern
that has persisted despite changes in descent rules. Unfortunately,
while the custom of dividing title to land appears to be wide-
spread, it cannot yet be consistently related to social stratification.
The coral islands of Butaritari and Makin
have traditionally
formed a single society ruled by a high chief. In precolonial times,
there were three well-defined status levels:
chiefs (the high chief
and his siblings), aristocrats, and commoners. Certain rights over
land were associated with each status level. The end of the old
social order may be dated from 1922, when a British colonial
officer altered the system of land tenure (Lambert l 966a: 658-
660). Social stratification and descent-group solidarity began to
crumble after that. But customs relating to adoption, guardian-
ship, and kinship behavior had changed relatively little up to 1961.
Hence the past tense is used to describe traditional political insti-
tutions, and the present tense describes practices actually observed
in the course of fieldwork.
There were ten autonomous villages on Butaritari and two on
Makin. A council of elders, under the presidency of an hereditary
headman, directed community affairs. Most of the land in the
neighborhood belonged to cognatic descent groups localized in the
village. A typical descent-group estate comprised at least one
village house site, several nonadjacent plots in the forest, and some
sunken gardens for growing atoll taro (Cyrtosperma chamissonis).
A few estates also included large sailing canoes, fish ponds, and
portions of the shore, all of which like land were inherited and
transferred. The membership of a descent group consisted of
persons who could trace a genealogical connection to the founder
through either men or women, and who had inherited a taro bed
in the group's garden. On the average, a native of Butaritari or
Makin probably belonged to five or six such descent groups. I have
called these nonexclusive cognatic groups "ramages," a term pro-
posed by Firth (1957:6) and Murdock (1960: 11 ). The equivalent
Gilbertese word is to utu.
Most ramages were divided into segments, each descended from
a child or grandchild of the founder. Within about three genera-
tions of the founding of the ramage, the segments became differ-
entiated into a localized "core" (ta kaainga) and several peripheral
branches. Some of the men of the core segment, and an occasional
woman, actually lived on the estate with their spouses and chil-
dren. The resident members constituted a kind of extended
occupying a single large house and sharing a daily food
supply. Persons belonging to the peripheral branches did not have
the right to reside on the estate.
The localized core of the ramage made greater use of the estate
than did the peripheral branches. The core members, whether or
not they resided on the house site, had the exclusive right of regu-
larly collecting coconuts, pandanus fruit, and breadfruit from the
The Cyrtosperma garden, unlike the forest lands, was par-
titioned among the sibling groups of the ramage. Ownership of a
taro bed was a criterion of ramage membership. Nonresident mem-
bers often permitted resident members to cultivate their shares.
The eldest male member of the core segment was normally man-
ager of the ramage estate and the group's spokesman on the village
council. The core in general took the initiative in most ramage
activities (Lambert 1966a:652). The presence of two core seg-
ments, each with its own house site, forest lands, and manager,
indicated that the ramage was in the process of fission. The non-
localized collateral segments gradually disaffiliated from the ram-
age, unless the core became extinct and the collateral segments
inherited its rights.
0 After the land reform of 1922, all branches
of a ramage were given equal access to the estate. It then became
expedient for people to claim membership in all the ramages with
which they were affiliated by descent (Lambert l 966a:658-660).
The sibling group, consisting of a set of middle-aged brothers
and sisters and their children, was (and is) strongly corporate. Full
siblings jointly inherit their parents' Cyrtosperma beds in several
ramage gardens. In the traditional society, they usually did not
formally partition their inheritance until old age, and they pooled
any new acquisitions. They made a single contribution of large
taro corms to a kinsman's life-crisis feast. They were again assessed
as a unit when one of their ramages prepared a gift for presenta-
tion to the high chief.
Siblings still rely on one another for unstinting assistance in
material goods and labor. A sibling's adolescent son or daughter is
often "borrowed" to help with the work in an overburdened
household. Yet it is acknowledged that brothers may dispute an
inheritance or be rivals over a woman. One brother is quick to
resent another's receiving a larger gift from their father. Disputes
arise also over alleged failures to play a full part in collective enter-
prises. Half siblings and first cousins, as members of the same ram-
age segment, join together in public affairs but assist one another
less than full siblings do. Cousins are also more likely than are
brothers to engage in open quarrels.
Nowadays, sons and daughters generally inherit land on the
same basis. In the past, a sister's share might be smaller than her
brother's, but she usually received rights to a taro bed in each of
her parent's gardens. Because she normally lived in her husband's
house, her taro beds were likely to remain in the hands of one of
her brothers, who was regarded as holding provisional title to
them. Whenever possible, the inheritance of a sibling set was in-
formally apportioned among pairs of siblings of opposite sex. A
man was responsible for making contributions of taro in his paired
sister's name as long as he cultivated her garden beds. As the
residual titleholder, she expected to receive a hospitable welcome
in her brother's home when she came to visit, and to find a refuge
there when she quarreled with her husband. The father and the
paired mother's brother usually sponsored separate life-crisis feasts
for a boy or a girl. If a woman's sons later took full possession of
their maternal inheritance, the woman's brother was relieved of his
special obligations and thereafter behaved toward his sister's chil-
dren as a classificatory father.
This section has dealt with relationships created by "descent,"
construing this term to include nonunilineal transmission of kin-
group membership (Firth 1963:23). On Butaritari and Makin,
kinsmen by descent possessed similar shares in a piece of land,
although some might live away from the estate and leave others to
cultivate it. They were also equal in rank and had joint responsi-
bilities toward the rest of the community. Authority was in the
hands of the men of the older generation and, to a lesser degree, of
the older brothers within a generation. The strongest authority,
that of a father over his sons, derived partly from the fact that the
sons benefited from the father's property and expected to inherit
it after his death. Brothers were thus under the same system of
authority and had parallel obligations toward one another. On the
other hand, the relationship between siblings of opposite sex often
involved divided title to garden plots and asymmetrical mutual
obligations; but in succeeding generations few distinctions were
made between the descendants of brother and sister. The next
section will show how adoption creates relationships resembling
those which stem from descent. A subsequent section will indicate
how relationships through guardianship lead to perpetually asym-
metrical obligations.
Butaritari-Makin appears to have two institutions, adoption and
guardianship, that correspond to adoption in the southern Gilberts
and Banaba. However, most of the adoption customs described by
the Maudes ( 1931) and by Lundsgaarde and Silverman in this
volume have analogs in the north, occurring in the contexts of
either adoption or guardianship.
In the Gilberts, except in the northern islands, a child may be
adopted either as 'son' or 'daughter', or as 'grandson' or 'grand-
daughter', depending either on his (or her) genealogical position or
on the adopter's relations with other kinsmen. The adopter re-
ceives a present of land from the child's natural parents and later
leaves some or all of his own land to the adoptee. On the two
northern islands, an 'adopted' child generally assumes the status of
a 'son' or 'daughter' and always inherits some land from the
adopter. The status of 'ward' in the northern Gilberts resembles
that of a child who is being reared by his grandparents-a common
practice in both the northern and southern islands. The guardians
may, in fact, refer to their ward as their 'grandchild'; unlike true
grandparents, they receive land in return for their trouble. They
need not leave property to their ward unless they also adopt him.
One of my informants summed up the distinction with the state-
ment, "You give property to your [natural or adopted I child, but
you receive property from your ward."
In Butaritari-Makin, adoptive relationships closely resemble
wnsanguineal ones, although an adopted child does not always
receive all of the rights which a legitimate son or daughter acquires
at birth. The Gilbertese term for 'adoption' is te natinati (from te
11ati 'child' in the sense of "offspring"). The corresponding transi-
tive verb is natina 'to possess (or treat) as one's child'. Nowadays,
a person wishing to adopt another has the latter's name entered
among the names of his heirs in the Lands Register. In former
times, an announcement in the presence of the interested kinsmen
was sufficient.
The use of the English word "adopt" implies that there is no
biological connection between the adopter and the child. But the
Gilbertese verb natina signifies merely that the social status of
'child' was bestowed on someone by a voluntary act.
people say that the father of an illegitimate child often decides, or
is compelled, to 'adopt' natina it, thereby permitting the adoptee
to inherit land on the same basis as his legitimate children. How-
ever, if the 'adopted' child is the fruit of a casual liaison and re-
mains with his mother, the father may neglect him completely and
contribute nothing to the child's support, contending that the
bastard is 'not a true child'. Such a man is severely criticized if he
asks the child to assist him when the child reaches adolescence. A
man without legitimate children may be eager to help rear an ille-
gitimate child, since the man will probably be dependent on that
child for support in his old age. If a woman has more than one
lover during the period when she is presumed to have conceived,
her baby may be adopted by two (very rarely by three) men, al-
though it is perfectly clear to the Gilbertese that there can be only
one genitor.
An adulterine child is almost always regarded as the social off-
spring of his mother's husband, unless she was divorced soon after
giving birth. The child acquires a second father if his genitor de-
cides to adopt him. He is then jurally under the authority of two
fathers, since he uses taro beds and coconut lands belonging to
both. He usually prefers one man over the other, however. The
adulterine children whom I knew personally were strongly at-
tached to their mothers' husbands, in whose homes they had
grown up.
Many adoptions, of course, have no connection with extra-
marital conception. The adopted child acquires another parent, or
another set of parents, without losing his rights to the estate of his
natural father and mother. There are several reasons for adopting a
child. The most common is to fill a hiatus caused by sterility or
premature death. A person without issue often adopts a sibling's
or a cousin's child (i.e., a classificatory 'child')
to care for him
in his old age and to inherit his share of the taro beds. Occasion-
ally a sibling's grandchild is adopted, in which case the act is called
tibuna 'to possess as a grandchild', instead of natina.
Many childless couples have two or more adopted children in
their homes. It is common for husband and wife each to adopt a
consanguineal relative in order to keep land rights within the ram-
age. Some spouses who feel particularly close adopt one another's
kinsmen. For example, the wife may adopt one of her husband's
brother's daughters, and the husband, one of his wife's sister's
sons. A few couples jointly adopt a child. A childless guardian
generally adopts his ward, either in place of or in addition to a
sibling's child.
Sometimes parents whose grown children have died adopt
young persons as replacements. The adoptee is either the spouse of
the deceased or a friend, of the same sex. The need for a replace-
ment is most keenly felt after a child's suicide, an act which is
often motivated by anger against the parent.
Adoption also provides a way for an adult to formalize his rela-
tionship with a boy or girl whom he has been treating as his own
child. Such an adopter is often the adoptee's distant classificatory
parent and is usually on close terms with the adoptee's family. For
example, a woman can assume the role of 'mother' toward another
woman's baby by nursing it. She will probably adopt it if it
belongs to the first descending generation of her ramage, or make it
her ward if it is two generations removed from her. A non-relative
would never be asked to act as wet nurse.
People may adopt a neighbor's child who has been spending a
lot of time at their house, since the youngster has already acted as
the adopter's son or daughter by helping with small household
tasks and by eating food the adopter has provided. Adults will not
adopt a visiting child if they dislike his parents, nor will they do so
if they have children of their own and the visitor is a sibling's son
or daughter. The loan of a child is, after all, the sort of assistance
one expects from other members of the sibling group, and no
special recompense is offered. At most, a classificatory child who
has helped with the household work over a period of time will
receive a taro bed when he or she marries.
'Strangers' frequently become members of the community
through adoption. Many of these immigrants are boys or girls
brought to Makin by kinsmen for extended visits. The adopter is
often related in some way to the child's host-for example, as the
guardian of one of the latter's own children. A grown man who
settles on Makin for any length of time, even as a schoolteacher or
a radio operator, is often adopted by a resident of the island. This
type of adoption occurs if neither the man nor his wife can claim
membership in a local ramage. A 'foreigner' who has married a
Makin man or woman is not usually adopted, because it is felt that
the native spouse's lands and kinsmen are at the couple's disposal.
Guardianship serves as an alternate means of assimilating immi-
grants. The stranger may become the guardian-but not the
adopter-of a native child, and receive provisional title to land or
taro beds in return.
The adoption of immigrants has probably become more impor-
tant during the colonial period, as the land under the personal con-
trol of chiefs and village headmen has decreased. A newcomer can
no longer obtain land merely by becoming guardian of the chiefs
child or by performing some nominal service for the chief himself.
But there is evidence that the adoption of foreigners is an ancient
custom. As early as 1895, for instance, a co-headman of Makin
Town adopted Nan Takirua, the son of immigrants from Abaiang.
Nan Takirua received a large taro garden from the headman, but
he was not a legatee in his adoptive father's estate (which was
divided among the adopter's natural children). The adoptive rela-
tionship was continued by the headman's eldest son, who reared
Nan Takirua's son and grandson.
This example indicates that an adopted child does not quite
become the equal of his adoptive brothers and sisters, particularly
if they are of high rank. He usually inherits somewhat less of his
adoptive parents' taro gardens than they do. Young people
adopted for friendship's sake are often satisfied with a bequest
from their adopter of a token taro bed. They are said to 'pity' the
natural children, whose inheritance would be appreciably reduced
if the adoptee were to receive a full share of the estate. An adult
adoptee may even cancel his adoption by surrendering his entire
share of the estate of his adopter. An adopted son and his descen-
dants are not eligible to succeed to the managership of the
adopter's ramage unless they are also affiliated to the group by
descent. This restriction seems to apply also to the sole heirs of
childless adopters. Nor can adoptees become village headmen,
since the status of headman is defined with reference to his man-
agership of a particular aristocratic ramage.
Butaritari-Makin adoption can be construed as a mechanism for
creating the social statuses of 'parent' and 'child'.
It provides
children where they are wanting due to sterility or premature
death, and parents for a child whose natural mother or father has
been left behind on a distant island. Adoption also creates a jur-
ally-binding parent-child relationship to validate certain biological
or behavioral relationships. The 'adoption' of an illegitimate child
by his genitor harmonizes social and biological fact. The adoption
of a child by a neighbor or a distant classificatory parent gives
assurance that both adopter and adoptee will continue to fulfill
their obligations. A parent's sibling usually does not adopt a 'bor-
rowed' child because the two already have binding obligations
toward one another.
My ethnographic census of Makin Island, carried out in 1961,
shows a surprisingly small number of adopted children. Of 639
children under the age of sixteen, only seven were living with
adoptive parents. These figures are somewhat misleading, however,
because most adoptions seems to be formalized only after the
child reaches adolescence, or even later. A childless adult generally
'borrows' a sibling's child, making the child his heir only after he is
assured of the young person's support in his old age. Some immi-
grants are adopted as adults. As expected, five of the seven house-
holds with young adopted children had no natural children of
either husband or wife. There was a total of thirty-five house-
holds without natural children of either spouse, and of these, all
but eight contained either adoptees, wards, or junior kinsmen.
While adoption establishes a parent-child relationship similar to
that resulting from legitimate birth, guardianship seems to imitate
the relationship between grandparent and grandchild. The ward is
usually the guardian's classificatory grandchild or classificatory
great-grandchild-that is, a kinsman two or three generations fur-
ther removed from a common ancestor. He is most likely the
guardian's parent's sibling's great-grandchild, or grandparents' sib-
ling's great-grandchild; a sibling's grandchild can be adopted but is
considered too closely related to be taken as a ward (Lambert
1964:236-238). The most important exception to this generaliza-
tion is the custom of allowing an immigrant to become guardian of
a local aristocrat's child.
