Pahari Polyandry: A Comparison

Author(s): Gerald D. Berreman

Source: American Anthropologist, New Series, Vol. 64, No. 1, Part 1 (Feb., 1962), pp. 60-75
Published by: Wiley on behalf of the American Anthropological Association
Stable URL: http://www.jstor.org/stable/666727
Accessed: 04-05-2017 04:53 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

American Anthropological Association, Wiley are collaborating with JSTOR to digitize, preserve and
extend access to American Anthropologist

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
Pahari Polyandry: A Comparison

GERALD D. BERREMAN

University of California, Berkede

pOLYANDRY has long been a popular subject for specu

sionally for research by anthropologists.1 Recently effort
origin and functioning of this rather unusual institution ha
mented by attempts to define it (Fischer 1952; Leach 1955;
Prince Peter 1955a). It can be most simply defined as that form
which a woman has more than one husband at a time.2 In frater
which is by far the most common kind, a group of brothers, re
tory, are collectively the husbands of a woman (or women).
This kind of polyandry has been reported from many par
(Westermarck 1922:107 ff.), but its best-documented and m
currence is in Tibet, described by Prince Peter (1955c: 176) a
most flourishing polyandrous community in the world today
Mandelbaum (1938:581 f.) notes that "in South India polyan
cially frequent occurrence. Six polyandrous tribes have been reported for
Cochin; the Nayars of Travancore and the Irava of British Malabar have this
form of marriage; while the Todas are the classic example of a polyandrous
people in the textbooks of anthropology." The Singhalese are known to prac-
tice polyandry to some extent (Leach 1955). In North India the Jats of the
northern Punjab, and especially those who are Sikhs, have been repeatedly
reported to practice polyandry (Briffault 1959:137; Kirkpatrick 1878:86;
Prince Peter 1948:215). The most consistent practitioners of polyandry in
India today are probably the residents of certain sub-Himalayan hill areas in
Himachal Pradesh, the northern Punjab, and northwestern Uttar Pradesh. It
is the polyandry of this relatively little-known region which I propose to dis-
cuss in this paper.
Non-Tibetan, Indo-Aryan-speaking Hindus inhabit the lower ranges of the
Himalayas from southeastern Kashmir across northernmost India and through
Nepal. These people are collectively termed Paharis ("of the mountains").
They constitute a distinct culture area bordered by the peoples of Tibet to the
north and by those of the Indo-Gangetic plains to the south. With the latter
peoples they share historical origins as well as linguistic and cultural affinites
(Berreman 1960). Among Paharis, polyandry has been reported in several
districts (cf. Das-Gupta 1921) and has been studied in some detail in Jaunsar
Bawar, a subdivision of Dehra Dun district in northwestern Uttar Pradesh
(Majumdar 1944; Saksena 1955). All of the Pahari areas in which it occurs are
in the western Himalayan hills and are inhabited by people who share the
Western dialect of the Pahari language (Grierson 1916:101). There are no re-
60

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
[BERREMAN] Pakari Polyandry 61
liable reports of Pahari polyandry east of Jaunsar Bawar and its immediate
vicinity, i.e., in the Central or Eastern Pahari-speaking areas.8
This paper is based on a study carried out among Central Pahari-speaking
people in Garhwal, a hill area adjacent to, and east of, Jaunsar Bawar.4 The
people of Garhwal, though they have not previously been studied, have fre-
quently been cited as nonpolyandrous by those who have written on polyandry
in Jaunsar Bawar. One goal of the research which led to this paper was to study
marriage in its total cultural context among the nonpolyandrous people of
Garhwal in order to compare that system with the polyandrous system of
neighboring Jaunsar Bawar as reported in the literature. The general hypothe-
sis with which the investigation began and which this paper will discuss
was that economic, demographic, or social-structural differences would be
found which would correlate with the occurrence of polyandry in Jaunsar
Bawar and its absence in nearby Garhwal. Further, some of the features found
in Jaunsar Bawar would correspond to those reported by people who have
studied polyandrous societies in other parts of the world.
The peoples of Jaunsar Bawar and Garhwal live under virtually identical
physical conditions and their populations and cultures are very similar, having
derived from a common source (Berreman 1960). Conditions for a comparative
study with polyandry as the dependent variable therefore seemed ideal. In both
areas the economy of the majority high-caste population is primarily agri-
cultural, with a secondary dependence on animal husbandry, while the low-
status artisan castes live by their craft specialities. Land is valuable but not as
scarce as in most of North India. All property is owned jointly by male mem-
bers of the patrilineal, patrilocal extended family. If property is divided among
brothers, they usually receive equal shares. Normally, however, brothers con-
tinue to hold the patrimony in common and division occurs in the next genera-
tion, among patrilateral parallel cousins. The eldest active male dominates in
the joint family but cannot compel younger men to remain within it. Marriage
takes place within the caste group and outside the clan and mother's clan. It
involves a payment of bride-price which must be returned if the marriage is dis-
solved unless the husband is clearly at fault. Where dowry is used it is excep-
tional and evidently of recent origin, having diffused from the plains. Levirate
is the rule upon a husband's death and payment must be made to his family if
his wife wishes to go elsewhere. These are general features of Pahari culture as
I know it and as it is reported in the literature.

