Académique Documents
Professionnel Documents
Culture Documents
in South India
The KaikkooLars of Tamilnadu State, South India, are a caste of the left-hand section liv-
ing in a social world that i s divided between left and right. Right-hand and left-hand sym-
bolic classification, a worldwide phenomenon (Needham 1973), is perhaps nowhere more
elaborate than in India. Although a fading memory in the Salem-Erode region of Tamilnadu
where I conducted fieldwork, this moiety division of castes into right and left (vaLangkail
iiDangkai jati) continues to influence caste behavior. The division’s character lends support
to the idea that the social structure of some civilizations i s dichotomized by the different
loci of power on which the artisan-merchants (the left-hand) and the landowning communi-
ties (the right-hand) base their strengths (see Beck 1970; David 1974; Stein 1980). Brahmans
do not belong to either section; but as priests, they occupy a ritual status that is distinctive
and superior.
There is much about the KaikkooLars, or Sengunthar Mudaliyars, that makes them
distinctive. Some of their features call into question current models of caste. Over the past
two decades, Marriott’s (1959,1968) transactional model and Dumont’s (1970b) dual model
have dominated discussion of Indian caste ranking. Both models depict castes as in-
tegrated, interdependent, unitary hierarchies; caste ranking i s seen as essentially a
homogeneous system in which all castes, from top to bottom, participate. Even at the level
of the excluded bottom-the untouchables-Moffatt (1979) demonstrates replication of
the system. It appears, however, that there are Hindu castes that do not f i t either model.
Heesterman (1973:99-100) notes that historically, castes of the right-hand section fit the
hierarchical, integrated model of jati stratification, but castes of the left-hand section do
not. Heesterman argues that the right-hand section is characterized by what Dumont calls
the kingly model, in which land-based power and interdependence articulate the hierarchy.
The left-hand section i s characterized by the priestly model, a stress on purity, and in-
Historical/y, castes in South India were divided into right-hand and left-hand
sections. This paper is an examination of the KaikkooLars, a caste of weaver mer-
chants belonging to the left-hand section. In many features this caste is distinc-
tive, the source of i t s uniqueness being the caste‘s social and economic
marginality that allows i t s members a degree o f separateness from the agrarian-
based village. The implications of this separateness for theories of caste are ex-
plored. It is argued that because anthropologists have concentrated their
research on the agrarian village, they have widely overlooked mercantilism as a
locus of power in rural India. [caste, right-hand/left-hand division, dual organi-
zation, symbolism, mercantilism, India]
Who are the KaikkooLars and to what extent are they part of an integrated, interdepen-
dent caste hierarchy, as Marriott’s transactional model of caste ranking depicts? The Kaik-
kooLars are a large, ancient (see Stein 1968:87; 1975:75; 1980:183-184) weaver caste living
in Tamilnadu State, South India. The term KaikkooLar refers to their occupation as weavers
(kai [hand], kool [shuttle], ar [people]), but they are also moneylenders and textile mer-
chants. They live throughout the state but can be found in particularly large concentrations
in certain cities and towns that are centers of weaving. The Salem-Taramangalam-Trichen-
gode-Erode region of the Salem and Coimbatore districts, where the fieldwork on which
this paper i s based was conducted, i s one such area. Although many families and certain
segments of the caste are vegetarians, most KaikkooLars are meat-eaters and make blood
sacrifices of goats, chickens, and pigs t o lineage gods. Vegetarians and nonvegetarians in-
termarry.
Such behavior i s generally taken as an indication that at least in Brahmanical terms, the
caste ranks low. That many eat the flesh of the pigs they sacrifice seems to confirm this.
Pig-eaters are untouchables, according t o the reasoning of most South Indians. Their low,
even untouchable, status seems confirmed by historical research as well. For example,
Frykenberg 1977; 1978:personal communication) believes that the KaikkooLars were
untouchable in mid-19th-century Tinnevelly because they were forcibly excluded from us-
ing the streets of the landed and dominant Vellalas. Another indication of low status was
the custom of dedicating daughters as deevi-daasis to Shiva temples, as recently as the ear-
ly 20th century. Since prostitution was associated with this practice, the custom is con-
sidered a sign of loose morals and low status.