Most people who initiate guardianship do so after their own
children have reached adolescence. They then rear a second fam-
ily, which includes the ward along with their natural grand-
children. The ward is treated with even greater indulgence than the
grandchildren are. The verb tibuna 'to take as a grandchild' is
sometimes used for the assumption of guardianship, as an alterna-
tive to tobana 'to nurse' and kaikawaia 'to rear to maturity'. The
self-reciprocal term tibuu 'my grandparent/grandchild' may also be
used by the guardian's descendants and the ward's successors (see
p. 282) to refer to one another. However, a ward, unlike a natural
or adopted child or grandchild, does not regularly inherit the guar-
dian's land or gain additional ramage affiliations from him.
The desire to rear an additional 'grandchild' is never the sole
reason for obtaining a ward. Guardianship, unlike adoption, is
always arranged very early in a child's life, sometimes even before
he is born. Prospective guardians have ties to the natural parents,
but seldom any attachment to a particular child. They are often
the close friends, as well as the distant parents or grandparents, of
their ward's parents. By becoming guardians, they transform an at-
tenuated consanguineal connection into a stronger relationship.
The children of wealthy or high-ranking people are much more
likely to acquire guardians than are those whose parents are poor
commoners. Wards are often of higher, and never of lower, rank
than their guardians. In the traditional society, it was practically
certain that most of a chiefs or aristocrat's children would be
taken as wards (Parkinson 1889:33). The guardians counted on
receiving both land and protection froin the powerful man in re-
turn for rearing his child and becoming, in a sense, his followers.
Even now, a man who has become relatively wealthy as a store-
keeper, or one who owns more garden beds than the average, will
be solicited by prospective guardians. Immigrants, who have tradi-
tionally ranked lowest on the social scale, can obtain both land and
a place in society by becoming guardians. They have to rear all of
their own children as well.
The child's mother or father gives the guardians provisional title
to several taro beds, and sometimes also to coconut land or a
canoe. The guardians may extend their provisional rights over any
canoe the ward acquires after he grows up. All this property is sup-
posedly transferred only to ensure a supply of food for the child.
However, like other provisional titleholders, the guardians have the
power to dispose of as they please the coconuts, taro, and fish pro-
duced from their ward's property.
They are, of course, respon-
sible for supporting the ward as long as he lives with them. In fact,
they are obliged to 'cherish' the ward by giving him preference
over their own children in the distribution of choice foods and
scarce goods, and in general by letting him have his own way.
After the ward has left their household, the guardians continue to
fulfill their obligations, either by rearing one of his children or by
presenting him with huge taro corms from time to time. Such gifts
are traditionally expected at life-crisis feasts and in time of illness.
Recently some families have been making them at Christmas as
well. It adds to an aristocrat's prestige to have a crowd of guar-
dians and their kinsmen arrive at his house carrying taro on their
shoulders. In former generations, guardians publicly honored high-
ranking wards on other occasions too. They did this, for example,
by spreading pieces of new cloth over a female ward's head as she
walked to a dance.
The ward is seldom incorporated fully into the guardian's fam-
ily. The only significant exceptions occur when guardians' mar-
riages prove sterile. A childless guardian will probably adopt his
ward, and thereafter treat him as he would a natural child. Other-
wise the true parents retain all their jural authority over the child.
For example, a ward's marriage always required the consent of his
or her father. The guardian may be consulted in advance, but he
has no power to prevent the wedding. After marriage, a male ward
generally resides on a house site belonging to one of his natural
parents; a female ward joins her husband. Both will derive ramage
affiliations and land rights from the natural parents only.
A ward of either sex should allow his or her guardian to rear
one of the ward's own children. Children may, in fact, be trans-
ferred in this fashion for four or more generations. The guardian's
obligations to successive wards are inherited jointly by all of the
guardian's children and grandchildren. The heirs also share provi-
sional title to the property which makes up the guardianship gift.
They thus constitute a sub-ramage, with special rights and duties
distinguishing them from the other branches of the ramage. Like
other ramages, the guardian's ramage-considered collectively 'the
guardians' -soon becomes differentiated into a core segment and a
p1.:riphery. Only members of the core utilize the guardianship gift,
and only they really concern themselves with the original ward's
grandchild. Consequently, if the child-givers wish to continue the
relationship into a third generation, they make a new gift of provi-
sional land rights to the localized core of the child-rearing group.
The guardians are, by this time, considered to be the full owners
of the original gift.
The guardian-ward relationship resembles in many ways other
Butaritari-Makin kinship ties. The households of the ward's par-
ents and of his guardians cooperate as near kinsmen in subsistence
and craft activities and render mutual assistance in emergencies.
The grown children of wards and guardians find a welcome in one
another's homes, especially when they travel to other villages or
islands. More specifically, the relationship of ward to guardian is
that of a junior to a senior kinsman. A child owes a lifelong debt
to the adults who reared him, whether they be his parents, his
foster parents, or his guardians. The younger person pays an in-
stallment on this debt whenever he hands over some item of per-
sonal property requested by the older person.
This obligation
weighs a little more heavily on wards than it does on natural chil
dren. The guardian, like a parent, is usually attentive to a grown
ward's needs; it is not polite to ask him directly for a gift.
It is with respect to land rights and formal obligations that the
guardian-ward relationship occupies a unique place in the kinship
system. The child-givers permit the guardians to assume provi-
sional title to coconut lands, taro gardens, or canoes. In exchange,
the guardians provide the food their ward consumes as a child, and
they bring him formal gifts, mainly of taro, after he grows up. The
guardians also 'honor' their wards and receive 'protection' from
them in return. The ward's parents, the ward himself, and the
ward's successors clearly occupy the superordinate status in this
relationship. Yet they are not entitled to make decisions for the
guardians; they do not exercise the authority of a father, an elder
brother, or a ramage manager. I have the impression that, because
their respective rights and duties are different, guardians and wards
are unlikely to feel, as do real and classificatory brothers, that a
gain for one implies a loss for the other. This attitude, along with
other features of the guardian-ward relationship, was carried over
into the relationship between aristocratic and commoner ramages.
In 1961, the number of children who lived with guardians was
only a little larger than the number of children living with adop-
tive parents: 25 out of a total of 639 individuals, or less than 4
percent of the population under age sixteen. This figure under-
states the social importance, and even the true frequency, of guar-
dianship. Some wards were probably living in their natal house-
holds at the time of the census: hardly anyone resides with his
guardians continuously from weaning till marriage, and an occa-
sional child never moves in with his guardians at all. The impor-
tance of guardianship is largely due to its hereditary character.
Many of the descendants of the original parties have obligations
stemming from guardianship even if they themselves neither grew
up with guardians nor reared a ward.
The aristocrats of the era before land reform were all descendants
of either the high chief Nan Teauoki (mentioned in Wilkes
1845:84) or the Makin Town headman, Nan Temango. Both of
these men probably lived toward the close of the eighteenth cen-
tury. A high chief or a Makin headman, unlike an ordinary ramage
manager, was succeeded by the oldest son of his principal wife
(Wilkes 1845: 85). Most of the numerous siblings and half siblings
of these two dignitaries received estates and founded aristocratic
ramages. Differences in land tenure make it convenient to distin-
guish a category of Butaritari aristocrats descended from Nan
Teauoki and a category of Makin aristocrats descended from Nan
Temango, even though local usage denotes both types of aristocrat
by the same term. Neither category was restricted to a single is-
land, although most of the lands belonging to Makin aristocrats
were located in the vicinity of their home village, Makin Town.
The great majority of estates were owned jointly by aristocratic
and commoner ramages. A shared title to land, of course, implied
a close social relationship. The principal kinds of aristocratic-
commoner relationships were alleged to date from the founding of
the ramages concerned. These relationships were as follows: co-
rcsidence, in which aristocrats and commoners occupied the same
house site, and "linkage," in which a commoner ramage was asso-
ciated with aristocrats residing in another village. More recent ties
between aristocrats and commoners have grown out of the guar-
dianship of an aristocratic child by people of lower rank. Although
the details of the aristocratic-commoner relationship varied con-
siderably from one estate to the next, the general outlines were
similar everywhere. The commoners, who held provisional title to
the estate, usually had primary responsibility for cultivating the
land. The aristocrats, as residual titleholders, were supposed to ask
the commoners' permission before collecting coconuts from the
forest lands or planting taro in a special garden bed.
The most intimate relationship possible between an aristocratic
and a commoner ramage was co-residence. Most Butaritari and
Makin aristocrats were, in fact, localized on estates also occupied
by commoners. Each of the two co-resident ramages had its ex-
tended family, composed of some members of the core segment,
their spouses, and their children; but both families lived in a single
large house and ate food cooked in a single oven. Aristocrats and
commoners appear to have behaved in everyday life as kinsmen.
Butaritari and Makin aristocrats acquired estates and co-resident
subordinates in different ways. The high chief assigned estates to
almost all of his own children and to some of his siblings' children,
who thereby became the founders of new Butaritari aristocratic
ramages. A few of the founders were sent to occupy ownerless
estates; most of them, however, received residual title to lands
belonging to ramages that had fought on the losing side in a war
over the high chieftainship. Even though some of the vanquished
ramages had been aristocratic, the victors had reduced them all to
a rank below that of native (nonimmigrant) commoners. These
unfortunate descent groups were now compelled to support co-
resident aristocrats.
A number of new Butaritari aristocrats were granted hereditary
offices, including the headmanship of every village except Makin
Town. An aristocrat who had been selected as village headman was
always endowed with a large estate and with commoners to culti
vate it.
The Butaritari aristocrats superimposed their own division of
the Cyrtosperma gardens on that of the co-resident commoner
ramages. Each of the original aristocrat's children and grand
children thus held residual title to taro beds belonging to one or
more commoner sibling groups. The commoners were charged
either with the support of the aristocratic extended family as a
whole or with providing tubers for the particular aristocrat whose
share of the garden corresponded to their own. The aristocrats,
besides helping with the routine tasks of cultivation, grew tubers
of extraordinary size in special beds made up of borrowed sections
of the commoners' regular beds. The aristocrat and the commoner
who had lent the aristocrat part of his taro bed brought these huge
tubers as gifts to the high chief and as contributions to the life-
crisis feasts of kinsmen. The aristocrats were forbidden to give the
tubers to any other commoners. All coconut land was controlled
exclusively by the commoner ramages, though aristocrats could
borrow plots to make copra.
Some of the founders of Makin aristocratic ramages occupied
land inherited from their father, the headman. Others were per-
mitted by the high chief to settle on estates belonging to com-
moners. Since these commoners, like the aristocrats themselves,
were members of ramages native to the island, they did not con-
sider themselves inferior to the latter. They insisted that they had
given the aristocrats rights to their lands only on the chiefs re-
quest. Hence they could not be compelled to raise taro for the
benefit of the aristocrats. Instead, each ramage was allotted about
half the beds to cultivate. The produce of the entire garden was
pooled for the support of the composite household. The coconut
land remained under the control of the commoners.
Immigrants from the southern and central Gilberts, and com-
moners who had fled Butaritari after the high chief had seized
their estates, were assimilated at the bottom of the social hier-
archy on Makin. Together with the fallen aristocrats who had
become co-resident subordinates of new village headmen, they
ranked somewhat below other commoners. An immigrant or ref-
ugee family often begged land from a Butaritari or Makin aristo-
crat. The newcomers might have to spend a generation or so in a
condition approaching slavery before they were allotted taro beds
on the same basis as other co-resident commoners. This method of
naturalizing immigrants produced some complicated social ar-
rangements. For example, a village headman's ramage might have
co-residential rights on several estates, some occupied by the origi-
nal owners, others by the descendants of immigrants or refugees.
A few estates were occupied by three ramages simultaneously-one
aristocratic, one native commoner, and one immigrant.
Commoners were exempt from all the burdens of co-residence
when residual title to their estates was held by aristocratic ramages
localized in other villages, especially Butaritari Town. I have else-
where called this relationship "linkage" (Lambert 1964: 234;
Lambert l 966b: 160-161 ). The relationship is usually said to have
been initiated by the founding ancestor of a native commoner
ramage. The link was created by making one of the high chiefs
children, the subsequent founder of an aristocratic ramage, the
protector of the commoners. The preferred linking aristocrat was
the son or daughter of a woman of the commoner ramage who had
married the high chief. Since membership in the chiefly status
level took precedence over other kinship ties, the high chiefs
wives did not pass descent-group affiliations and land rights to
their children in the ordinary way. A chiefs wife's brother thus
acquired a perpetual provisional title to lands that the chiefs wife
would normally have inherited. They owed her and the chief regu-
lar gifts of food in return. By becoming a linked aristocrat, the
chief's child in effect maintained his or her mother's residual title
to the estate. When the chiefs child was unrelated to the com-
moners, the linked relationship more closely resembled guardian-
ship. Other linkages seem to have been established by aristocrats
who granted separate estates to co-resident commoners. A single
aristocratic ramage might be linked to commoners in several vil-
lages, and have co-resident subordinates besides.
The linked commoners had complete control of their estate.
Except on special occasions, the aristocrats received none of the
taro, breadfruit, or coconuts that the land produced, nor could
they reclaim the land itself if the commoners had long been settled
there. Under the traditional system of tenure, aristocrats were in
fact so little concerned with the lands of their linked commoners
that they were ignorant of the boundaries of the taro beds. The
aristocrats had to be shown these boundaries when the Lands
Commission altered custom by giving them half of their former
subordinates' gardens.
The commoners' provisional title to the estate included rights
over any canoe belonging to a linked aristocrat. A man who had
taken over an aristocrat's canoe generally permitted other mem
bers of his home ramage to use the vessel. The residual owner of
the canoe, if he lived in the commoner's village, received part of
each day's catch. A guardian who controlled his adult ward's
canoe paid a similar "rent" in fish.
A principal duty of linked aristocratic and commoner ramages
was to extend hospitality to each other's members. When Buta-
ritari aristocrats visited the village of Kiebu on Makin Island, for
example, they stayed with their linked subordinates, who provided
them with food and did not ask them to do any work. Aristocrats
were never more than temporary guests in the homes of their
linked commoners, however. The very lavishness of their reception
made them reluctant to prolong their stay and so impose an un-
necessary burden on their hosts. Some aristocrats repaid the hospi-
tality they received by erecting a house in their home village for
the use of visiting commoners.
Associated ramages of aristocrats and commoners, whether co-
resident or linked, acted as single corporate groups whenever cir-
cumstances required the transfer of land or canoes. Since part of
the jointly-owned estate had to be surrendered as compensation
for murder, adultery, or theft, both ramages suffered when a mem-
ber of either committed a serious offense. If a plot had been seized
as a penalty for a relatively minor transgression, an older aristocrat
might approach the head of the ramage that had taken it and plead
for the return of the land. The aristocrats were expected to assume
leadership when the commoners became embroiled in feuds with
outsiders and in the seizure of land.
At least part of the aristo-
cratic ramage was entitled to share in any land acquired by the
commoners as a result of a dispute. The same rule applied to most
instances in which outsiders had voluntarily transferred land to the
In contrast, aristocrats who acquired real estate were not ob-
liged to share possession with their linked or co-resident com-
moners. Small plots were usually left in the hands of any co-
rcsident or linked commoners who happened to live nearby. But
larger acquisitions tended to be reserved for new dependents.