POLYANDRY IN JAUNSAR BAWAR

In Jaunsar Bawar fraternal polyandry has been described as "the common
form of marriage." Indeed, it does seem to be the preferred, but not the exclu-
sive, form. Monogamy, polygyny, and fraternal polyandry, including a com-
bination of polyandry and polygyny approximating fraternal "group mar-
riage," appear in the same villages and even in the same lineages (Majumdar
1944:167 f.). Nonfraternal polyandry is not reported. In the one village for
which figures are available, Majumdar (1955b: 165) reports that of 57 families,

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
62 American Anthropologist [64, 1962

39 percent are polyandrous with more husbands than wives, 10 percent are
polyandrous with an equal number of husbands and wives, 12 percent are
polygynous, and 39 percent are monogamous.5
In this society a polyandrous union occurs when a woman goes through a
marriage ceremony with the eldest of a group of brothers. This man represents
the group of brothers, all of whom thereupon become the woman's husbands.
Subsequent wives may be taken, especially if the first one is sterile or if the age
differential of the brothers is great. If so, the wives are individually married in
a ceremony with the eldest brother and are shared by all, unless one or more
brothers wish to break away from the joint family. No brother can remain a
member of the joint family and claim exclusive rights to a wife. The eldest
brother dominates with respect to the wife or wives, but he has no exclusive
sexual or reproductive rights. A woman considers all of the brothers to be her
husbands. Children recognize the group of brothers as their fathers; they call
all of them "father" and inherit from all as a group without regard to paternity
or maternity within the polyandrous family (Majumdar 1944:178; 1953: 179).
In cases of division of the family, paternity may be assigned by lot, by mother's
designation, or by order of birth (Majumdar 1944:144 f.). This is "true"
fraternal polyandry similar to that reported among the Iravas of Central
Kerala by Aiyappan (cf., Aiyappan 1935:114 if.; Leach 1955:182; Gough
1959:34).

MONANDRY IN GARHWAL

Majumdar (1944:168) has pointed out that "the Garhwalis d

serve polyandry but the Jaunsaris do." While people of Jaunsar B
knowledge their polyandry quite readily and defend this custom, the idea of
polyandry is rejected by Garhwal residents. I neither found nor heard of any
case of polyandry in the area of my work. Of a total of 300 marital unions for
which I accumulated complete information in one Garhwal village, 85 percent
were monogamous and 15 percent were polygynous. There every family is
careful to secure a wife for each of its sons, and each son normally goes through
the marriage ceremony with his own bride.6
Although there are strong negative feelings about polyandry in this region,
sexual relationships within the family are not greatly different from those
among fraternally polyandrous families of Jaunsar Bawar. The situation is
very similar to that among the Kota as reported by Mandelbaum (1938).
Brothers have the right of sexual access to one another's wives. Despite these
rights of fraternal ciscisbeism, every man has his own wife and each child its
own father. There is never ambiguity on this point. Brothers share their wives'
sexuality but not their reproductivity. As long as a wife fulfils her sexual ob-
ligations to her husband and does not indicate a preference for another, she is
normally available to all of her husband's brothers, but her children are the
children of her husband only.
In assessing the hypothesis with which this study began I will look briefly
at some of the factors which have been advanced in the literature as causal for,

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
BERREMAN] Pahari Polyandry 63

predisposing toward, or correlated with polyandry and consider them with

reference to the societies being described here.

FACTORS ASSOCIATED WITH POLYANDRY

Economic factors
Contemporary discussions frequently emphasize econ
counting for polyandry.
E. R. Leach (1955:183 ff.) believes that polyandry" . .
ciated with an institution of dowry rights," and has h
adelphic polyandry is consistently associated with system
well as men are the bearers of property rights." In suc
guished from those in which property is exclusively in the
marriage "establishes a distinct parcel of property right
If two brothers share one wife so that the only heirs of the brothers
that wife, then, from an economic point of view, the marriage will tend
the sibling pair rather than tear it apart, whereas, if two brothers
children will have separate economic interests, and maintenance of t
in one piece is likely to prove impossible (Leach 1955:184).

Polyandry thereby also serves "to reduce potential host

brothers." Without polyandry there would be a tenden
brothers to break up the joint family in order that each gr
pursue its own economic interests.
The Pahari evidence contradicts this hypothesis. Dowry is not part of the
traditional Pahari marriage transaction which is, in fact, dependent upon
bride-price for validity (Joshi 1929:50 f.). Moreover, and more importantly, a
woman has no property of her own except in most unusual circumstances and
she forfeits even her jewelry if she divorces her husband. Children remain with
their father or his family when a marriage dissolves. Therefore, in the Hima-
layan hills, children of brothers who share a wife have no different economic
interests as a result of that fact than do children of brothers each of whom has
his own wife.7
A widely cited economic advantage of fraternal polyandry is that it keeps
family property, especially lands, intact in a patrilineal, patrilocal group
(Westermarck 1922: 185 f.). It accomplishes this by restricting the number of
heirs and by keeping them together around a common wife. This virtue of
polyandry is cited for Ceylon (Prince Peter 1955b), Tibet (Prince Peter 1955c),
and the Himalayan hill area (Saksena 1955:33; Stulpnagel 1878:133).8 It is an
advantage that is claimed by Jaunsar Bawar people themselves (Majumdar
1944:168). Where this economic function is served by polyandry it is attrib-
buted to the desire to keep intact the property of the wealthy, to the necessity
to keep the property of the very poor from dropping below the subsistence
level, or to both (cf. Prince Peter 1955b: 169; Stulpnagel 1878:133 ff.).
If fraternal polyandry were practiced consistently, there would be no
patrilateral parallel cousins, and it is they who generally divide land in the
Himalayan hills. If the number of wives were appreciably less than in non-