Although the KaikkooLars appear t o have been untouchables in some regions, in others
they seem t o have occupied a more exalted but s t i l l low status. Today, KaikkooLars are
staunch Shivites, but Stein (1968:87) notes that they had “the ritually pregnant task of offer-
ing coconut t o the deity” at the Srirangam temple (a Vishnivite temple at Tiruchirapalli) 600
to 900 years ago. KaikkooLars commonly appear to be juxtaposed to the dominant land-
owners of their region, who are of the right-hand section, but t o have little in-
terdependence. At Srirangam, “the temple accountant [koyil kanakkan] was a member of
the dominant Sudra agricultural community of Tamil country, the Vellalas.. .” (Stein
The Sengunthar Mudaliyars are thus not directly connected to the land, and their overall economic
position in Endavur is not strong. Despite its relatively high rank, the caste in no way shares in the
dominance of the village; its only potential effect on village-level politics is in the sheer number of
votes its members can bring to village-level elections. Nor has it any strong vertical relations by
which it can enforce authority over the Harijans; there are no Harijans in paNNaiyaaL [traditional
subordinate labor tie with a member of a dominant caste] tied-labor relations to members of the
Sengunthar Mudaliyar caste.
The KaikkooLars trace their origins t o Shiva, Parvati, and Murugan. When Shiva i s
creating Murugan, his consort Parvati is frightened. In her flight, nine jewels fall from her
ankle and take the form of nine beautiful women, whom Shiva impregnates with his gaze.
Parvati, enraged with jealousy, curses the nine women so that the infants remain unborn so
long that they grow mustaches, a sign of virility and of their future warrior status. Eventual-
ly, Parvati relents and the nine original KaikkooLar warriors (navavirahaL) are born.
KaikkooLars see Murugan as their caste’s primary god. In early sankam poetry, Murugan
i s depicted as having an ambivalent character-on the one hand benevolent, while on the
other, cruel and fierce. He possesses women and is worshiped with the sacrifice of goats
and cocks, with the use of intoxicants, and with frenzied dance. He is a hunter who
becomes a warrior. Later, with the influx of northern influences, Murugan becomes
Brahmanicalized and i s depicted as a son of Shiva. He i s a peaceful god, as Shiva is, but he
i s also a warrior who vanquishes Surabatma, an evil giant. According t o the myth, Shiva i s
plagued by Surabatma, who i s imprisoning gods. Shiva wishes him destroyed and asks
Murugan to do so. Murugan enlists the aid of a general, Virabahu, whose army is composed
of the nine original KaikkooLars. Together they pursue Surabatma and repeatedly cut off
his head, only to have him sprout another. Eventually, Surabatma has only one head re-
maining and hides by taking the form of a tree. But Murugan detects him and splits him in
two with his spear. From the two halves emerge the cock and the peacock, which Murugan
takes as his emblems, along with the spear.3
The KaikkooLar use the symbols of the cock and the spear on their caste flag to repre-
sent their ties with Murugan. They also call themselves Sengunthars-red (sen), spear
(kuntappadai), people (ar)-again in affirmation of their ties with Murugan, who is known as
a red god. In other words, when KaikkooLars answer the question ”Who are we?“ they do so
by saying they are Murugan’s people (see, for example, the caste memorium published for
the 1976 Erode Meeting).
The KaikkooLars are unaware of the cross-fertilization of their traditions, which is evi-
dent in the Tamil and Sanskritic origins of their beliefs. They are staunchly anti-Brahman
and pro-Dravidian in their attitudes, while simultaneously firmly identifying themselves
with the Sanskritic Shivite tradition and with a Brahmanicalized Murugan. Nonetheless,
there are striking parallels between the early worship of Murugan and the KaikkooLar wor-
ship of small gods and certain beliefs. KaikkooLars perform blood sacrifices to small gods,
and some KaikkooLar women foretell the future while in a state of trance when possessed
by a small god. In terms of personal interests and daily life, these small gods are more im-
portant to worshipers than are either Murugan or Shiva, because they affect the worshipers’
immediate lives and are worshiped in the context of lineage and family; Murugan and Shiva
do not affect lives so immediately. KaikkooLar mantravadis (black magicians) use potions
and matrams t o benefit and curse others, and they also use intoxicants. Until recently, Kaik-
kooLars dedicated women dancers as a feature of their worship of Shiva and Murugan. The
parallels of these customs with early Murugan worship are so striking that it is not too in-
cautious t o suggest that many features of early Murugan worship are s t i l l carried on by
KaikkooLars in the context of this second locus of the sacred. KaikkooLar beliefs and iden-
tity, therefore, are split between two traditions: Brahmanical Shivism and non-Brahmanical
ritual practices.