Occasionally aristocrats even reassigned provisional title to por-
t ions of their co-resident commoners' own property. By doing so,
howt:ver, they risked being deserted by the commoners, who could
usually obtain fresh land elsewhere or subsist on the estates they
daimed through other ramage affiliations.
An aristocrat with land at his disposal was likely to grant provi-
sional title to the guardians of his children. This made it possible
for immigrants from the south and families who had been deprived
of their land to acquire new estates by taking the children of local
aristocrats as their wards. The newcomers might later rear a child
and a grandchild of their first ward, but their status eventually
became indistinguishable from that of other co-resident or linked
commoners. Sometimes the ward was detached from the donor
ramage and founded a new aristocratic descent group with residual
title to the guardianship gift (Lambert 1964:253-254). I suspect
that this form of guardianship increased in importance in the late
nineteenth century, when aristocrats were no longer able to keep
immigrants in servitude.
Some children of chiefs and aristocrats were reared by their
classificatory grandparents, whose rank was only slightly inferior
to their own. A few of these same children also became the wards
of immigrants, since their high rank permitted them to have more
than one guardian. Although a native guardian, such as an aristo-
crat's classificatory grandparent, always received provisional title
to some land, this guardianship gift was not regarded as a new es-
tate, and the guardians did not found a new ramage subordinate to
the ward. The child served, instead, as a bond between established
and independent descent groups. A child of very high rank might
become the ward of his classificatory grandparent's entire ramage,
or even of co-resident ramages of aristocrats and commoners. The
guardians usually constituted only a localized ramage segment,
however. The relationship persisted only so long as the guardians
reared a lineal descendant of the original ward in each generation
(Lambert 1964: 236-238). The ward acquired no rights over the
bulk of his guardians' estate unless the guardian's ramage segment
died out. In the nineteenth century, when sterility was more prev-
alent than it is today, a child of a high chief or village headman
might in fact inherit from several guardians (Kramer 1906:320).
It is evident that the relationship between aristocrats and com-
moners could take any of a number of forms, even within a small
On some estates, members of the headman's ramage
Jived off the labor of commoners, to whom they denied all auton-
omy in community affairs. On others, aristocrats and commoners
worked side by side as near-equals. Still other estates were under
the sole control of commoners, who met their linked superiors
mainly at feasts and on occasional visits. Further complications
ensued because aristocrats, and especially the headman, could
create new bonds of dependency by granting land to the guardians
of their children. The next section will examine the common
characteristics that underlay this diversity of relationships.
The correspondences between guardianship and the aristocratic-
wmmoner relationship exist mainly on the level of latent struc-
ture. The overt ideology of social stratification does not insist that
aristocrats be treated as wards. As we have shown, the old men of
Makin state that most of the commoner ramages were associated
with particular aristocratic descent groups from their foundation
onward, either because one ramage settled on the other's land, or
because a commoner woman bore the high chiefs child, or be-
cause the commoner's ancestor had 'taken' a chiefly child to be his
'protection'. My informants rejected the suggestion that 'taking'
the high chief's child had been equivalent to becoming the child's
guardian. They pointed out that the commoners were associated
with an entire aristocratic ramage, not merely with a line of suc-
cessive wards, and that the aristocrats' children were seldom reared
by either co-resident or linked commoners. On the other hand,
recent immigrants who actually became the guardians of a head-
man's child entered into a relationship with the headman's ramage
that was practically indistinguishable from the older commoner-
aristocratic relationships.
l have tried to show that all provisional titleholders, whether
they were commoner ramages or the guardians of particular wards,
had essentially similar rights and obligations. Just as a guardian
received provisional title to a piece of land from one of his ward's
parents, so all commoners were perpetual caretakers and culti-
vators of an estate to which aristocrats retained residual rights.
The commoner ramage, like a guardian, had a provisional land-
owner's obligation to extend hospitality to the aristocrats and to
provide them with food if necessary. The similarities between the
ward-guardian and aristocrat-commoner relationships extended
even to the terminology employed by the two parties and to
behavior not directly connected with land-holding.
Members of co-resident or linked ramages considered one an-
other as kinsmen. The commoners were spoken of as the 'parents'
or 'grandparents' of the aristocrats. The latter were sometimes
called the 'children of the land' ateein te aba (the word teei or
ateei means 'young person' or 'junior kinsman' generally, but not
specifically 'offspring' or 'relative of the first descending genera-
tion', for which the term is nati; the reciprocal of teei is unim 'aane
'old man' or unaine 'old woman'). The self-reciprocal term tibu
'grandparent/grandchild', which could denote a guardian or ward,
was also used in reference to the co-resident aristocrat whose taro
bed matched that of a particular commoner. Now that hereditary
status levels are becoming obsolete, the young descendants of asso-
ciated aristocrats and commoners often believe themselves to be
distant consanguineal kinsmen. One such relationship came to
light in the course of an ethnographic census. A young man of an
aristocratic ramage who had provided a home for a girl of a for-
merly co-resident immigrant ramage described her as his 'mother',
without being able to trace his genealogical connection to her.
Evidently a family tradition of friendship and cooperation af-
fected his behavior even after its basis had been forgotten.
The mutual obligations of associated commoners and aristocrats
were more strongly institutionalized fifty years ago than they are
today. These obligations were probably less extensive than those
owed to near kinsmen, to guardians, or to wards. However, com-
moners performed much the same services for aristocrats that they
performed for their own kinsmen outside the sibling group and for
the first- and second-generation descendants of their wards. An
aristocrat living elsewhere was taken to the home of associated
commoners when he fell ill, and he was cared for there. An aristo-
crat could request the assistance of linked commoners whenever
there was a great deal of work to be done in his Cyrtosperma beds,
even though these might be part of estates in which these particu-
lar commoners had no direct interest. Commoners could be asked
to help build a house for the aristocrats to whom they were
linked. An aristocrat's consanguineal kinsmen also assisted him
with his house-building and gardening. Neither linked commoners
nor distant kinsmen could be compelled to participate in these
activities, and their labor had to be repaid by a meal at the end of
each working day. Commoners probably received help on occasion
from aristocrats, although they were 'ashamed' to ask the aristo-
crats for it.
Reciprocal attendance at life-crisis feasts has always been an
obligation of kinship. When a man sponsors a feast for his child or
for his sister's child, all members of his own sibling group are
expected to attend, and the other sibling groups of the ramage seg-
ments to which he belongs are expected to send representatives.
Participation by people outside these ramage segments is confined
to friends of the sponsor who can claim consanguineal or fictive
kinship with him. Everyone who attends the feast contributes at
least one large taro corm. Members of associated aristocratic and
commoner ramages have traditionally come to one another's
feasts. Even though many other customs of the past have lapsed,
this obligation is still taken seriously. It is noteworthy that Buta-
ritari aristocrats formerly attended the life-crisis feasts of their
linked commoners as cooks rather than as contributors. My infor-
mants explained that the co-resident commoners, as provisional
owners of the gardens on the aristocrats' home estate, would have
resented the use of their taro for the benefit of persons unrelated
to themselves. But there may have been another reason. The
aristocrats owed access to land and canoes to both co-resident and
linked commoners; they did not, however, owe them gifts of food.
Aristocrats exercised little authority over any commoners ex-
cept those belonging to some co-resident ramages. A father or
ramage manager certainly demanded more obedience from his
dependents than most aristocrats did from their nominal subordi-
nates. Like the donors of guardianship gifts, they could not easily
disturb the rights of well-established provisional owners of an
estate. The logic of divided land rights thus left the estate, and
especially the coconut land, under the control of the commoner
manager. Even co-resident aristocrats were reluctant to give com-
moners instructions about growing taro.
Residual titleholders were paid formal respect in acknowledg-
ment of their superior rank. It will be recalled, for example, that
guardians are more willing to satisfy the whims of their wards than
of their own children. Commoners deferred to the manager of a
co-resident aristocratic ramage by allowing him to sleep and sit in
the place of honor under the middle rafter on the eastern side of
the house. Aristocrats also had the prerogative of receiving the
heads of sharks caught by either co-resident or linked commoners
who happened to live in the same village. Failure to honor an
aristocrat was considered to be supernaturally dangerous, but
there is no recollection that any secular punishments were im-
Marriage between members of either co-resident or linked ram-
ages was normally prohibited, even though ramages were not ex-
ogamous groups, and status-level endogamy was merely preferen-
tial, not prescriptive. Persons belonging to the same ramage could
marry if they were at least four generations removed from a com-
mon ancestor. It was also possible for aristocrats, especially those
descended from a remote chief, to marry commoners with whom
they did not share an estate. The mutual exogamy of associated
ramages was thought necessary to prevent individuals from oc-
cupying contradictory statuses. At least in theory, no suitable or
consistent attitude could be displayed toward a person who had
inherited residual rights to an estate through one parent and pro-
visional rights to the same estate through the other. It was believed
that the child of a 'mixed' marriage would simply be treated by
the commoners as a consanguineal kinsman and denied the respect
due an aristocrat. The same objections are still raised against mar-
riages between descendants of guardians and wards, or between
persons who are an unequal number of generations removed from
a common ancestor and therefore stand toward one another as
classificatory 'parent' and 'child'.
The data given in this section indicate that the relationship
between aristocratic and commoner descent groups was patterned
after two kinds of interpersonal relationships-that between a
ward and a guardian and that between a married woman and her
paired brother. Some aristocratic and commoner ramages ex
plained their joint ownership of an estate by a tradition that their
respective founding ancestors had actually been ward and guar-
dian, or sister's son and mother's brother. Even ramages that
lacked this tradition employed some of the terminology of guar-
dianship (e.g., the 'grandparent/grandchild' term for the joint
owners of a taro bed) to refer to their aristocratic or commoner
associates. The behavior expected from members of an associated
ramage was not precisely that expected from an individual guar-
dian or ward. However, most of the reciprocal obligations of
aristocrats and commoners resembled those imposed on persons
related by guardianship. In a general sense, superiors and subordi
nates treated one another as distant kinsman.
Since similar patterns of behavior were prescribed for wards and
guardians, for sister's sons and paired mother's brothers, and for
associated aristocratic and commoner ramages, it is convenient to
introduce a collective designation for all three relationships. I will
call these relationships "complementary," because each party per-
formed different services for the other. The parents of a ward, the
woman who had married into another ramage, and the ccrresident
or linked aristocrats were all considered donors of usufructuary
rights to land. In return, the guardians, brothers, or commoners
provided gifts of food. There was often an exchange of intan-
gibles: when the donors of land rights were of higher rank, they
received deference and offered protection. In all these respects,
the complementary relationships contrasted with the relationships
between brothers and between segments of the same ramage.
All of the complementary relationships implied a division of
land rights between residual and provisional titleholders. The
holders of provisional title had exclusive use of the property but
undertook certain obligations toward the residual owners. The
provisional titleholders could sometimes take for their own use
lands, or canoes, belonging to the residual titleholders. Both pr<>-
visional and residual titles, together with their attendant obliga-
tions, were theoretically inherited by all the descendants of the
initial holders.
A complementary relationship endured only so long as descen-
dants of the original parties continued to divide rights to land.
Either owner normally inherited full title to the property if the
other died without issue. For example, an aristocratic or com-
moner ramage became sole owner of an estate at the extinction of
the other descent group possessing rights to the land. An aristo-
cratic-commoner relationship might otherwise last indefinitely.
The complementary relationships between persons were generally
of more limited duration than were those between groups. Unless
a woman had married the high chief, one of her sons usually took
control of her taro beds, thereby ending his special relationship
with her paired brother. Guardians became the sole owners of the
original guardianship gift after about two generations. They con-
tinued to rear descendants of the initial ward only if they received
provisional title to additional land.
The major obligations of the provisional titleholders all entailed
contributions of food, especially taro. Guardians, of course, pro-
vided food for the wards growing up in their households; the guar-
dianship gift was conventionally described as 'the source of the
ward's food', as if to emphasize this aspect of the guardian's role.
Certain ramages of commoners grew taro for co-resident aristo-
crats who were called their 'grandchildren' (meaning 'wards' in this
context). Nonresident residual titleholders received gifts of food at
life-crisis feast and other occasions. The modern practice of bring-
ing taro to some adult wards at Christmas is an example of how
provisional titleholders fulfill this obligation. In the past, the high
chief's wives' brothers, who were the permanent provisional
owners of their sisters' taro beds, had to bring regular gifts of food
to the chiefs household. Another old custom was recorded in
Parkinson's statement that "tenants" had to deliver a specified
number of coconuts to the aristocrats (1889:99). Gifts of food
from the residual to the provisional titleholders were less frequent.
Wards, who were residual titleholders, contributed taro to their
guardians' life-crisis feasts, but some aristocrats attended their
linked commoners' feasts as cooks. Finally, any residual title-
holder received unstinted hospitality in the homes of people who
cultivated his property.
Guardianship, co-residence, and linkage established close ties
between groups or individuals of different rank. Ideally, such rela-
tionships were advantageous for both superiors and subordinates.
Several episodes in the traditional history of Butaritari and Makin
show that a high-ranking ward could count on the loyalty of his
guardians when he was under attack by his enemies. A grown ward
or an aristocratic ramage were in turn expected to protect com-
moners against locally-powerful men who threatened their lives or
property. The residual titleholders were sometimes treated with
formal respect, but actually exercised little authority over their
The relationships among segments of the same ramage reflected
the kinship ties of their immediate founders, who had usually been
brothers. Although the ramage had certain rights over the estate as
a whole, each segment had exclusive control over its ancestor's
portion. The mutual responsibilities of the segments were quite
similar. However, the segments were distinguished from one an-
other by complementary ties that crossed ramage boundaries.
Through guardianship, a segment could acquire land that had not
formed part of the ancestral estate. Larger segments attained the
status of autonomous ramages by establishing a link with a new
aristocratic ramage, especially one founded by a child of one of
the commoner segment's women.
In the matrilineal societies north and west of the Gilberts, com-
plementary relations obtain between father and son, and between
a lineage and the children of its men. Goodenough states that on
Truk, a form of gift called niffag creates a state of divided owner-
ship when applied to productive property. The provisional owner
is obligated to give the residual titleholder a share of the produce
or of the sale price of the property, but he inherits full title even-
tually (l 951 :37-39). An individual matrilineage member holds
provisional title to the land he cultivates-title which derives from
the lineage as a corporate body. Children also hold provisional title
to some land belonging to their fathers' matrilineages. These land
rights serve to differentiate the children and their matrilineal
descendants from the rest of their lineage, just as the acquisition
of a guardianship gift helps to differentiate ramage segments in the
northern Gilberts (Goodenough 1951: 42-45 ).
The relations between a Trukese chief and the people of his
district correspond closely to those between parties in divided
ownership, with the chief as residual owner. Assumed patrilateral
ties connecting the chiefly lineage with other lineages make the
district a social, as well as a territorial, unit (Goodenough 1951:39,
In the Marsh alls, divided ownership results from the transfer of
land, called in ninnin, from the father's matrilineage to his chil-
dren. The transfer is not absolute, since the land is said to revert to
the father's matrilineage after a few generations (Spoehr 1949: 167-
168; Tobin 1956: 16-19, 23-24). It is interesting that Marshallese
chiefs granted land rights to commoners who adopted the chiefs'
Matrilineages on the atoll of Lamotrek, west of Truk, make gifts
of land to the children of male lineage members and to adopted
l:hildren. The plot reverts to the donor if the recipients abandon it.