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
64 American Anthropologist [64, 1962

polyandrous societies this would reduce, absolutely, the number of offspring

and hence heirs. Either or both of these would theoretically tend to reduce
fragmentation of property, expecially if, as will be discussed below, polyandry
were to reduce the frictions which lead to break-up of joint families.
The over-all effect of polyandry for family property retention in Jaunsar
Bawar is tempered by the fact that not all marriages are polyandrous; of those
that are, many involve a plurality of wives. While in the Jaunsar Bawar village
cited above 49 percent of the families are polyandrous, 61 percent have as
many (or more) wives as husbands and hence no reduction in the number of
heirs. Occasional polygyny or monogamy among brothers in a lineage might
wipe out the advantage, for property retention, of generations of polyandry.
On the other hand, in the nonpolyandrous Garhwal village three of 16 land
allottments have remained intact in the joint jamilies to which they were
assigned nearly 150 years ago. Unfortunately, land fragmentation figures for
Jaunsar Bawar are not available to compare with those of the Garhwal village.
My guess is that they would not show significant differences.
Polyandry has often been attributed to economic hardship which neces-
sitates cooperative work among brothers for survival (Kapadia 1955:71;
Majumdar 1944:168). The expense of obtaining and/or maintaining a wife
and of supporting a family has been cited as an important factor in contribut-
ing to polyandry in several contexts. Stulpnagel (1878:133) and Kapadia
(1955:71) mention the difficulty of raising a sufficient bride-price and the con-
sequent necessity for several brothers to combine to purchase a single wife.
Majumdar (1955a:95) notes the similar difficulty of providing the costly
jewelry which a Pahari woman requires.
Bride-price marriage, though it is the rule in Jaunsar Bawar as elsewhere in
the Himalayan hills, is not a likely motivation for polyandry since the amount
is proportional to the wealth of those involved. Moreover, permanent unions
may be established without payment at all. In the nonpolyandrous areas men
do not often go unwed because of bride-price. Precisely the same points are
applicable with regard to the bride's jewelry.
Heath (1955) has suggested that polyandry is related to "... sex special-
ization in which the woman makes only an insignificant contribution to sub-
sistence." This explanation could not be farther from the facts found in the
Himalayan hills, including Jaunsar Bawar.9 There women contribute as
heavily as do men to subsistence, and a wife is an economic asset (Majumdar
1944: 171). In the Garhwal village which I studied, need for additional field
labor was cited as a reason for securing an additional wife in eight of twelve
current cases of polygyny. As one villager remarked, "Here two wives are better
than one because they do much of the work. In your country and on the plains
the husband has to support his wife so a second wife is a hardship and a
luxury." In polygynous families the agricultural and pastoral labor is divided
among wives just as in polyandrous families it is divided among husbands,
except that in both cases plowing is reserved for men and certain household
tasks for women. Saksena (1955:33) notes that in the difficult economic cir-

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
BERREMAN] Pahari Polyandry 65

cumstances of Juansar Bawar it often takes several men to support a single

wife and family: "In order to make life successful a system of life in keeping
with the demand for joint labour within a village had to be evolved. The wide
practice of polyandry seems to be the outcome of this demand." However,
polyandry is only one means of enlarging the work force of the family. In
Garhwal (and in many families of Jaunsar Bawar) the same end is achieved
by polygyny, by adoption of sons, by hiring agricultural servants, or prefer-
ably by having several sons.
An advantage of polyandry may be that it tends to keep the ratio of work-
ing adults to children high in the family, just as it keeps the number of heirs
low. In the polyandrous village mentioned above, this would apparently not
be true for the 61 percent of all families who have one or more wives per hus-
band. However, in that village about 20.5 percent of the population is ten
years of age or under, while in the nonpolyandrous Garhwal village which I
studied about 28 percent of the population is in this age bracket. This sample
is far too small to yield significant conclusions, but it does not contradict the
contention that there are fewer children in polyandrous communities than in
nonpolyandrous ones-an advantage in an economically hard-pressed area.
It must be noted in respect to all of the economic arguments for polyandry,
that polyandrous Jaunsar Bawar is no more hard-pressed than nonpolyan-
drous Garhwal, and that Paharis in general are economically more secure than
many people of North India, despite their reputation for poverty (Berreman
1959:102).

Social factors
Security of wife and family in the prolonged absence of the husband has
been noted as an advantage of polyandry among such martial peoples as the
Jats of the northern Punjab and the Nayars of South India (Prince Peter
1948:223; 1955b:169; Westermarck 1922:193). Likewise, it has been cited as
an advantage to Paharis who travel considerable distances to tend lands and
cattle and are therefore absent from their homes for extended periods (Ka-
padia 1955:72). Brothers can arrange to protect a common wife in such cir-
cumstances where an individual could not. This advantage accrues equally
in formal polyandry and in wife-sharing. In Garhwal a man may be sent to
accompany his brother's wife on a trip or while she works in the forest or even
to live with her in the absence of her husband to insure that she will have no
liaisons with men outside the family.
A more fundamental social function of polyandry, and one of the benefits
most widely acclaimed by both observers and practitioners of polyandry, is
the maintenance of intrafamilial amity, i.e., it reduces quarrels among brothers
(Aiyappan 1937; Carrasco 1959:36; Leach 1955:185; Prince Peter 1948:224;
1955a:181; 1955b:170; Saksena 1955:33). In India, joint family dissolution
is frequently attributed to friction among wives who enlist the support of
their respective husbands, with resultant fraternal strife.?? Polyandry is said
to minimize fraternal conflict by eliminating this source, though jealousy over

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
66 American Anthropologist [64, 1962

the common wife or wives is also reported (Mukherji 1950). As was indicated
above, Leach attributes decreased friction in fraternally polyandrous families
to the identity of economic interests among their members. Reduction of fric-
tion might be achieved in part by the simple reduction in number of heirs
which polyandry theoretically accomplishes and the consequently decreased
number of potential disputants in the family. Unfortunately, no data such as
frequency of joint family dissolution are available with which to test this
alleged advantage of fraternal polyandry.