The descent group using the land must present offerings of food to
the donor lineage when the latter holds a funeral or repairs a canoe
house. According to Alkire, "This gift symbolizes the recognition
of the residence group that the donor lineage actually retains re-
sidual rights to the land" (1965: 50-54, 114 ). In several eastern
Micronesian societies, then, complementary relationships outside
the corporate descent group seem to adhere to the same pattern.
Only the kinship idiom in which they are expressed varies accord-
ing to the rule of descent.
Some of the other petty states of Micronesia resemble Butari-
tari-Makin in that one particular descent group, or two intermarry-
ing groups, hold the highest rank and fill all the principal offices.
The aristocratic groups possess superior rights to land, which they
permit commoners to cultivate in return for formal gifts. Except
for Truk, however, there is little evidence that the relationship
between status levels is modeled on a kinship relationship. For
example in the Marshalls, a paramount chief, a lesser chief, and the
head of a commoner matrilineage all hold certain rights to most
pieces of land (Spoehr 1949: 162-163; Tobin 1956:6). Marshallese
chiefs do not necessarily claim any kind of kinship with their sub-
jects, although some of the lesser nobility is composed of the sons
or daughters' sons of chiefs (Spoehr 1949:76). On Ponape, the
chiefly status level consists of two series of four matriclans each.
The two series intermarry and alternately provide the district
chief, so that each incumbent of the highest office is a patrilateral
kinsman of his predecessor. A wide gap separates the nobility from
the common people (Garvin and Riesenberg 1952:202). The ab-
sence of cross-status kin ties in these societies may be connected
with the chiefs' autocratic powers, which exceed those institution-
alized on Truk. This would fit Sahlins' generalization that in the
most highly-stratified Polynesian societies chiefly ramages deny
any kinship with commoners (1958: 161-167).
This chapter is based on data collected on Makin Island from June, 1959
to October, 1961. The research was sponsored by the Tri-Institutional
Pacific Program under a grant from the Carnegie Foundation. I am grateful
to Vern Carroll and Henry Lundsgaarde for their comments on earlier ver-
sions of this chapter, and to Joan Miller and Terence Turner for their help
with the final draft.
2 This point has been made for Polynesian societies by Sahlins ( 1958: 140-
151 ). The expression of political dominance in a kinship idiom is not
limited to Oceania. Among the Yao, a Central African society, a village
founder's relationship to the headman of his parent village is perpetuated
by his successors. Mitchell writes: "This relationship, although phrased in
a kinship term and moulded to some extent by the norms of kinship be-
havior appropriate to the term used, is in reality political, for it fixes the
relative positions of two villages through their headmen" (1956: 122). In
other societies, matrilateral cross-cousin marriage provides a basis for dis-
tinguishing superior and inferior groups (Leach 1961 ).
3 I have previously used the term "fosterage" for this relationship (Lambert
1964). However, contributors to this volume agreed to reserve "fosterage"
to describe more informal arrangements.
4 A connection between fictive kinship and social stratification has often
been reported from other parts of the world. For example, South Slavic
godparenthood shows a number of interesting parallels to north Gilbertese
guardianship. In parts of Serbia, Macedonia, and Bulgaria, the right to act
as godparent to members of a patrilineal extended family is inherited in
the male line. The major difference between the Slavic and Gilbertese insti-
tutions is that in the former case the godparents are of a rank higher than
that of the children they sponsor (Hammel n.d.).
5 A Trukese father has almost no authority over a grown son (Goodenough
1951: 112; Schneider 1961 :224).
6 Butaritari Island is generally called "Makin" on American maps; Makin is
"Little Makin" or "Makin Meang," when it is shown at all. The two main
villages have the same names as their respective islands. I use "Butaritari
Town" and "Makin Town" to distinguish them.
7 I follow Sahlins (1958:2-3) in calling categories of rank in a kinship so-
ciety "status levels" rather than "social classes." The use of "status levels"
for the Gilbertese categories calls attention to the fact that an individual
can belong simultaneously to two descent groups of different rank.
8 Butaritari-Makin residential groups do not quite fit Murdock's concept of
the extended family (1949:32-35), since, while young married couples
normally reside with kinsmen, they do not all reside with the parents or
mother's brother of either spouse. It would be more accurate to use the
term "clan," which Davenport (1959:562) and Murdock (1960: 11) have
applied to the localized form of a ramage. Unfortunately, the use of a
familiar word in a relatively unfamiliar sense seldom clarifies matters.
9 In a previous publication (Lambert l 966a: 654), I wrote that the extended
family was called ta kaainga and that it had exclusive rights to harvest
coconuts and fruit from the ramage estate. A new look at genealogical
materials has led me to realize that rights were not so narrowly defined.
The definition of kaainga in this chapter is virtually the same as that given
by Goodenough in his account of Gilberiese social organization (1955:
10 This was the main, though not the only, reason why ramage size remained
more or less constant from generation to generation. For other limitations
on ramage membership, see Lambert 1966a:655-658. The Butaritari-Makin
ramage resembled the Samoan "sept," which Ember describes as a non-
unilinear descent group with a single localized segment (1962:965-966).
11 Natina is used in a different sense in expressions such as e rangi n natinaki
iroun tinana 'he was spoiled by his mother' (literally, 'he was treated very
much as a child by his mother').
12 Gilbertese classify consanguineal kinsmen by generation and sex. Only six
terms are used; they can be translated 'grandparent/grandchild', 'father',
'mother', 'sibling of the same sex', 'sibling of the opposite sex', and 'child'.
In practice, distinctions are made between lineal, near collateral, and dis-
tant collateral kin.
13 Adoptive siblingship, which Silverman reports as common among the
Banabans, is very rare on Butaritari and Makin. I recorded only three cases;
all involved young people wishing to formalize their friendship.
14 Lundsgaarde (chapter 9) gives a similar motive for adoption on Tamana.
15 The motives for guardianship are discussed at greater length in Lambert
16 The following data are taken from Lambert 1964:244-245.
17 On Butaritari and Makin, personal property includes all man-made objects
(except canoes), as well as livestock, food, and money. Unlike real prop-
erty, these items are always owned by individuals. A gift of personal prop-
erty does not establish a new social relationship.
18 Since a child was affiliated with the descent groups of both parents, it was
possible for someone to belong to the aristocratic ramage on one estate
and to the commoner ramage on another. It is probable that an indi-
vidual's social standing was mainly determined by his rank at his place of
residence. The problems posed by the overlapping memberships of dif-
ferent status levels were resolved by periodic wars over the succession to
the high chieftainship. The victor limited aristocratic rank to his children,
his siblings, and his siblings' children (Lambert n.d.).
19 Cf. Parkinson's general statement about stratified societies in the Gilberts:
"But there is no difference [between social classes] in everyday relation-
ships; lord and vassal, te tokker [te toka 'aristocrat'] and te bei [te bai
'commoner', 'slave') sometimes sleep on and under a single mat, together
drink de karuoruo ['sweet toddy'] and de mangina ['sour toddy'], or pre-
sent dances or games in common; there is no difference in dress either, but
still everyone knows how to make his rank count in a crisis" (Parkinson
1889:99-my translation).
20 Before the introduction of firearms, most Gilbertese battles were essen-
tially duels between heavily-armored champions. Aristocrats were con-
sidered better warriors than commoners because they had undergone a
long period of practical training and magical preparation.
21 I have elsewhere summarized the affiliations of the twenty ramages (four
aristocratic and sixteen commoner) which owned estates at Kiebu on
Makin Island around 1920 (Lambert 1966a:648).
J. L. Fischer
This chapter considers adoption on Ponape.
Present tense is used
to refer to the period of my fieldwork from 1950 to 1953. What
adoption may have been like in the period before major contact
with modern nations is not a major concern of this chapter, except
insofar as certain native beliefs and values may have persisted from
ancient times and thus may have been an important influence on
recent practice.
Ponape is one of five high island groups in the Caroline Islands,
an elongated cluster of islands stretching over a couple of thou-
sand miles and constituting the geographical core of the culture
area of Micronesia. The Ponapean language is related to a number
of other Micronesian languages, which form a branch of Eastern
Austronesian. The total land area is about 135 square miles. The
native Ponapean population in about 1830 is thought to have been
about 15,000, but it progressively declined to a low of about 2,000
by 1855. Following that it was rather stable, eventually showing a
gradual increase to about 6,000 by 195 0.
The climate is tropical oceanic with an average twenty-four-
hour temperature range from 75 F in the middle of the night to
85 F in the middle of the afternoon. Rainfall is abundant
throughout most of the year. Ponape is a fairly healthful place
compared to many other tropical areas. The principal endemic
diseases were yaws, a mild variety of filariasis, and several intes-
tinal parasites. A number of other widespread tropical diseases,
including malaria, were absent, as were the epidemic contagious
diseases-smallpox, measles, mumps, chickenpox-characteristic of
the Old World continental areas.
Subsistence economy on Ponape traditionally involved a num-
ber of common Oceanic crops. In order of importance they
were yams, breadfruit, bananas, coconut, and three kinds of taro.
Sweet potatoes, cassava, and a few other crops introduced after
foreign contact had also acquired some importance by 1950. Pre-
contact domestic animals were limited to the dog and the chicken,
both of which were occasionally eaten, although the major source
of protein was fish and shellfish. Pigs have been kept since the
nineteenth century and are now a popular feast food. The subsis-
tence economy has persisted, augmented by some wage labor and
production of copra.
At the time of first recorded contact Ponape was divided into
five independent petty states. Each of these had an elaborate hier-
archy of titleholders who had little real political power and who
functioned mainly in an elaborate system of feasts and etiquette.
The states survive today as the modern "municipalities" of the
island under the Trust Territory government. The feast and title
system has also persisted, with increasing emphasis on its social
aspects. The island has about twenty named exogamous matri-
lineal clans. Many of these are further subdivided into named sub-
dans, and many of these subclans contain a number of localized
ma trilineages.
The principal cultural agents of the major world powers have
been: American and other Western whalers and traders from about
1835 on;
American and other Protestant missionaries from about
1850 on; the Spanish colonial government from 1885 to 1899,
during which period Catholic missionaries became permanently
established; the German colonial government from 1899 to 1914;
the Japanese military and mandate government from 1914 to
1945; the American military and Trust Territory government from
1946 to date. These foreign agents have had a variety of goals, and
power to pursue these goals. On the whole, however, Ponapean
culture as a distinct entity has persisted through numerous changes
in foreign control. For example, the Ponapean language has per-
sisted as the mother tongue of all Ponapeans, matrilineages and
matrilineal exogamy have persisted, and much of the mythology
has been retained in spite of a general conversion to Christianity.
Moreover, the Ponapeans themselves have become aware that
there is a set of practices called 'Ponapean customs' tia ken
Po-nipei, which are distinct from the ways of the various 'people
from abroad' men wait, and which on the whole they have some
interest in preserving, at least as long as they remain a distinct socio-
cultural group without full opportunity for participating in a
Western-style life and economy. Ponapean custom in the sense of
practices and associated beliefs as identified by the Ponapeans
themselves is not strictly limited to survivals of precontact culture
as determined by careful ethnographic study, although in fact it
consists mostly of such survivals. Thus the ways for handling and
cutting up pigs at feasts are generally regarded as a part of
'Ponapean custom', although pigs were first brought to Ponape by
foreign ships in the nineteenth century. Pigs, however, have been
present on Ponape long enough to be regarded as part of the native
culture by most people.
Some Ponapean customs more than others have been the con-
cern of the various foreign cultural agents. Adoption has been
a matter of concern both for missionaries and for colonial adminis-
trators, and may well have changed in response to acculturative
pressures. It becomes difficult to be sure, therefore, how repre-
sentative recent practices of adoption under Ponapean custom are
of the actual precontact custom. However, the current concepts of
Ponapean custom are in any case highly relevant to current adop-
tion practice.
In assessing the importance of the various foreign cultural
agents, we must note that all adult Ponapeans at the time of my
fieldwork had been exposed to several of these, and that the
power positions of these agents had varied at different times. All
adult Ponapeans had grown up under some form of colonial
administration, and the various colonial powers had shifted their
policies from time to time corresponding to changes of key
personnel or to shifts in home-country politics. In the brief
American period there had been by 1953 a shift from naval "mili-
tary government" to naval "civil administration" under United
Nations trusteeship to administration under the Department of the
Interior. Each of these shifts involved important changes in the
policies and procedures of the territorial government, which ef-
fectively exercised ultimate political power. The concept of a
body of Ponapean custom, therefore, served to give life a certain
stability and continuity which it would have lacked if the people
had been oriented to follow closely the various changes proposed
by the foreign cultural agents.
Ponapean custom is said to include a variety of fosterage arrange-
ments. That discussed in this chapter and translated henceforth as
"adoption" is pwek-seri or pwekipwek, meaning literally 'lifting a
child' or 'being lifted'. This is the closest parallel to Euro-
American conceptions of adoption, but it is not identical with
them. Other arrangements for fostering children may be men-
tioned for contrast.
One type of fosterage involves children of deceased siblings of a
man or woman, especially a man's brother's children or a woman's
sister's children. These are already classified in the kinship termi-
nology as 'children' and it is considered that a sort of secondary
parent-child relationship exists from birth. This simply becomes
more strongly activated in case of natural parents' death.
A second type of foster child is the 'stepchild' serin peidiar
(perhaps meaning literally something like 'discovered child'), that
is, a spouse's child by a prior union. Such a child might be the
product of an earlier marriage or of a premarital affair. If the step-
child were young, the new spouse might adopt him but would not
be bound to do so, and would, I think, be fairly likely not to. This
is because one of the main reasons for marriage is to produce chil-
dren, and one of the main reasons for adoption in the Ponapean
sense is to acquire a child when the marriage has failed to produce
children. A new spouse would want to see whether he or she could
have children by the partner before adopting the partner's child by
some other person.
Occasionally a man may marry a woman who is already known
to be pregnant by another man, or a married woman will conceive
a child by a man other than her husband. Such children when born
during the marriage are regarded in Ponapean custom as legitimate
children of the woman and her husband. The husband may in fact
resent the adultery or premarital relationship, but as long as he is
married to the woman when she gives birth he has no customary
right to refuse to recognize the child as his own, and he is not
supposed to take out resentment on the child. In the traditional
kinship terminology a woman's husband and his real and classifica-
tory brothers are all her husbands or 'spouses' pwoud; the same
applies to her sister's husbands. Similarly, a man's wife's sisters
and his brothers' wives are his 'spouses'. Any man and wife in a
classificatory spouse relationship had secondary sexual rights in
each other, although it was considered better form for a man to
have sexual relations with his elder brother's wife and a woman
with her elder sister's husband instead of the reverse age relations.
Many of the relationships which Euro-Americans and many
modern Ponapeans would consider adulterous would not have
been so considered in the past, but even children conceived in a
disapproved extramarital relationship were considered the legiti-
mate children of both spouses if accepted by the husband, as
shown by his failure to divorce his wife before the time of birth.
Generally, overt jealousy about one's wife is considered 'unmanly',
although the feeling is common and sometimes gives rise to fights.
Probably foreign influence has increased open expression of jeal-
ousy, although it undoubtedly did not originate the feeling.