Socio-economic factors
Radcliffe-Brown defined the unity of the sibling group as "its unity in rela-
tion to a person outside it and connected with it by a specific relation to one
of its members," and he said that "it is in the light of this structural principle
that we must interpret ... adelphic polyandry.... " (Radcliffe-Brown
1941:7 f.). Prince Peter (1955a:181) has suggested that "the economic func-
tion" of polyandry "intensifies the unity and solidarity of the sibling group."
The missionary Stulpnagel (1878:135) commented that in the Himalayan
hills "polyandry is ... in reality nothing more than a mere custom of com-
munity of wives among brothers who have a community of other goods."
Majumdar (1944: 172) has made the same point with regard to property and
polyandry in both Tibet and the Himalayan hills where he has described mar-
riage as a "group contract." This corresponds closely to the explanation for
fraternal wife-sharing among the Kotas given by Mandelbaum (1938:575 ff.),
who describes it as one manifestation of a general principle of "equivalence of
brothers" which shows itself in the sharing of labor and property, and which
is maintained because (and as long as) it is economically worth while. Leach
refers to similar "corporate polyandry" among the Iravas of Central Kerala
as described by Aiyappan (Leach 1955:182).
In Jaunsar Bawar and Garhwal, a group of brothers has the kind of unity
to which Radcliffe-Brown referred. It is expressed prominently in economic
matters, but also in ritual and social relations. The unity is especially appar-
ent in the relationship between a group of brothers and their wife or wives.
Marriage in these areas is in a sense a group transaction in which the family
pays collectively for a woman and acquires her economic, sexual, and repro-
ductive services. All three kinds of services are shared by a group of brothers
in Juansar Bawar. In Garhwal, the first two services are shared by the brothers
while the third, reproductive capacity, is granted to one brother exclusively
during his lifetime and is passed to another on his death by the practice of
levirate. Kapadia (1955:66) has discussed in some detail the Pahari woman as
the "property" of her husband(s) and the implications of this concept.
The economic arrangement helps explain the community of interest in the
wife, but it leaves unexplained the difference in marriage pattern between
Jaunsar Bawar and Garhwal, and it leaves unanswered the question of why
groups in other parts of India with a similar community of property among
brothers do not tolerate either fraternal polyandry or wife sharing.

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
BERREMAN] Pahari Polyandry 67

Psychological factors
Psychological functions of polyandry have been little discussed in the liter-
ature and I have no new data on this subject. Prince Peter's suggestion
(1955a:181) that polyandry satisfies repressed incestuous desires seems ten-
uous at best.

Traditional factors
Most people attribute their customs to tradition. In India, polyandry is
widely attributed to specific traditions, notably those embodied in the religious
epic, Mahabharata, which tells of the exploits of the five Pandava brothers
and their common wife Draupadi. Almost every group that practices fraternal
polyandry in India attributes the practice to that precedent, and usually to
an intimate association between themselves and the deities of that epic (cf.
Kapadia 1955:52 f., 75, 92 f.; Prince Peter 1948:223). Paharis are well known
as devotees of the Pandavas who roamed these very hills in their legendary
travels. This tradition in the Himalayan hills has led to such statements as that
of Munshi (1955:i) who says that Jaunsar Bawar culture represents "a fossil
of the age of the Mahabharata."
The historical origins of polyandry in the Himalayan hills have been spec-
ulated upon at some length by Saksena. Mayne is quoted as having suggested
that polyandry was adopted by the Indo-Aryan invaders of India from the
aborigines or neighboring polyandrous people, and Majumdar seems to share
this view (Saksena 1955:30). Among neighboring people most often cited as
possibly influential are the polyandrous Tibetans with whom Paharis have
long been in occasional contact. Saksena holds the widespread view that poly-
andry in this area is a remnant of the culture of early Indo-Europeans who
came to India via the Himalayan hills. Support for this opinion is found by its
proponents, not only in the polyandry of the Mahabharata, but in other Hindu
classics and ancient records wherein polyandry and other traits characteristic
of the hills, such as animal sacrifice, meat-eating, freedom of women, widow
remarriage, and lack of caste rigidity are mentioned without disfavor (cf. Brif-
fault 1959:138 f.). Saksena summarizes his view in the following words:
... A polyandrous belt can be traced extending from Jaunsar-Bawar through Kangra Valley to
Hindu Kush and even beyond. This led Briffault to remark, "The practice of polyandrous marriage
is among the Indo-Aryans of the Panjab associated with other survivals of a more archaic and
tribal order of society, which are culturally identical with the usages of the polyandrous people
of Hindu-Kush, whence the invaders came to India " (Saksena 1955:30).
It is, therefore, evident that polyandry was an institution not unknown to the early Aryan
settlers in the Western Himilayas from where it gradually spread southwards, and is even now
the accepted form of marriage among the Rajputs and Brahmans of Jaunsar Bawar. To quote
Briffault again: "The highland regions of the Himalayas are but a residual cultural island which
preserves social customs that had once a far more extensive distribution. The institutions which
are found there were once common throughout the greater part of Central Asia" (Saksena
1955:32).

Thus, it is possible that polyandry was an acceptable form among the an-
cestors of the Central Asian invaders who are presumed by many to be an-
cestral to present-day high-caste Paharis. It is also possible that it was

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
68 A merican Anthropologist [64, 1962

adopted by Paharis, or some groups of them, from aborigines (often thought to

be the ancestors of low-caste Paharis) whom they presumably met and cul-
turally absorbed in this area. It may have been adopted as a result of contacts
with the polyandrous Tibetans. It could well have been a regional develop-
ment, probably in the western Himalayan hills. Its precise origins have been
obscured by time and are not now a fruitful subject for inquiry. More promis-
ing is the subject of the present functioning of polyandry and its economic
and social structural implications among those who practice it.

Demographic factors
In most discussions of polyandry, the possible influence of the sex-ratio
has been mentioned (Aiyappan 1935: 118; Majumdar 1944: 168; Prince Peter
1955b: 173 f.; Westermarck 1922:158 ff.), along with explanations to account
for any disparity of the sexes found in association with it (e.g., Rivers 1906:
520 f.). Heath (1955) has suggested that polyandry is generally related to a
shortage of women. The consensus of most modern writers is typified by
Kapadia (1956: 70) when he states that "sex disparity is likely to perpetuate,
though it does not necessarily give rise to, a polyandrous pattern."
Data on this subject from the Himalayan hills are suggestive but incon-
clusive. While North India shows a general surplus of males over females,
polyandrous Jaunsar Bawar has an unusually great shortage of females: 789
per 1000 males as compared to the Uttar Pradesh state ratio of 922.1n Adja-
cent nonpolyandrous Garhwal has a striking and very unusual (for India)
surplus of females: a ratio of 1110 in one district and 1149 in the other. These
contrasting sex ratios extend back as long as census figures have been avail-
able. The two small sub-districts of Garhwal (both adjacent to Jaunsar Bawar)
for which polyandry has been reported are the only parts of Garhwal in which
there is a relative shortage of women, with ratios of 942 and 965.
Thus, in the areas of immediate interest here there is a gross correlation
between polyandry and a shortage of women and, conversely, between mon-
andry and a surplus of women. From the point of view of explanation, the sig-
nificant fact is that in the Himalayas there is not an equal distribution of the
sexes among both polyandrous and monandrous groups as those who reject
the sex ratio as an explanation would expect. Neither is there a simple short-
age of women in the polyandrous areas in contrast to an equal distribution in
the monandrous areas, as those who consider polyandry to be an adaptation
to an unusual sex ratio might expect. Instead there is an unusual and unequal
sex ratio among both the polyandrous and monandrous groups, with the in-
equaltiy in each case apparently favoring the marriage system of that group.
Under these conditions one system cannot be considered prima facie to be
"natural" and the other deviant.
Figures for larger regions are more ambiguous. The entire Western Hi
layan area, throughout which polyandry has a scattered distribution, sho
consistent though (for North India) not an unusual surplus of males, w
the Central Himalayan region, where no polyandry has been reported, shows

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
BERREMAN] Pahari Polyandry 69

a more nearly equal distribution of the sexes. The latter may be a relatively
recent trend, however, as the proportion of women has increased quite steadily
from a ratio of 955 in 1901 to a ratio of 1019 in 1951.12

CONCLUSIONS

In describing and attempting to accou

rules of Jaunsar Bawar and Garhwal, a m
looked by previous commentators: that
personal connotations of the two system
In view of this fact and of the nonunive
practiced, the systems are not as differ
expected. Polyandry and monandry- in
polar types of marriage systems as has
was supposed at the initiation of this re
minor variations on a central theme, n
benefits to a group of brothers who ha
and who share other rights and property
lent, and show their unity as a group,
reproductive capacity (i.e., the "title"
other it is not. In both groups any one of
logical father of a particular offspring
father is shared; in Garhwal it is exclusive
polyandry and monandry in this area. It
values but a less drastic difference in the
been anticipated.
In view of the over-all similarity of Pahari cultures, contrasts between
polyandrous Jaunsar Bawar and monandrous Garhwal have not appeared as
clearly as was expected when the research began. However, some conclusions
pertaining to the original hypothesis can be stated:
Features of polyandrous societies reported in other parts of the world
correspond only partially with those found in Jaunsar Bawar. The Pahari
case contradicts the hypotheses that virtual economic uselessness of
women, and dowry or property rights held by women, are universal cor-
relates of polyandry.
As evidenced by a comparison of marriage and family relations in
Jaunsar Bawar and Garhwal, fraternal polyandry may be advantageous
but is not inevitable when there is a shortage of women and when a low
proportion of children in the family is economically advantageous.
"Equivalence of brothers" in economic matters and in relation to the
sexuality of their wives may be advantageous when, in a patrilocal society,
husbands leave their wives for extended periods, but there is no evidence
to show the superiority of polyandry over wife-sharing in such circum-
stances.
It seems logical, but could not be demonstrated by this comparat
study, that fraternal polyandry would be advantageous when cost of

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
70 American Anthropologist [64, 1962