A further foster child relationship is one which has no particular
Ponapean name, as far as I know, but which might simply be
labeled "foster child." This would include those orphaned children
of remote and junior relatives or of immigrants without local
family ties who are sometimes cared for by wealthy families with-
out becoming full-fledged adopted children. These people when
they become adults would be regarded as 'servants or retainers'
Another type of child-adult relation somewhat suggesting adop-
tion occurs when an older couple or person growing too weak for
heavy physical labor asks for a young child to come and live with
them and help in work around the house. This child would usually
be a relative and might be a classificatory child, but he would be
old enough to be of some use and would require little care other
than verbal guidance by the old people. Such a child would not be
regarded as adopted but, if he or she stayed with the old person
until death, would normally be given some share of the inheritance
as a reward. He might also receive payment for his services to the
older person in the form of special instruction in esoteric skills and
lore-house or canoe carpentry, medicine, agricultural and other
magic, clan mythology (even if the clans differed), etc. Such a
child would retain a closer connection with his true parents than
an adopted child would, and he might well live quite near them.
I encountered two cases in which high-ranking chiefs were said
to have 'adopted' an elder sister's first son, i.e., their first-ranking
heir in the matrilineage. Under the provisions of the German land
laws as interpreted by the Japanese, this meant that these sororal
nephews, as officially adopted first sons, were in line to inherit the
chiefs land. It seems probable that, traditionally, avunculocal
residence was practiced more by the prospective heirs of the heads
of the higher-ranking matrilineages. According to tradition there
were certain plots of land that were permanently associated with
certain political titles, which were again the property of certain
high-ranking lineages. Very possibly the shift to avunculocal resi-
dence usually took place at or after adolescence for a high-ranking
heir, and the adoption in these two cases may have been partly a
special device for insuring the customary matrilineal inheritance of
land and title under the reformed colonial land laws. In both
instances neither chief had a true legitimate son, however, so he
would have had some reason to adopt a child in any case. Further
information on this point would be desirable. The link between
property and titles has now largely lapsed.
One 'high-ranking chief' nan mwareki reported that he had sent
his young son to his parents' house, and that it was a Ponapean
custom for children born to a high chief while he held office to be
turned over to the chiefs parents for care, providing the parents
were still living. Apparently this would not be regarded as a case of
adoption, and there would be a good chance that the grandparents
might die after a while, in which event the child would return to
his true parents' house. In general, children rank in order of birth,
but the later child of a high chief, born after the father has
achieved his full rank, is regarded as having special prestige and at
least for some purposes (e.g., title inheritance of his own clan
titles) outranks his older siblings, born while the father was in a
lower rank. While title inheritance is basically matrilineal, there is
a dual system of 'lines' of high titles, and the highest-ranking
chief of the first line is supposed to take a wife from the matri-
lineage holding the top ranks in the second line of titles. This
makes it possible for a high chief in one line of titles to be the
father of the high chief in the other line, and the grandfather of an
heir to his own title back in the first line. Such an alternation of
generations between the two lines of titles in the patrilineal line is
considered desirable by the Ponapeans. The fosterage arrangement
in which the high chiefs highest-ranking child is sent to the chiefs
parents' home would enable the child to receive instruction in the
duties of the position in the other line to which he would one day
probably succeed.
In the ideal customary sense, an 'adopted child' serin pwekipwek
on Ponape is a child who is given by his true parents (biologically
and socially) to another couple, with the understanding that this
foster couple will raise the child and provide for him as if he were
their own, and that in return the child will owe them the services,
including care in old age, which a child would normally give to his
true parents. No special ceremony appears to be required. The
child is adopted as soon as his two sets of parents have explicitly
discussed the matter and the child has been transferred. Giving a
child for adoption is by tradition wholly in the power of the par
ents, or in the power of the mother for illegitimate children born
out of wedlock.
It is considered best if the child is adopted at an early age, as
soon as the child can be weaned, or before weaning if a wet nurse
happens to be available. Weaning might be hastened somewhat to
facilitate the transfer. I was told recently by a Ponapean Peace
Corps instructor that the transfer often occurs when the child is a
few months old. Possibly earlier transfers are now facilitated by
the availability of canned milk and bottles. It is thought that chit
dren adopted after infancy will be more likely to have a poor rela
tionship with their adoptive parents, although there is some opinion
that relations with adopted children are likely to be difficult in
any case. Some informants say that children taken at a later age
are "not really adopted children" in the Ponapean sense, but this
is a matter of some disagreement. Certainly most cases that
claimed to be true adoption involved a transfer of the child
immediately after weaning.
Most often the adoptive parents do not have children of their
own at the time of adoption, although occasionally they may have
another adopted child or even a true child, and sometimes a
couple without true children will adopt several children. The adop-
tive parents are usually but not always somewhat senior to the
true parents. If so they are probably in a position to command
some deference from them, as well as to help them if they choose.
Most commonly one of the true parents is a younger sibling of one
of the foster parents. Other close relationships recorded by my
wife or myself include adoption of a child by a true parent's
younger sibling (mother's younger sister), by grandparents, and by
own older sibling (sister), though these seem infrequent. Birth
order is important on Ponape. Older siblings are supposed to have
authority over younger ones and are generally given a larger share
of the family inheritance. A childless older sibling is in a strong
position, therefore, to request a child for adoption from a younger
sibling, although the older sibling does not have any absolute right
to "requisition" the infant, and there is no customary explicit
sanction which he can exercise to enforce his request. Even first-
born children may be requested and given for adoption.
Ponapeans commonly say that people prefer to adopt boys
rather than girls, because a boy will stay with his foster parents
and help them when he grows up, while a girl will marry out. Ac-
tually, when the family members are in good health a fair number
of households are nuclear family households.
Important people,
such as chiefs and rich men, are likely to have some younger
people, including younger married relatives, living with them. A
married child is also likely to live with a feeble old couple. A resi-
dence census which I took in three parts of the island showed con-
siderable variation, but indicated a preponderance of married sons
over married daughters living with parents. However, everyone
n:cognizes the possibility that if a couple has only daughters they
may have a married daughter come and live with them and inherit
the land.
Moreover, if prospective foster parents would prefer to
adopt a boy, by the same token the true parents would be more
reluctant to provide a boy for adoption, so that many childless
couples would be likely to take a girl instead. Certainly girls as
well as boys are adopted in fair numbers, although precise statis-
tics on this point are lacking.
This preference for boys seemed strange to me in view of the
importance of matrilineal descent in the society, so I asked a
number of old informants whether people had always had this
preference or whether it was a recent development in response to
the attempts of the German and Japanese governments to institute
strict patrilineal inheritance of land. In reply to my leading ques-
tion, these informants generally cited an Oedipus type myth about
the conqueror of Ponape, Iso-kelekel, who is said to have insti-
tuted the precontact political order by deposing a tyrannical ruler
of the whole island. Supposedly Iso-kelekel ordered that his own
son should be killed at birth. His wife instead gave the child to a
relative, who secretly raised him. This was cited as an instance of
someone in olden times who did not want a son. However, even
this is not such a convincing example, since the boy, who inciden-
tally grew up to have an incestuous relationship with his father's
sister, was forgiven by his father, although he repeatedly provoked
his father, and ended up by receiving honors and a share of his
father's kingdom. None of these informants cited any statements
about a preference in the recent past for daughters. Moreover,
there are a number of customary privileges which boys have with
respect to high-ranking fathers which are not shared by the matri-
lineal political heirs of these same men. These suggest a kind of
institutionalization of the preference for sons and an indulgent
father-son relationship.
As noted above, the primary relationship of the true parents to
the adopting parents may be through husband or wife on either
side. While I do not have a large number of cases recorded in
de tail, it is my impression that all combinations of linkage
between true and foster parents are fairly common. Note that
according to Ponapean custom, while the child belongs to his
mother's matrilineage and clan, the father has custody rights in
case of divorce, and he usually exercises them. In other words, the
rights of the married father over his children are clearly recognized
by custom, although postwar American influence had begun to
change this by 1953.
The initiative in an adoption generally comes from the adopting
parents. Few people actively seek out foster parents for their
child; rather they give the infant up under pressure, with mixed
feelings. Even people who are poor and have many children and
little land do not generally take much initiative in putting up their
children for adoption, although such people are especially likely
to receive requests for their children. A few young married women
and some women with illegitimate children are said to 'throw away
their children', i.e., readily agree to the adoption of their children.
The true parents retain an interest in their child after adoption.
The child is expected to know the identity of his true parents and
the true parents continue to see their child after he has been
adopted. If the true parents feel that their child is not being
treated properly they have the right to take back their child, al-
though in fact this is rarely done, since it would imply severe criti-
cism of the foster parents, one of whom is usually a senior relative
of one of the true parents. If the true parents wish, they may leave
some inheritance for any of their children who have been adopted
out, although they are not under much pressure to do so. If the
true parents have no other living children at the time of their
death, their child who has been adopted out will automatically
have some claim on the inheritance of his true parents.
Clan membership is said to be unaffected by adoption. The
child remains a member of the clan of his true mother. Fairly
often, of course, if the adoptive mother is a sister or matrilineal
cousin, she will be of the same clan as his true mother. It is
not, however, out of the ordinary for the adopted child to be of a
different clan from that of both adoptive parents. By rule of clan
exogamy, true children are always of a different clan from their
fathers, and fathers are likely to retain custody of their children in
the event of divorce; among the common people divorce has
traditionally been frequent. This means that nuclear families con-
taining stepchildren of another clan from that of the father and
stepmother have been common.
Ponapeans, in talking about adoption, tend to phrase the re-
lation between adoptive parents and adopted children in terms of
'matrilineages' kainek (Lieber, personal communication). This is
more a manner of speaking about kin relationships than a require-
ment of lineage identity between a child and his "new" mother or
father. Thus if there were two brothers married to women of dif-
ferent lineages, and one brother adopted the other's child (a rea-
sonably common case), many Ponapeans would no doubt explain
this as an example of the operation of the matrilineage principle,
even though the child belonged to a different lineage from either
of his adoptive parents. In this case, of course, the true father and
the adoptive father would belong to a single matrilineage. More-
over, the adopted child retains some connection with his true
mother, and through her to his original matrilineage and clan.
Before the imposition of foreign colonial rule, the matrilineages
were units for purposes of vengeance in feuds. The ties by adop-
tion between matrilineages in a community, which usually rein-
forced preexisting ties by marriage, probably helped somewhat to
maintain peace in the community.
In spite of the fact that adopting parents are most often child-
less when they adopt a child, and, if so, promise that the adopted
child will be treated as their true child, they sometimes later give
birth to a child of their own. It is recognized that they are likely
to prefer the true child and to favor him in inheritance, and many
Ponapeans believe that this is proper. Ponapeans allow for some
difference in treatment even of true children according to the
parents' preferences, in spite of the general rule that children
should rank in order of birth for inheritance. This is functional in
terms of the importance of personal achievement in the political
system: aged parents will best be taken care of by a successful
child, so they may wish to throw their support and resources
behind a bright and hard-working later child instead of a dull and
lazy first child.
If a couple has an adopted child and later has a true child it is
likely that they will succeed in establishing a better relationship
with the true child, who has not had the trauma of changing
parents in infancy and who benefits by the greater experience in
child-rearing his parents have gained by the time he is born. Ac-
cordingly disputes between true and adopted children are likely to
arise about the family property after the death of the foster
parents. During the Japanese period, governmental decision was
apparently based on whether the adoption had been officially
registered by the local Japanese policemaster or not. If registered,
the adopted child had the full rights of a true child of the same
birth order. By 1953 the American administration had not estab-
lished a clear set of precedents for handling disputes of this sort,
although later, in 1957, a law passed by the Fifth Ponape Congress
and promulgated by the District Administrator for Ponape Island
specified: "For purposes of hereditary succession, an adopted
child shall be considered as younger than any natural child, except
that the oldest adopted son, older than real son, shall be entitled
to one-third of the value of the estate .... " Thus the adopted
child's somewhat disadvantaged position was formally ratified.
It can be argued that the disadvantage of the adopted child is
not as great as it may seem. Although firstborn children are some-
times given for adoption, it is much more common for younger
children to be given. These children would stand to inherit little
from their true parents anyway, in view of the practice of primo-
geniture, which fits in both with Ponapean tradition of emphasiz-
ing birth order and with the land laws of the colonial governments.
Moreover, the true parents of children given for adoption are often
junior in their own family and will then probably have little land
to give even to their eldest son. There is probably a better than
fifty-fifty chance that the adopting parents will not have a true
child later, in which event the adopted child will probably inherit
as if he were a true child of his foster parents, except that he will
not have the same position in the mother's matrilineage.
As far as is known, neither the local Protestant church nor the
local Catholic church have given ritual recognition to adoption; in
fact, they have made some attempts to discourage it in the normal
case where the mother or the true parents are alive and in good
health. Some missionaries have seen a connection, correctly, I
think, between the frequency of adoption and the emphasis on
sibling ties as opposed to marital ties, and they regard adoption as
a renunciation of the responsibility and importance of marriage on
the part of the true parents. Indeed, it seems plausible that one
unstated reason why young couples will sometimes give up one or
more children for adoption at the beginning of their marriage is
that they are not getting along too well together and are wonder-
ing whether they should be divorced. Even if they expect to re-
main together, perhaps they find the added responsibility of a
child something of a burden. Adoption, then, may sometimes
bl.!come a device for prolonging a sort of premarital state for the
true parents; again, one might think of it as a means of gradually
sliding into the responsibilities of marriage.
In reviewing fieldnotes on land disputes involving adoption, I
have noticed that informants speak often of either the husband or
the wife as the adoptive parent. This suggests that perhaps in cases
of divorce an adopted child might remain in the custody of the
parent initiating the adoption, who would usually be the parent
related to the child most closely by blood. I have little informa-
tion on the fate of adopted children in the event of the divorce of
the foster parents, but it deserves investigation.
This manner of speaking also suggests that perhaps occasionally
an unmarried person might adopt a child, or one spouse might
adopt a child without the other spouse's participation in the adop-
tion. I suspect that the latter is unlikely, but that the former may
occur occasionally. I recall one man from the low island of Ngatik
who said he had adopted a son before marrying. This island is cul-
turally related to Ponape but is nevertheless regarded as a distinct
society with its own set of customs. Probably a married couple
with a child would refuse to yield it for adoption to anyone ex-
cept another couple who showed considerable interest, but there
are a fair number of illegitimate children born, and unwed mothers
of these children may be less anxious about keeping them, since
they would be something of a deterrent to marriage.
We will try to summarize the functions of adoption on Ponape as
suggested by a comparison with other tribal societies of Oceania
and elsewhere. One of the factors which seems to be conducive to
the development of widespread adoption on Ponape and in many
other Oceanic societies is the small, insulated social arena with a
stable climate. This has several consequences. First, it means that
serious disturbances from outside invaders are rare and that the
inhabitants have the opportunity to work out stable sociopolitical
arrangements which are suited to their technological level and
local resources. Secondly, an optimum population level is reached
in a few generations which must somehow be maintained as an
upper limit as a part of the socio political arrangements. Thirdly,
given a rather dense population with dependence on some form of
horticulture, there is a need for preventing the grossly uneven
accumulation of inherited land rights through the working of ran-
dom irregularities in genealogies. This problem of random accumu-
lation of land rights becomes especially severe if childless couples
must always return their own land rights to some collateral rela-
tive, or if there is some preference for one sex as holder and re-
ceiver of land rights, so that couples with children of the wrong
sex are restricted from passing on rights to their children.