taining or maintaining a wife is high, and also when in a patrilineal society

property upon which livelihood or wealth depends is unusually scarce
and limited. The latter is an advantage frequently cited by people of
South Asia who are fraternally polyandrous.
No evidence was adduced with regard to the psychological implications
of polyandry nor the value of polyandry as a means toward intra-family
cooperation and consensus. The latter is, however, widely held to be an
adjunct of polyandry and is a necessary one if the economic advantages
of unity of family property in such societies are to be fully realized.
Characterization of polyandry as an extension of the principle of
equivalence of brothers, especially in economic matters, is valid for Jaun-
sar Bawar as well as for other Himalayan hill areas and probably for Tibet.
It does not, however, characterize formal polyandry in contrast to fra-
ternal wife sharing. This is evidence by its applicability to the Garhwal
Paharis and to the South Indian Kotas, neither of whom allow a woman
a plurality of husbands or endow a child with more than one social father.
Both polyandrous and nonpolyandrous Paharis share a favorable attitude
toward the sharing of wives and property among brothers. This equivalence
of brothers may be a predisposing but not sufficient precondition for formal
polyandry. Certainly this attitude characterizes most fraternally poly-
androus people in South Asia.
The question of why Jaunsar Bawar people are polyandrous and Garhwal
people are not has not been answered. The answer undoubtedly lies in a com-
bination of cultural-historical factors, including the advantages which one
system may have relative to the other in a particular context (cf. Cooper
1941:55). Without going too deeply into conjectural history, some possibili-
ties may be considered.
Pahari culture functions satisfactorily under either polyandry or monan-
dry. Whatever the history of polyandrous and monandrous institutions of
Pahari marriage, they likely proved differentially advantageous among their
practitioners or potential practitioners. Each would presumably have per-
sisted most among those groups to which it proved most advantageous or
least disadvantageous. Advantages could take the form of the approval of
neighbors, economic well-being, social integration, etc.
Accurate historical data on the origins and contacts of Himalayan peoples
are lacking, and according to my evidence the distribution of polyandry as
contrasted with wife sharing in this region cannot be explained in terms of
associated economic or social structural features. It is therefore not unreason-
able to seek a partial explanation in the one apparently significant difference
which does appear between polyandrous and monandrous groups of the area:
the sex ratio. The sex ratio could have been a potent factor in the acquisition
and/or retention of one system or the other. For example, when external pres-
sure for abandonment of polyandry grew as a result of increasing administra-
tive, religious, and social contacts with people of the plains of India, polyandry

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
BERREMAN] Pahari Polyandry 71

persisted most in Jaunsar Bawar, the area in which the sex ratio favored it to
the greatest extent.
Of the several advantages which can be cited for polyandry or monandry,
a crucial one could have been the social and economic advantage which derives
from insuring the availability of family life for every adult. These are societies
in which it is difficult as well as almost unheard of to subsist without a family.
The sex ratio might tip the scale toward polyandry or monandry on the basis
of this advantage.
The weakness in this argument is that it depends on a disparity in the
sexes as an antecedent condition, and this cannot be demonstrated. Some ob-
servers claim that male Garhwal residents emigrate in great numbers to work
as servants on the plains. It is extremely doubtful that this occurs frequently
enough to account for the sex ratio, but no data are available with which to
verify or disprove the suggestion.?3 The same can be said of military service
as a possible explanation. Some claim that selling of Jaunsar Bawar women
to plains people has resulted in the shortage of women there, but this, too,
doubtless occurs too infrequently to account for the sex ratio. Moreover, in
Jaunsar Bawar, the ratio of the sexes among children is as uneven as that of
adults. This suggests as "causes" of the paucity of females, female infanticide,
for which there is no evidence; or neglect of female children, which is less
unlikely (cf. Majumdar 1944:171). Rivers (1906:520 f.) was among the first
to point out that such practices can as satisfactorily be attributed to the effects
of polyandry as they can be described as its causes.
To explain the origin or distribution of polyandry and monandry in the
area would therefore require data which are not available: culture history and
census data from earlier eras. The futility of seeking causes without knowledge
of the attendant conditions is well known.

Why there is polyandry in Jaunsar Bawar and not in G

not a question that is likely to be answerable now, or th
very relevant. A comparable question would be that of
people speak Western Pahari while Garhwal residents s
These are relatively minor differences; the culminations
contacts, and of drift from a common base. They have
choices over considerable periods of time. The choices ha
the cultural context of economic and social equivalence of
contractual nature of marriage wherein a bride is "purc
into which she marries. Both of these conditions are compat
polyandry. The choices which have led to regional differenc
terns have been made in response to conditions (limitat
ties) within and without the groups involved directed in
haps, by certain advantages which followed from them
conditions which influenced them are now largely unknown
research there are in the Himalayan hills no simple func
polyandry as contrasted to monandrous wife-sharing ex

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
72 American Anthropologist [64, 1962

sex ratio is so correlated. The important correlations are those of specific cul-
tural content: polyandry is one feature of an over-all cultural pattern of the
Western Himalayas which contrasts in a number of details with the over-all
pattern of the Central (and probably the Eastern) Himalayas, one feature of
which is the absence of formal polyandry. The present distribution of these
patterns is apparently the result of regional divergence from a common and
relatively homogeneous culture; a divergence made possible in part by relative
regional isolation.?4 The same processes which resulted in divergence of such
features as language, dress, and worship facilitated the present distribution of
marriage regulations (Berreman 1960). Therefore, regional variation in mar-
riage regulations is no more fundamental nor surprising than other cultural
differences in these hills and is to be understood as being of approximately the
same order.

Observers feel compelled to question and explain polyandry wherever it

occurs because it is unusual. One might equally fruitfully question the occur-
rence of polygyny. The factors leading to one are probably no more consistent
and compelling than those that lead to the other. Polyandry, like polygyny, is
evidently not a sufficiently unitary phenomenon to be explained in the same
terms everywhere.?5 It may have certain advantages or functionally related
correlates in some areas and not in others. That they are not universal does not
mean, of course, that they are not significant.
There may be conditions which correlate with fraternal polyandry on a
widespread cross-cultural scale. However, these probably take the form of
effects of the functioning of polyandry, or prerequisites for polyandry, rather
than specific causes which inevitably lead to polyandry.