The slow recovery of the population of Ponape following the
decrease in the first half of the nineteenth century, after the intro-
duction of contagious diseases from the whalers and other foreign
sources, suggests that the cultural and ecological adjustment of the
society imposed some kinds of controls on fertility which have
persisted until recent times. What these were we do not know
exactly, but we can be sure that Ponape and most other small,
isolated societies in Oceania had such controls of one sort or an-
Given some system of fairly effective controls on fertility,
one would expect that through random operation of these con-
trols a fair number of couples would be left with no children and
others might have a fairly large number of children. This is espe-
cially true if infanticide, abortion, or conscious efforts at contra-
ception are not a part of the system, for if they were a couple
might be able to decide to have a small number of children and
stop. I have no extensive statistics on a large number of couples,
but in reviewing fieldnotes I have noticed a number of couples who
were childless or who had an only child, and others with as many
as twelve children. Hambruch and Eilers report, as did our own in-
formants, that infanticide was not practiced, even for defective
infants ( 1936:87). Contraception and techniques of coitus inter-
ruptus are not reported and apparently were not used tradition-
ally. Techniques of abortion were known but disapproved, and
probably used mainly by a few unmarried girls. On the whole, the
ethnographic data sustain the picture of a population in which
fertility was controlled ecologically rather than consciously, in
which the fertility of couples was variable and randomly dis-
tributed, and in which children were generally valued.
Children are desired in Ponape for several reasons. The women,
especially girls of marriageable age, appear to take pleasure in
1:aring for young children, and volunteer for "baby-sitting" jobs
uround the community, carrying the babies around well after the
ugc when the child can walk, singing songs to them, and playing
with them.
When children get older, especially as they approach adoles-
cence, they are useful in helping with the work of daily living and
become increasingly useful as they reach adulthood. As the par-
ents reach old age and become too feeble to take care of them-
selves, it is expected that one or two of the married children will
remain with the parents. Some sort of arrangement with some
more distant relative can always be made if a couple has no chil-
dren, but it is considered much better to be cared for by children
whom one has raised and with whom one has a strong affectional
tie. Many foster parents say they adopt a child to have someone to
care for them in old age.
Ponapeans are deeply attached also to their native places and
have an affection for the land they have worked. This appears to
be a sentiment which antedates the German land reform, when
"private" ownership was supposed to have replaced ownership by
the chiefs. Even before that time, it appears that people normally
had effective control over the lands which they had planted and
cared for. Apart from this, the average Ponapean was mythologi-
cally attached to his home community: he knew its legendary
history and the sites of mythical events, and his kin relation to the
mythical actors in these events. He therefore wanted to have a
personal successor to live in the community after he was gone.
This type of sentiment of local attachment is one which requires
generations to build up to full strength, and is likely to be espe-
cially strong in a small island community with lengthy political
stability. It is also likely to be strong in a community dependent
on agriculture for subsistence. Another common explanation
which Ponapeans give for adoption is to "have someone to inherit
my land." This seems to be a way of expressing attachment to the
land and the local community. It is not thinking of the land as an
economic resource so much as thinking of it as an object of affec-
tion, as we might think of a family heirloom. This reason for adop-
tion is one which is strong and widespread in Oceania, but might
be weaker in societies with less settled political organization or
with less settled subsistence activities, such as herding societies.
We have considered some reasons why Ponapeans would have a
strong desire for children and why a fair number of couples might
be childless while others have a sizable number of children. We
must now consider the reasons why some parents would agree to
give their children for adoption to others. The case of the unmar-
ried mother may be easiest to understand. There has traditionally
been much sexual freedom for young people on Ponape, as on
many Pacific islands, and this seems to have had the effect of
reducing rather than increasing fertility. If an unmarried girl does
have a child, no great stigma is attached to her; but still the child
restricts her mobility and her freedom to have affairs, from which
marriage proposals normally eventuate. Moreover, when she does
get married it is likely that her premarital child will be in an in-
ferior position relative to her children by her husband. Therefore,
the unmarried mother may be willing to give the child for adop-
tion to a childless couple, who may have the prospect of providing
for him in a better fashion than she could expect to do by herself,
at the same time freeing herself from the responsibilities of child
care. Newlywed couples wishing to prolong the period of adoles-
cent freedom might also be fairly willing to give up a child for
adoption, especially if the couple were not yet well adjusted to
each other and were considering divorce. Many islands in Oceania
are noted for their relative sexual freedom, and, as has been sug-
gested above, this may have functioned to help population control
by reducing male fertility, as well as by its less disputable role in
the transmission of gonorrhea and other venereal diseases which re-
duce fertility.
Divorce is probably one important source of children for adop-
tion, and again may be related to the high value on sexual free-
dom: if one has freedom to pursue extramarital affairs, one may
sooner or later find a "better" spouse. When a man receives the
custody of the children in a divorce, he is likely in fact to turn for
help to his mother or a sister, who may adopt the children.
Another reason for the willingness of parents to give a child for
adoption is the relative strength of ties between blood kin com-
pared to affinal kin and the importance of seniority within the kin
group. While it is my opinion that the nuclear family is more
important on Ponape than it is in many Oceanic societies (cf.
Bascom 1965: 18), ties between parents and adult children, and
between adult siblings, remain rather strong, and the junior rela-
tive is expected to show deference to the senior. At the same time
a person's family does not consider the tie of a junior member to
his spouse to be so important, and older members of the family
would therefore not feel much reserve about asking a junior mem-
ber for a child to adopt. The emphasis on blood kin ties is wide-
spread in Oceanic societies as in many nonliterate societies. It is
related, among other things, to a weak occupational differenti-
ation and low occupational and geographical mobility, and to a
lack of development of law and impartial institutions for settling
disputes. This means that family members, even after marriage, are
likely to live close enough to each other to maintain contact, and
that the family members are available to each other to help as
needed in matters of defense and daily living. In such a situation
junior family members will be under considerable pressure to re-
spond to the requests of senior family members.
Ponapeans who give their children for adoption are often think-
ing to some extent of the welfare of the child to be adopted as
well as the remaining children, if any. Adopting parents often use
the inducement that they will be able to give the child more land
than the true parents can. If one of the adopting parents is a senior
sibling to one of the true parents, he will probably have received a
larger inheritance of land from his parents to begin with. And if
the adopting parents have no children, or only one or two chil-
dren, while the true parents have several or many children, the
inequality of inheritance becomes compounded. While the German
land reform prescribed patrilineal primogeniture for the inheri
tance of land, in fact it seems to have long been a Ponapean ideal
that all children should be provided with some land of their own,
and in practice this has been largely achieved in one fashion or
another. A child with inadequate land or without land at all would
not starve, since he would be able to live on a relative's land and
work on it and receive the necessities of life. However, his position
as a dependent would not be favorable, since he would not have
much in the way of luxuries and would not be in a position to
compete in the feast and title system. A couple with many chil
dren and little land for them who refused to give out some of the
children for adoption would be in the position of economically
hampering all of their children. Traditionally, and even at present,
there was practically no way of making a living without land. Most
specialist occupations were not and are not full time. Moreover,
land was not generally available for purchase.
One of the features of adoption in Ponape, which seems a little
unusual in Oceania generally, is the emphasis on infant adoption.
Another somewhat unusual feature is the reported retention of
original clan and lineage membership by the adopted child. Both
of these features may be related to a somewhat greater emphasis
on the nuclear family and on individual achievement than is found
in many Oceanic societies. The Ponapean settlement pattern con-
sists of isolated small farmsteads; there are no true village clusters,
although there may be a sort of hamlet or extended family cluster
around the house of an important chief. There is a fair amount of
competition between men in the feast and title system. Their work
is evaluated by the senior chiefs, who are influenced in part by
considerations of hereditary rank but are also supposed to take
into consideration industriousness and public service. In this sys-
tem a wife's status depends much on her husband's, and the hus-
band himself is in competition with all other men to some extent,
including even his own brothers and other close kin, although
there are devices limiting the competition within the kin group.
One of the devices for limiting competition within the lineage is
the strict hierarchical order of hereditary rank according to own
birth order and mother's birth order. The members of the lineage
have an interest in supporting each other's political advancement
in the community, but they can effectively cooperate only to the
extent that the order of internal hereditary rank is not violated:
that is, a junior person in the lineage should not seek a political
title higher than that held by a senior member, without the latter's
full approval.
Presumably members of the lineage are jealous of
their own rank and their children's rank (in the case of the
women), and would resent the introduction of new persons of
rank higher than theirs into the lineage. Ultimate authority over
lineage matters would generally be in the hands of the senior male.
If, hypothetically, a woman asked her lineage chief to approve of
her adopted child as a lineage member, such an action would pre-
sumably offend other lineage mothers, junior to her (especially
those just beneath her in rank), and their children. This difficulty
probably explains why the acceptance of adopted children into
their foster mother's lineage does not occur.
It might be thought that the elaborate system of political titles
would require that lineages could adopt new members to hold
titles in the event of lack of a suitable heir, but this does not
appear to be crucial. While there is a tendency toward matrilineal
inheritance of titles, this is not so important except for the highest
titles. The highest titles are generally held by the largest clans. The
subclans within the clans, the lineages within the subclans, and
individuals within the lineages are all carefully ranked by recent or
legendary birth order, so for these high titles there is a long list of
potential heirs which is not easily exhausted. For most political
titles there is no feeling that the title must absolutely stay in one
matrilineal group, since ultimate authority to assign the title rests
with the highest chief. Even if the list of matrilineal heirs for one
of the higher titles should become exhausted, it would be easy to
choose another successor on the basis of merit and seniority in the
line of titles leading up to the vacant title; for each high title has a
ranked series of lower titles leading up to it through which the
matrilineal heirs to the highest title are supposed to pass. The
actual holders of the titles in each ranked series may come from
different clans, although fewer clans are represented in the higher
ranking titles than in the lower-ranking titles.
Returning to the question of emphasis on infant adoption, an-
other reason for this may involve the settlement pattern of rela-
tively small families on isolated farmsteads. With small household
size, children become important as companions for their parents.
At the same time, to introduce a new member into such a small
group can be more disturbing for it than for a large extended
family. To bring in an older child would be to bring in someone
who had already learned a distinct microculture in another house
hold, but a baby would be less of a problem, since he would be
able to learn the microculture of his foster parents as he grew up.
One hypothesis for comparative study might be that in societies
with larger households, where the burden of socialization can be
spread over a number of older members, adoption of older chil-
dren and young adults would tend to be more popular, since the
microcultural adjustment of the new members could be more
easily managed.
In this chapter we have described recent practices of adoption
on Ponape from a functional and ecological viewpoint. While fur-
ther field data would be desirable, it appears clear that adoption
serves important socioeconomic functions in bringing about a
rough redistribution of land in each generation, and in providing a
form of personal care for the aged. These are functions of adop-
tion widespread in Oceania.
Adoption on Ponape generally is limited to infants and is not
especially important as means of calculated self-advancement on
the part of either the foster parents or the adoptee. The adoption
is a transaction which mainly concerns the true parents, foster
parents, and adopted child, and does not especially concern the
matrilineages of any of these. Preservation of lineage continuity as
a motive for or function of adoption is not reported.
A number of possible explanations for the special characteristics
of Ponapean adoption were offered. Hopefully, comparison with
reports of adoption elsewhere in the Pacific will enable further
testing of these and alternate explanations.
Most of the data for this chapter were obtained during my employment on
Ponape by the U.S. Administration of the Trust Territory of the Pacific
Islands, first as District Anthropologist and later as Assistant District Ad-
ministrator. The data were reviewed and the chapter written while I held a
Special Fellowship of the U.S. Public Health Service, National Institute of
Mental Health, at the Center for Advanced Study in the Behavioral
Sciences, Stanford, California, during the year 1965-1966. I wish to ex-
press my gratitude for support received from all these sources. Some eth-
nographic reports of special relevance to the question of adoption and
Ponapean social structure include the following: Bascom 1965; Fischer
1957a, 1957b, 1958, 1964; Hambruch and Eilers 1934; lmanishi 1944;
Riesenberg 1949.
2 This last figure excludes immigrants from other islands and their descen-
dants who have settled in distinct colonies in certain parts of the island,
but includes immigrants who have become integrated into Ponapean
3 A disastrous smallpox epidemic in 1854 introduced by an abandoned crew
member from a whaler killed off about half the population, but smallpox
then disappeared and so far has not returned.
4 The number of whalers declined after about 1870.
5 For example, a residence census in August, 1950, of Depek Island, one
of the sections of the state (currently "municipality") of U , shows six
couples living in households with no other couple, twelve couples living in
households with one other couple (i.e., six such households), and five
couples living together in a single large household.
(1 For part of the Japanese period it was considered that the son-in-law had
to be adopted in Japanese style and officially inherit the land for his wife,
but this is no longer true.
7 Both Catholic and Protestant missionaries, but especially the Catholic,
have attempted to discourage divorce. Traditionally, marriage among the
Ponapean nobility was marked by ceremony and was considered stable.
However, the nobles could practice polygyny. The common people, who
were permitted only one wife at a time, were married simply by publicly
living together and declaring themselves married. Many marriages begin in
this fashion today and are confirmed in the church only if and after they
have been thoroughly tested in practice. Some other church marriages end
in divorce and remarriage in the eyes of the community, but are reconsti-
tuted briefly when one of the parties is believed to be on his deathbed. To
have two or three marriages in youth is rather common for Ponapeans.
Divorce is less common after the birth of children, but by no means un-
known. Some people have had many spouses in succession, but such
people are not typical and are regarded as a little strange by others.
8 The Ponapeans do not attach much shame to a mother of illegitimate chil-
dren and perhaps even value somewhat the proof that the woman is fertile,
but a man marrying her later would not necessarily want to take responsi-
bility for an illegitimate child born well before his marriage. Also, while
virginity is not expected in a wife, general promiscuity in a woman is not
considered desirable, and several illegitimate children around the house
could be taken as a sign of promiscuity.
9 The disease of filariasis may have played an important part, since it is well
suited to working in a feedback system. Overt symptoms of filariasis are
usually slow to develop, and usually occur only after repeated infection
over a period of many years. The disease is not generally fatal but can
eventually result in male sterility. The ftlarial parasite must return to a
mosquito to go through its full reproductive cycle, so the extent of pathol-
ogy depends on the number of bites by infected mosquitoes, which in turn
depends on the availability of infected people for the mosquitoes to bite,
and on the strength of their infection. Obviously, in a densely settled area
filariasis can spread more easily and the infection can be more intense in
individuals and develop more rapidly. At the same time, widespread and
heavy infection leads to widespread reduction of male fertility and a o ~
pensating decline in birthrate, until the population is thinned out enough
to permit a decline in rate and intensity of infection.
Intestinal parasites may have a similar relation with population density
and crowded living conditions. Winton Boyd, a Navy physician, reported
in 1951 that of 258 pregnant women examined over a two year period 204
(79 percent) had evidence of one or more kinds of worms in stool sample1,
Perhaps the proportion infected would be even higher for women of the
same age who had difficulty in conception, since heavy worm infestation ii
reported to suppress or delay the menstrual cycle.