NOTES

1 This paper was read in abbreviated form before the Fourth Annual Meeting of the Kroeber
Anthropological Society, May 21, 1960, at Berkeley. The research was carried out in India during
1957-58 under a Ford Foundation Foreign Area Training Fellowship and is reported in full in:
Berreman 1959. I would like to thank David Mandelbaum for his helpful comments on the
manuscript.
2 Following the recent definition of marriage put forth by Gough (1959:32), a husband may be
defined as a person who is in a relationship to a woman such that a child born to her under circum-
stances not prohibited by the rules of the relationship is or may be publicly acknowledged to be
that person's child and is accorded full birth-status rights common to normal members of the
society or social stratum into which it is born.
3 Polyandry has been reported in the Rawai and Jaunpur sub-districts of Tehri-Garhwal,
immediately adjacent to Jaunsar Bawar (Kapadia 1955:63). Those portions wherein polyandry
is found are doubtless the western border areas which fall into the Western Pahari sub-culture
area, or on its peripheries.
4 By Garhwal, I mean the districts of Tehri-Garhwal, Garhwal, and the hill sections of eastern
Dehra Dun district (other Central Pahari districts are Almora and parts of Naini Tal). The re-
search reported here was in a hill area of western Garhwal, overlapping Tehri-Garhwal and Dehra
Dun districts (Berreman 1960). The area can legitimately be lumped with Garhwal because its
residents are culturally of Garhwal. Their ancestors came from interior Tehri-Garhwal, they con-
sider themselves to be Garhwalis and are so considered by others. Generalizations in this paper
about Garhwali marriage and family relationships are valid for western Garhwal and only in-
ferentially for the rest of Garhwal.

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
BERREMAN] Pahari Polyandry 73

i Note that a plurality of husbands constitutes polyandry and the number of wives is consid-
ered irrelevant in the definition implicit here. Most discussions of the advantages of polyandry
imply either that only one wife is involved or at least that husbands outnumber wives.
The incidence of polyandry reported by Majumdar for this village is high if, as seems probable,
he is referring to conjugal family units. If all sets of adult real brothers currently living in the Garh-
wal village I studied had formed fraternally polyandrous conjugal families, then 43 percent of all
conjugal families in that village would have been polyandrous. This is comparable to the propor-
tion of polyandrous families reported by Majumdar. Fraternal polyandry would therefore seem
to be the preferred pattern of marriage in Majumdar's village, with an incidence about as high as
possible. Monogamy probably occurs most often among men with no brothers. Plural wives are
probably secured in either case primarily to remedy a shortage of labor or heirs in the family, as
they are in Garhwal.
6 1 witnessed one Garhwal marriage in which an elder brother substituted for the groom. This
arrangement was devised to avoid the consequences of incompatibility in the horoscopes of the
intended bride and groom rather than to effect a polyandrous union. The intended husband took
over after the ceremonies. One might speculate upon a polyandrous precedent for this device but I
could find no evidence to support such a speculation. More probably this incident reflects the
general equivalence of brothers in Pahari culture. David Mandelbaum has pointed out that it
illustrates not only the ritual and social equivalence of brothers, but also their personal non-
equivalence in relation to the supernatural. Nothing could be more personal than the horoscope
and in that respect the brothers were significantly not equivalent.
7 Majumdar (1944:173 ff.) and Kapadia (1955:73, 83) have argued rather unconvincingly
that high-caste Paharis were once matrilineal or heavily influenced by matrilineal people, evidently
in the belief that this is more compatible with polyandry than is a purely patrilineal tradition.
This is in line with the belief of McLennan and others that polyandry is associated with matrilin-
eality. Leach (1955:183), who hypothesizes that inheritance of property through females as well as
through males is consistently associated with polyandry, implies that only patrilineality of "an
ambiguous and rather uncertain type," and not "patrilineal systems of the more extreme type,"
can be associated with polyandry. Although Pahari patrilineality is not extreme, it is so with
regard to inheritance of property. Leach's hypothesis is not supported by my research nor by the
evidence presented by Majumdar or Kapadia.
8 An exception, according to Prince Peter (1955b: 171 f.), are the Todas. He asserts that they
share no property in the family. However, Rivers (1906:558 ff.) describes the house as specifically
belonging to a group of brothers who share a wife, and he mentions that although buffaloes are
largely individual property, "in practice, owing to the fact that brothers usually live together, a
herd of buffaloes is treated as the property of a family of brothers, butwhenever the occasion arises
there are definite rules for the division of the buffaloes among them." Such rules are undoubtedly
to be found in all polyandrous societies, as they are among the Paharis.
9 The Tibetan evidence, too, contradicts this as a general explanation of polyandry. Car-
rasco (1959:35, 68) describes the important and productive economic role of women in Tibetan
society. It also seems doubtful, according to his data (Carrasco 1959:36 f.), that Tibetan women
are invested with property rights frequently enough to support Leach's hypothesis concerning the
relationship between such rights and polyandry.
10 This explanation undoubtedly contains a large element of rationalization. It serves to pre-
serve an ideal of fraternal amity in the face of a good deal of actual fraternal strife by blaming it
on wives who are essentially outsiders in the family and who most often come from alien villages.
11 Sex ratios given here are figured as they are in the Census of India, i.e., number of females
per 1000 males. This is the reciprocal of the usual ratio given in the United States Census.
All figures are for rural areas, i.e., excluding towns of over 5000 population in most cases, and
are drawn from various volumes of the 1951 Census of India.
i In nonpolyandrous and almost entirely non-Pahari Dehra Dun district adjacent to both
Jaunsar Bawar and Garhwal, the shortage of women (ratio of 759) is even greater than in Jaumsar
Bawar, and in nonpolyandrous Naini Tal to the east the sex ratio is only 728. These two areas
border on the hills, but their populations are largely derived from the plains. Dehra Dun, at least,

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
74 A merican A nthropologist [64, 1962
has been relatively recently settled and the sex ratio is affected by the presence of tea plantations
and other innovations atypical of the hill areas.
13 In the immediate area of my research there was neither a surplus of women nor a signifi-
cant amount of out-migration by men.
14 This divergence may have been of polyandry from a monandrous base in the Western
Pahari area, or of monandry from a polyandrous base in the Central and Eastern Pahari areas. The
change need not have been a difficult or disorganizing one in view of the over-all similarity of the
cultures involved and the apparent compatibility of both polyandry and monandry in these cul-
tures.

Gough (1952:86) records a case of significant structural change without discontinuity in

South India: "the Nayar system has, over a period of two hundred years, changed from a very ex-
treme form of matriliny into a 'bilateral' system with only a weak tendency to matriliny; but the
latter system developed imperceptibly out of the former."
Prince Peter (1955c:183) notes that in Ladakh, Muslim converts dropped polyandry almost
overnight, apparently without seriously affecting other aspects of their culture.
16 The following parallel comment by Westermarck (1922: 206) was discovered by the author
after this article was in press: "To explain in full why certain factors in some cases give rise to poly-
andry and in other cases not is as impossible as it often is to say exactly why one people is monog-
amous and another people polygynous."

REFERENCES CITED
AIYAPPAN, A.
1935 Fraternal polyandry in Malabar. Man in India 14:108-18.
1937 Polyandry and sexual jealousy. Man 37:104.
BERREMAN, G. D.
1959 Kin, caste, and community in a Himalayan hill village. Doctoral dissertation, Cor-
nell University.
1960 Cultural variability and drift in the Himalayan hills. American Anthropologist
62:774-94.
BRIFFAULT, ROBERT
1959 The mothers. London, George Allen & Unwin Ltd.
CARRASCO, PEDRO
1959 Land and polity in Tibet. Seattle, University of Washington Press.
COOPER, JOHN M.
1941 Temporal sequence and the marginal cultures. The Catholic University of America
Anthropological Series, No. 10.
DAS-GUPTA, H. C.
1921 A short note on polyandry in the Jubbal State (Simla). The Indian Antiquary
50:146-48.
FISCHER, H. T.
1952 Polyandry. International Archives of Ethnography 46:106-15.
GOUGH, KATHLEEN
1952 Changing kinship usages in the setting of political and economic change among the
Nayars of Malabar. Journal of the Royal Anthropological Institute of Great Britain
and Ireland 82:71-87.
1959 The Nayars and the definition of marriage. Journal of the Royal Anthropological
Institute of Great Britain and Ireland 89:23-34.
GRIERSON, GEORGE A.
1916 Linguistic survey of India, Vol. IX, Part IV. Calcutta, Superintendent of Govern-
ment Printing.
HEATH, DWIGHT B.
1955 Sexual division of labor and cross-cultural research. Paper read at the Fifty-fourth
Annual Meeting of the American Anthropological Association, Boston. Refrenc in

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms
 BERREMAN] Pahari Polyandry 75
The world of man, by J. J. Honigmann,
374.
JOSHI, L. D.
1929 The Khasa family law in the Himalayan districts of the United Provinces of India.
Allahabad, The Superintendent, Government Press.
KAPADIA, K. M.
1955 Marriage and family in India. London, Oxford University Press.
KmRKPATRICK, G. S.
1878 Polyandry in the Punjab. The Indian Antiquary 7:86.
LEACH, E. R.
1955 Polyandry, inheritance and the definition of marriage. Man 55:182-86.
MAJUMDAR, D. N.
1944 The fortunes of primitive tribes. Lucknow, The Universal Publishers Ltd.
1953 Children in a polyandrous society. The Eastern Anthropologist 6:177-89.
1955a Family and marriage in a polyandrous society. The Eastern Anthropologist 8:85-
110.
1955b Demographic structure in a polyandrous village. The Eastern Anthropologist 8:161-
72.
MANDELBAUM, D. G.
1938 Polyandry in Kota society. American Anthropologist 40:574-83.
MUKHERJI, ANtA
1950 The pattern of polyandrous society with particular reference to tribal crime. Man in
India 30: 56-65.
MuNSm, K. M.
1955 Foreword. In Social economy of a polyandrous people, by R. N. Saksena. Agra
Agra University Press.
PRINCE PETER OF GREECE AND DENMARK
1948 Tibetan, Toda, and Tiya polyandry: a report on field investigations. Transactions
of the New York Academy of Sciences, Series II, 10:210-25.
1955a Polyandry and the kinship group. Man 55:179-81.
1955b The polyandry of Ceylon and South India. Actes du IV' Congres International des
Sciences Anthropologiques et Ethnologiques, Vienne (1952) 2':167-75.
1955c The polyandry of Tibet. Actes du IVe Congres International des Sciences Anthro-
pologiques et Ethnologiques, Vienne (1952) 2':176-84.
RADCLIFEE-BROWN, A. R.
1941 The study of kinship systems. Journal of the Royal Anthropological Institute of
Great Britain and Ireland 71:1-18.
RIrERS, W. H. R.
1906 The Todas. London, Macmillan and Company Ltd.
SAKSENA, R. N.
1955 Social economy of a polyandrous people. Agra, Agra University Press.
STULPNAGEL, C. R.
1878 Polyandry in the Himalayas. The Indian Antiquary 7:132-35.
WESTERMARCK, EDWARD
1922 The history of human marriage. Vol. III. New York. The Allerton Book Company.

This content downloaded from 103.18.0.20 on Thu, 04 May 2017 04:53:56 UTC
All use subject to http://about.jstor.org/terms