Diet, especially protein level, may have worked in a feedback system
controlling population density. The principal source of protein on Ponape
is fish, and the ratio of accessible lagoon to land area is less favorable than
on most coral atolls in the region. The protein level in the diet has prob-
ably been rather poor compared with most other islands in Micronesia.
The supply of proteins would have been worse in times of greater popula-
tion, when the area of settlement was pushed farther inland, and when
there was greater pressure on the accessible fishing grounds. The effect of
dietary deficiencies may be aggravated by chronic infection with yaws,
which was found in most of the population in latent form until controlled
by widespread treatment with penicillin after World War II.
Patterns of sexual intercourse may have contributed to male infertility.
Daily intercourse is reportedly typical of many married couples and prob-
ably of many unmarried young people as well. Some authorities state that
this frequency will reduce male fertility by failing to allow the sperm time
to mature sufficiently.
The ritual operation of removing one testicle from all adolescent males
may have resulted in infections and reduced fertility in a number of men,
although this conjecture can no longer be tested since the practice was
stopped in the early part of this century. Incidentally, the operation itself
is known elsewhere in the Carolines as a therapy for testicles excessively
painful or swollen from filariasis. I suspect it was at one time performed
solely for this purpose on Ponape; if so, it may have become ritualized for
adolescents in a period of denser population when filariasis was a more
serious health problem than it has been recently.
10 It is uncertain whether gonorrhea was present on Ponape before contact
with whites, but it has almost certainly been present since the first half of
the nineteenth century. Perhaps before then, other types of infections of
the female genital tract were transmitted by sexual contact.
11 If a higher chief outside the lineage offers a junior member a high title, the
latter will accept the title; but he will not be in a good position to request
help from his lineage members in making the title-payment feast and in
fulfilling other obligations if they feel that in seeking the title he has been
too ambitious, to the disadvantage of senior lineage members.
Ruth Gallagher Goodenough
Adoption is a common practice on the Truk island of Romon um.
A census that covered three generations of recorded births, num-
bering 520 individuals, revealed 57 verified cases of adoption. This
represents a crude adoption rate (the number of adopted children
as a percentage of the total population)
of 10.9 percent for this
particular population. A survey of present-day households dis-
closed that two out of three households with children have at least
one adopted person in them.
Older people assert that adoption was not always so common,
and the age distribution within the 57 cases of our sample appears
to bear these informants out: 47 of the 57 adopted individuals
were under thirty years of age.
When pressed to explain why
adoption should be more common now than formerly, one old
woman said that it is because people are nicer now. In her view,
their Christian training was making them more loving and helpful
to one another. There is no question that a commonly given rea-
son for adoption is the desire to help someone with a large family,
but a closer look at the data suggests that a skewed fertility rate
may be more relevant than any improvement in character or
As Table 19 shows, children are unevenly distributed among the
women of childbearing age. Of the 175 women in the three senior
TABLE 19 Fertility of Women on Romonum
No. of Named No.of
Children Borne Women* %
0 56 32
1 40 23
2 33 19
3 or more 46 26
TOTAL 175 100
*All married women on Romonum in the last three generations.
generations, over half had either no children or only one child,
while hardly more than a quarter of the women had three or more
children. It is clear that some women have had many children. This
helps to account for an overall population increase of 32 percent
between 1947 and 1964. For a study of adoption however, it is the
one-third who have had no children, or the half who have had at
most one child, who are most significant.
It is interesting to speculate on why so many women in this
group appear to be infertile. All but one of the women in Table 19
had been married at least once; birth control, infanticide, and
abortion are not practiced by the Trukese; and chastity among
women is not highly valued. There is nothing in the obvious social
practices of the Trukese to account for the high rate of childless-
ness. We are led to look at other factors in the situation, notably
those having to do with public health.
In chapter 11, John Fischer discusses some of the physical
causes of diminished fertility. The conditions he cites for Ponape
might apply equally to Romonum. Filariasis may contribute to
male sterility, particularly in later adulthood. Dietary deficiencies,
especially those endured during wartime occupation may well have
lowered fertility, and intestinal worms may be a continuing factor.
However, one suspects that the particularly high rate of childless-
ness in relatively recent times is more likely to have been the result
of a higher instance of venereal disease.
Venereal infections are
more apt to result in sterility for some women, leaving other, non-
infected women free to bear as many live children as the limiting
factors of diet and the like permit.
We may conjecture that successive civilian and military occupa-
tions on Truk introduced,
or at least helped to spread, venereal
infections, which in turn reduced the number of live births and
rendered a number of men and women infertile. If this is true, we
might expect that public health measures after World War II, in-
cluding mass penicillin inoculations against yaws, would have elim-
inated much of the venereal infection from the islands. Thus the
generation that has just come into childbearing age should show an
increase in the number of women who bear live children.
It seems apparent that the high incidence of adoption in recent
times is associated with this skewed fertility distribution. Where
some women have many children and others have none, adoption
is a way of redressing the imbalance. If this is indeed the situation,
one might predict that the rate of adoption will fall off appreci-
ably in the immediate future, back to a level that reflects a more
balanced fertility picture.
This last point rests on the observation that there are apparently
no strongly operating factors to sustain such a high rate of adop-
tion other than childlessness or the occasional orphaning of chil-
dren. Informants themselves give only three reasons for adoption.
The first is to provide parents for a child who has lost one or both
parents through death or incapacitating illness. In our sample this
reason accounts for 13 of the 5 7 cases. Of the remaining 44, which
we may think of as the group of "voluntary adoptions," the
second reason, attributed to exactly one-half, was that the adopt-
ing parents wanted to help a relative who had too many children;
the reason attributed to the other half was simply that the adopt-
ing parents wanted a child. No other reasons were suggested and,
moreover, whichever of the three reasons was stated in explana-
tion of a particular case, as the following table shows, the over-
whelming number of adopting couples in all groups were in fact
No. of Children
in Adopting Family
No. of Families
2 or more
When the adoption practices of Romonum are compared with
those of societies in which adoption is clearly an important part of
a larger social game, the lack of institutionally elaborated signifi-
cance becomes quite apparent. Considerations of social rank and
political advantage in adoption are minimal on Romonum. Where
social leverage enters as a factor, it is used when asking for a child:
one tries to ask someone who would find it hard to refuse. A male
informant stated that he might ask anyone for a child in adoption
but that some people would find it more difficult than others to
refuse the request. Those who would find it most difficult to
refuse would be his own siblings or a close personal friend with
whom he had a strong 'brother' relationship. Next in order would
be the members of his own matrilineal lineage. The third group in
this descending scale would be the close members of his father's
lineage (father's brother, father's sister, or father's sister's child),
and below these he placed the more remote members of his
father's lineage, along with the children of the men of his own
lineage. He would have much less leverage with more distant mem-
bers of his matrilineal sib, who presumably would find it easier to
refuse his request.
Adoption is not used by the adopting parents as a pretext for
extracting favors from the real, or natural, parents, nor, except for
one strongly disapproved instance, were there reports of the real
parents trying to wangle subsequent favors from the adopting
parents. Other differences are equally on the side of minimizing
the institutional significance of adoption. Grandparents do not
adopt grandchildren in this society. Children are not exchanged by
families to ensure that mutual obligations are lived up to. Adop-
tions are not used to strengthen weakening social bonds caused by
absence or indifference. Donor parents do not give an accompany-
ing piece of land or goods with their child to establish for him a
social base independent of the adopting parents. Finally, for what
it may be worth as negative evidence, adoption has not been elabo-
rated as a significant theme in Romonum's folk tales.
In general, Romonum lacks a stratified social system with
highly formalized rules and observances for changes of rank or
status. This is reflected in adoption. A very few cases of adoption
seem to have been deliberately calculated to keep land holdings
within a sib; in two instances the highest-ranking male child of a
lineage was adopted by the last surviving male of a dying, related
lineage of the same sib. But such considerations are unusual in our
data. In the absence of any other significant rationale, we may be
permitted to look at the high rate of adoption in recent times
simply as the application of a presumably traditional social form
to meet a recently augmented social need.
The term for adoption on Romonum, mwuumw[L, means 'breaking
apart' or 'breaking the tie'. It is used to describe situations in
which the real parents, or their surviving kin, relinquish the child
to an adopting couple to raise as their own.
The adoption is not
accompanied by formalities, other than the bringing of gifts of
food or small articles by the adopting parents when they take the
child. Adoption is almost always between consanguineal kinsmen
of some degree, although such relationship is not essential. There
were five instances of adoption where there were no blood ties
between the child and either of the adopting parents.
For convenience we shall designate the member of the couple
who has the closest kinship tie to the real parents as the "adopting
parent." This does not mean that one of the adoptive parents acts
alone in the adoption. The evidence points, on the contrary, to
considerable accord between the parents on the matter of adop-
As shown in Table 20, the adopting parent is usually already in
the kinship category of 'mother' to the child. In more than half
the adoptions, moreover, the adopting parent is a member of the
child's own lineage
(see Table 21).
It should be noted that although in more than half the adop-
tions the adopting parent is a member of the child's lineage, this
by no means assures that the child's lineage by adoption will be
the same as his lineage of birth; when the adopting parent is a
man, the child becomes a member of the adopting parent's wife's
lineage. Actually, in 38 of the 57 adoptions in our study, the
child's adoptive mother was of a different lineage from that of his
real mother. This fact has important bearing on the way in which
adoptions work out.
Another table will help to make clear an additional point about
who is related to whom in an adoption. In Table 22, the cases have
been divided into two groups, depending on whether or not the
adoption was "voluntary."
TABLE 20 Relationship of Adopters to Adoptees.
No. of Adoptions
Subtotal Total
Adopting parent 'mother' 34
'Mo' in Ch's lineage 19
Mo's uterine Si (IO)
Other (9)
'Mo' in Fa's lineage 12
Fa's uterine Si (6)
Other (6)
'Mo' in Mo's Fa's lineage 2
'Mo' in Fa's Fa's lineage
Adopting parent 'father' 13
'Fa' in Ch's lineage 8
Mo's uterine Br (4)
'Fa' in Fa's lineage 5
Fa's uterine Br (2)
Other (3)
Adopting parent 'Sister' (in Ch's lineage) I
Adopting parent 'Brother' (in Ch's lineage) I
Adopting parent e[ekur' of Ch's lineage 2
Adopting parent efekur of Fa's lineage I
Adopting parent's relationship unclear 2
Adopting parent unrelated by blood s
a A lineage's ifekur are the children of its men. The eflkur of one's own lineage may be
classed as 'children' without regard to generation, or they may be classed according to
generation. The efikur of one's father's lineage may be classed with 'siblings' with regard
to generation, or may be classed according to generation.
bThe total exceeds the actual number of 57 adoptions because in two cases both adopt-
ing parents were immediate kin of the child on his father's and mother's side.
TABLE 21 Lineage Relationship of Adopters and Adoptees
Lineage of Adopting Parent
In Ch's lineage
In Ch's Fa's lineage
An efekur of Ch's lineage
An e[ekur of Fa's lineage
In parent's Fa's lineage
Relationship unclear
Not related by blood
No. of Adoptions
*In two cases the child was immediately related to both adopting parents.
It is evident that the commonest kin tie is one involving the two
mothers. Very often they are sisters, cousins, or occasionally aunts
and nieces. The next most common tie, as Table 22 shows, is
between the real father and the adopting mother. It can be seen
that the real mother is related to the adopting parents in 33 cases
and the real father in only 20. Similarly, the adopting mother is
related to the real parents in 36 of the cases and the adopting
father in only 16. Thus, women are more often the most closely
TABLE 22 Relationship of Real and Adopting Parent to Child
Related Parents Adoptions due to 'Voluntary'
Real Adopting Death or Illness Adoptions Total
Mother Mother 3 18 21
Mother Father 3 8 11
Father Father 4 5
Father Mother 4 11 15
No kin relationship
between parents 2 3 5
Unknown 0 2 2
TOTAL 13 46 59*
*Two cases involved a close double-kin tie.
related parents in an adoption; if we consider only the voluntary
adoptions, this tendency is even more pronounced. However, it
would be a mistake to assume from this evidence that women have
a greater "say" in adoptions. On the contrary, fathers in our sam-
ple frequently are the primary agents in particular adoptions. It
appears rather that residence patterns, which favor the relation-
ships between mothers, are responsible for this tendency. A closer
look at both adoption and fostering practices will help to illustrate
this and other points.
The Trukese recognize the distinction between 'adopting' mwuuti
and 'fostering' tumwunuuw. Although in borderline cases the basis
for the distinction is not always quite clear, they have no trouble
agreeing on who is adopted and who is not. The clearest basis for
an adoption is the contractual one between two sets of living par-
ents. Here the donor parents actually surrender not only the child
but their nemeni, their 'authority' or 'control' over him, to the
adopting parents. Fostering, on the other hand, is an expected kin
obligation in the case of death or family misfortune. Kinsmen step
in to care for dependent children whenever their parents cannot
provide for them. Occasionally this kind of expected fostering
becomes an adoption when the tie between the fostering parents
and the child has had a chance to grow unusually strong. This
comes about most commonly if the fostering parents are childless
and the child is very young when he comes into their care. Then
the fostering relationship not infrequently turns into what the
Trukese consider an adoption: 'The tie is broken'. The parents
then look upon the child as 'their own' and others in the society
recognize the claim.
In Romon um at the time of our study, fifteen individuals of all
ages were considered to be or to have been foster children rather
than adopted children. As a group, the foster children differed
from the adopted children in that they were always orphans; they
were generally more than five years old when their parents died;
there were no conflicting lineage affiliations involved in their situa-
tion; and whatever their age, they knew who their real parents
were. From the point of view of the guardians, foster care is an
obligation toward the children of kin. In adoption the emphasis is
on creating a new parent-child tie.
In addition to adoption and foster care, many children are given
over to the temporary care of other adults when for some reason
their own parents cannot take care of them. The reason may be
the illness of a parent or sibling, or commonly, the advent of a
new baby. Whatever the occasion, the adults are usually closely
related, and everybody understands that the arrangement is a tem-
porary one.
The lack of clear-cut distinction between adoption and foster-
ing, and between fostering and temporary care, reflects the non-
exclusiveness and flexibility in the attitude of the people of
Romonum toward the nurture of children. The Trukese are people
who value children highly. New babies are always welcome, and
everyone seems to enjoy holding them and carrying them around.
Within a gathering of kin, babies are treated almost as common
property, being passed about freely from one adult to another. On
one occasion, during a church service, I observed a fretful baby
who was first fed by his young mother and then passed along to
seven different people, all of whom in turn held and played with
him before he was finally carried out on the shoulder of his grand-
father. Relatives of all ages and of either sex simply reach out and
take a baby from one another; nobody bothers to ask permission.
It is important to see the practice of adoption on Romonum
against this background. The conditions of life are such that all
children are cared for at one time or another by adults other than
their real parents. They live in households that may number as
many as twenty people, and they visit other households sometimes
overnight, sometimes for days or weeks at a time. The coming of a
new baby is a favorite occasion for long visits. When the baby is
due, the expectant mother, accompanied by her husband and chil
dren, goes to live with an older female relative; or, if it is more
convenient, one or another of her sisters or aunts may move into
the household and stay from a few months to a year to help the
new mother in caring for both the new baby and the older chil
dren. These preferred housing or visiting arrangements which bring
related women together help to account for our observation that
women are more often the most closely related parents in adop-
Other reasons for shifting from one household to another are
common. A child who has misbehaved may run off to sleep for a
night or two with some of his kin in another house until his par-
ents are no longer angry with him. Another may sleep with rela-
tives who can provide him with space in a sleeping net if his
parents cannot. Boys are routinely expected to move out of
their parents' households at puberty because of the brother-sister
taboos. Children usually return to their own parents to be fed, but
everyone is expected to share food with a visiting child, and in fact
it is a matter of pride with a family temporarily well-supplied with
food to provide for the children of their kin who may be around
at mealtime.
These conditions surrounding child care provide an important
part of the context within which voluntary adoptions-those not
dictated by strict necessity-take place.
Sometimes an unborn child is spoken for by a couple who are
childless or who may have only one child. In the latter situation,
the sex of the child is often important. A couple might say, "If
your baby is a girl we would like to adopt her as a sister for our
only son" or, if their only child is a daughter they might ask for a
boy "to be a brother to our daughter."
When the parents have
worked out an understanding in advance, the adopting couple
brings presents to the real parents. Food is exchanged, and after
the baby is born soap and cloth are brought for the child. Clearly
the adopting parents feel obligated to the natural parents, but the
gifts that are given appear to be only small tokens.
If the two sets of parents live on the same island, the adoptive
mother and the baby get acquainted over the months when the
baby is still breast-fed. Adoption and weaning frequently coincide,
usually when the baby is about nine to ten months old. Trukese
mothers occasionally wean their babies by giving them to a relative
for a week or two until the child gives up trying to nurse, so it is
not considered a particular hardship for a baby to be given to his
adoptive mother to wean.
As might be expected, adoptions between parents who live on
different islands present the most traumatic situation for the child.
Here, where the adoptive parents have not been able to visit con-
tinually with the real parents, the child is taken away by virtual
strangers and weaned by them. Fortunately this kind of adoption
is infrequent, but one such child who was observed for the year
following his adoption at ten months of age, showed the classic
symptoms of maternal deprivation described by Bowlby (1953).
The most common adoption experience, however, appears at least
on the surface to be not much more traumatic than the ordinary
weaning experience of the nonadopted child.
Arranging to adopt a child before or shortly after his birth, in
the absence of any crisis in the lives of his real parents, requires
that the adopting parents have some kind of social leverage vis-a-
vis the real parents. The adopting parents in such instances are
usually older, securely situated people who are not only in a senior
relationship to the parents but also clearly in a position to do well
by the adopted child. As tokens of their future intentions, and
also to obligate the real parents before the baby arrives, they bring
little presents to the family, particularly to the pregnant mother.
After the birth they bring things like soap, baby clothes, or canned
milk to the baby.
In family crises, arrangements must be made with less deliber-
ation. When the mother of an infant dies, the child is generally
taken by one of the mother's attending female relatives, preferably
a childless sister. If no sister of the mother is available, the father
may give the child to one of his own female relatives, again, most
often a childless one. In any event, the father ordinarily remar-
and gives up his primary interest in and control over the
child. If the mother of the baby is too ill to care for him or nurse
him the same procedure is followed,
and the woman who is
called in to care for the child generally adopts him as her own.
If a father dies, or is seriously ill, one or more of his young chil-
dren may be adopted by relatives in order to help the mother. In
particular, female relatives of the father, as an expression of their
love for him, will off er to adopt one of his children. One woman,
struggling to take care of her family while her husband was away
for several years working on Ocean Island, gave two of her numer-
ous children for adoption by her husband's childless sisters. This
helped to relieve her situation, and it was considered an acceptable
solution, since her husband, on his return, would not object to his
sisters' having the children.
As we have noted, helping a relative who has many children is
given as one of the three main reasons for adoption. Here, as in
other kinds of adoption, the adopting parents are characteristically
childless, but their stated reasons for adopting tend to emphasize
altruism rather than self-interest. Such adoptions often come
about after the adopting parents have lived with the real parents
and helped them with their children. They may take care of the
youngest child when his mother becomes pregnant again or over
the period when a new baby is born. It is not a large step from this
kind of helping-out to an actual adoption; and a mother of many
children appears to be relieved, rather than threatened, when such
a couple agrees to take on the full care of a particular child. It is
not always easy to feed and clothe a large family of growing chil-
dren, and parents with many children often look upon adoption
not only as a help to themselves but as an opportunity for the
child. It is interesting that ten living mothers gave up two or more
children for adoption, accounting for a total of twenty-two chil-
dren in our sample.
It is clear from the age distribution of our sample, as shown in
Table 23, that adopting parents prefer younger children.
The reason for preferring young children is that the younger the
child the more likely he is to look upon the adoptive parents as his
real parents. If the child is less than two years old, the adoptive
parents try to keep him from knowing his real parentage, even
though both sets of parents may be living in the same household.
Other people cooperate to a surprising degree in keeping a young
TABLE 23 Ages of Children at Adoption
Age No. Age No.
l yr. or less 32 5 yrs. 2
l3 mos.-2 yrs. 10 6-7 yrs. 0
2-3 yrs 7 8 yrs. l*
4 yrs. 4 unknown 1
*The one instance of a child adopted at eight years of age came about when a land-rich
lineage, without junior members, adopted a girl from a related line to inherit their
lineage lands. It is one of only three cases in our sample in which inheritance of property
was noted as a significant factor.
child from knowing he is adopted. There are stories of adopted
children accidentally finding out the truth by overhearing an adult
conversation, but there are no accounts of anyone's learning the
truth through taunts, or through someone's callously or mali-
ciously telling him. The real parents are expected to cooperate in
keeping up the fiction when the child is small. One mother, who
had given up her ten-month-old baby to her husband's cousin,
came to visit him a month later, and although the visit was per-
mitted, the cousin's wife scolded her, saying, "Why do you want
to make things hard for me by coming when your baby still re-
members you?" The object is to give the adoptive parents a real
chance to establish a firm parent-child tie.
Because of this early conspiracy of silence, many adopted chil-
dren reach adolescence without finding out that they have been
adopted, but, none in our sample reached the age of sixteen with-
out learning of their adoption. By this age, and for reasons that we
shall soon examine, it becomes important to know one's real iden-
tity. For some adopted children the knowledge comes as little sur-
prise and without appreciable shock, particularly if they have been
living and visiting closely with both sets of parents and if there
is no conflict of lineage affiliation. For others the news can be
very disturbing. One elderly woman, who with her husband had
adopted an infant when his mother died, told of the poignant
scene when she informed the fifteen-year-old boy about his real
identity. She felt she had to tell him because she and her husband
might die soon, and she was afraid that he would not be treated as
well by the surviving members of her lineage as he would be by his
real older siblings. As she spoke of it, and of the pain it caused the
boy, there were tears in her eyes.
By now it should be apparent that adoption on Romonum
is not necessarily for keeps; one's real identity remains intact,
whether as a latent social fact or as an expressed one. One very
good reason why children find out that they are adopted by the
time they are fifteen or so is that they are still subject to the incest
regulations and brother-sister avoidance taboos of their real iden-
tities. By the time they are ready for courtship, they need to know
who they are in order to abide by the taboos. These restrictions
between real brother and sister also obtain between brother and
sister by adoption, but violation in the latter relationship is a
"small sin" rather than a large one.
To see the kinds of complications that can arise from the
adopted child's dual identity, consider the situation in which
a woman adopted her brother's daughter, and ten years later
adopted her brother's son's daughter. The two girls, sisters by
adoption, were in fact aunt and niece. The older one, when asked
whether she looked upon the younger as sister or as brother's
daughter, said: "Well, as little sister, some of the time, when I help
her out; but I cannot ask my brother's daughter to help me wash
my husband's clothes, although she can wash my sons' clothes
since she is their 'sister'."
There are a number of situations in which an adoption may be
nullified or temporarily suspended. The first is on the initiative of
the real parents in the first critical weeks after the child has been
given up. If they find that the separation is just too hard to en-
dure, they may come to take the child back. The few recorded
cases in which this happened involved the adoption of children
from another island, and it may be supposed that one of the fac-
tors making for difficulty was the completeness of the separation.
Terminating an adoption in this way appears to be difficult for
everyone, but particularly for the adoptive parents. Sometimes it
requires the prestige and offices of a common elder kinsman to get
the baby back. If it is to be done at all, it must be done early in
the adoption, before the adoptive parents have had time to make
much of a physical investment in the child; the longer they feed
and clothe the child the more they bind the adoption and the
harder it is for the real parents to get the child back.
Another reason for terminating an adoption is any apparent
neglect of a child by his adoptive parents. Although this occurred
only once in our sample of fifty-seven, adoptive parents are sensi-
tive to the possibility, and to avoid shame and loss of the child
they usually try to be conspicuously good to an adopted child.
Besides feeding him and clothing him well, they frequently give
him generous gifts of land and property. Adoptive parents who
subsequently have a child of their own are watched with particular
care to see that they continue to treat their adopted child well.
As the child grows older, another period that is critical for
adoption becomes apparent. By late adolescence the child begins
to become an economic asset to his parents, and although his emo-
tional ties are firmly established with the adoptive parents, his real
parents may think it profitable to try to get him back. The two
such cases on record both went to the courts and were settled
when the adoptive parents drew up a bill for what they had put
out for the child's care. In each instance, when the child's real
family was asked to reimburse the adoptive family and could not,
the case was considered closed and the child remained with the
adoptive parents.
Another kind of crisis that may result in a total reversal of
adoption comes upon the death or divorce of the adoptive parents.
If the real parents are still alive, the child ordinarily returns to
them. Otherwise, the child most commonly stays with a sister of
the adoptive mother, and thus becomes a foster child. In two
instances of divorce however, the adoptive father, although he was
the unrelated parent in each case, was permitted to keep the child
by express permission of senior members of the child's own fam-
ily. In these two cases, it was obvious that the real family members
had an ultimate say over the disposition of the child and were
using it to express disapproval of their kinswoman's role in the
divorce, and also to reward the husband for his faithful role as
father to the child.
A temporary suspension of the adoption may occur if the child
becomes seriously ill. His adoptive parents notify his real parents
and, if possible, take him to his real parents so that they can share
the responsibility of caring for him. Similarly, when an adopted
girl is about to have a baby, she returns if possible to the home of
her real mother for the event, where she is attended by both her
real and adoptive female relatives. The reason given is that the
adoptive parents want the real parents to see that they are doing
everything possible for the child. In the event the child should die
they cannot be blamed for neglect.
The last point at which a crisis may occur in an adoption comes
when the child must resolve a conflict in lineage affiliation. Of our
recorded adoptions, 38 of the total 57 entailed such a conflict.
When the child approaches his late teens he is old enough to take
on some lineage rights and obligations; these include giving his
labor for lineage feasts or other social affairs, using lineage lands,
joining in work parties, and so on. He must eventually choose one
or the other of the two maternal lineages available to him. The
important factor in the decision between the two seems to come
down to a matter of their relative strength, which in turn depends
both on land resources and the number and industry of the lineage
members. The adopting lineage may have the advantage in land
but it rarely has the advantage in people. The adopted child comes
typically from a large family to a small one, and the pull back into
the real family of siblings is strong in this society, where the para-
mount economic and social unit is a sibling group.
It is understood by everyone that the option is the child's. The
adoptive parents may desire that he stay in his adoptive mother's
lineage, but they cannot require him to, nor can they secure his
stay with them in any way, except possibly by giving him so much
land that he has no need to align himself with his own lineage
mates. In our sample two grown women maintained almost equal
ties with both lineages, but, aside from these, all the others who
had a conflict of lineages chose to affiliate with one or the other
lineage by the time they were twenty years old. Interestingly, al-
most twice as many chose to join the real mother's lineage as
chose to join the adoptive mother's lineage. If the adoptive lineage
was chosen, it appeared to be for one of three reasons: the real
lineage was no longer functioning; the child had been adopted
from another island-in effect a variant of the first; the child stood
to gain significantly more in land, wealth, and family by staying
with the adopting lineage than by returning to his real lineage
mates. The last reason accounts for two cases of adoption in which
the principal motive was to provide an heir for a lineage that was
dying out.
It should be understood that the warm relationship between the
adoptive parents and the child usually endures despite the fate of
the adoption or the child's choice of his natal lineage over the
adoptive one. One adopted boy had to be returned to his real par-
ents when he was about four years old because of the death of his
adoptive mother. He kept up a close relationship with his adoptive
father, however, and when the latter died the sixteen-year-old
boy's grief appeared to be as real as that of any of the man's own
kin. A good child, when he is grown, will continue to take care of
his adoptive parents, even if he is identified as now belonging to
the lineage of his real kinsmen. He may contribute, in a limited
way, to his adoptive mother's lineage and will give food and other
gifts to his adoptive parents.
It can be seen from the foregoing discussion that the relation-
ship of adoption, which is meant to be a stable and permanent
one, is in fact quite fluid and even reversible. However, the emo-
tional bonds of adoption do in fact persist and, over the years,
provide the participants with what is, by Romonum standards, a
quite satisfactory parent-child relationship. It is not true that
adoption actually 'breaks the tie'. Adoption cannot wipe out a
child's claim to his own identity and lineage any more than it can
abrogate rules of incest, but it does give the individual an alterna-
tive set of obligations and rights for as long as he chooses to exer-
cise them, as well as giving him parents to care for him in his grow-
ing years.
The concluding section of this chapter will deal with motivation
and reward in the adoption practices we have been describing. The
first relevant question concerns the motivation of the adopting
parents, since they initiate most of the cases of voluntary adop-
tion. After that we will consider adoption from the point of view
of the child's interests, and finally the needs or motivations of the
real parents.
Why do childless couples on Romonum adopt children? The
usual reasons given are either that they want a child or that they
want to help a relative who has too many children. The first rea-
son continues to beg the question, while the second, which may
serve as a socially sanctioned rationalization, is not otherwise very
informative. After all, one might more appropriately help a rela-
tive with a large family by giving him trees or land, or by working
for him.
It is rarely necessary for an individual to have children in order
to have an heir. In most cases one's lineage mates will have pro-
duced enough children to carry on the lineage as a social entity.
Why not enjoy the freedoms and privileges of childlessness if one
has no children? Why is it that virtually every childless couple on
Romonum who are over thirty have adopted or fostered a child?
Some possible answers to these questions are to be found in
what is already known of Trukese personality structure and Truk-
ese values. A review of the literature on the personality dynamics
of the Trukese on Romonum reveals some points of similarity
with the picture of Hawaiian adoption as a kind of dependency
inversion drawn by Howard, et al. (chapter 2). Certainly the de-
pendency trauma of the Romonum child is severe enough between
the ages of two and five; punishment is just as arbitrary and unpre-
dictable as that reported for the Hawaiians, and verbal reinforce-
ments for correct behavior are similarly infrequent. But there are
enough differences in the situation of the Trukese on Romonum
as compared to the Hawaiians to lead one to suspect a significant
difference in dynamics. In the Hawaiian study, adoption appears
to be a somewhat compulsive attempt to fill an essentially unap-
peasable need in the psyche of women. It works satisfactorily for a
while, as long as the child is truly dependent, and then it becomes
increasingly disappointing as the child matures. From what we
have seen of the way adoption works on Romonum, we suspect
that Trukese adoption brings a broader spectrum of rewards to the
adopting parents, and may perhap