KaikkooLars do not exhibit a homogeneous religious ideology; theirs embodies an inner
conflict that i s the result of the bifurcation of the sacred in Tamil society into two chief
Identity i s in some ways a process, ongoing and constantly changing. As society changes
and its place in society changes, a caste modifies i t s image of who it is. This process com-
Beck (1970, 1972), in her elaboration of the right-handlleft-hand division, argues that
there are two prestigious models for behavior in the South Indian village: the kingly model
associated with landownership and economic power and the priestly model associated with
the Brahman who always occupies the ritual apex of the caste hierarchy. Lacking land, she
argues, left-hand castes follow the Brahmanical model for behavior. Beck notes, however,
that low left-hand castes tend to combine the two models, a reflection of their own am-
bivalence as to which model to follow, since they rank high in neither. KaikkooLars, ac-
cording t o her reckoning of ritual rank, are the highest of the left-hand castes (Beck
1970:786). If Beck’s logic were correct, the KaikkooLar should closely follow a Brahmanical
mode; but according to her analysis, their social customs are mixed. Indeed, five of the
seven left-hand castes have this mixed background. Only the Cooli Aacaari and the Komut-
ti Cettiyaar are “Brahmanical.” Beck’s ordering, therefore, does not tolerate close scrutiny.
Another weakness of her argument i s the manner in which she measures kingly and
priestly custom. Beck examines ten customs, five associated with the local landowning or
kingly caste and five associated with the local Brahmans, the priestly caste, taken to repre-
sent kingly and priestly custom, respectively. She scores castes according to the degree to
which they hold each of the ten customs and derives a scaling of adherence that she plots
on a continuum between the kingly and priestly poles. Beck seems to be correct in her
recognition of the importance of the customs she uses to determine her scaling. However,
her technique obscures the importance of a local tradition that i s revealed only when the
values and views of individual castes are examined. Beck’s model does not determine
whether the left-handlright-hand distinction is basically a priestly-kingly contrast; instead,
it polarizes the sections as such. In fact, the dichotomy is not so clear cut, as we have
noted.
When ideology and behavior are considered, Beck’s polarization i s greatly weakened,
because, as we have seen, the KaikkooLars of Salem-Erode (also part of KonkunaaDu, the
region in which Beck worked) are definitely not priestly in several contexts. Beck’s model is
too simple to account for the apparent complexity, despite i t s appealing neatness. In her
book, Beck (1972:266) recognizes this ambivalence but does not attempt to explain it.
Finally, there i s the Mutaliyar [KaikkooLar] community, a relatively high-status group which is
caught squarely i n the middle of these right-left division differences. Their ancestry is martial, and
although members of this caste officially belong t o the left, they have come t o identify with the
right in many respects. Thus the Mutaliyar can be seen to vacillate between the two poles of social
organization outlined in this work. In a sense they may be seen as striving for a neutral position like
that of the Brahmans and the PiLLai, both of whom are generally agreed t o stand above this exten-
sive bifurcate structure.
Because she emphasizes the essentially Brahmanical character of the left-hand division,
her model i s troubled by exceptions, the most apparent of which are the Mudaliyar.
The explanation I propose has two parts. First, the ambiguity in custom that Beck notes
among left-hand castes occurs not because of confusion about which model t o follow, but
because their sacred cosmos has two distinct loci that require different forms of worship.
The exterior locus i s Sanskritic and Brahmanical, and so i s KaikkooLar worship of Shiva and
Murugan. Sengunthar Shivite priests are vegetarian, wear the sacred thread, and shave their
foreheads in the Brahmanic fashion. Both alcoholic and sexual abstinence are valued, as i s
control of the passions. But when KaikkooLars are concerned with the sacred locus of the
interior, meat eating, blood sacrifice, spirit possession, and the worship of small gods are all
prominent. The KaikkooLars thus follow both a priestly model and a Dravidian tradition.
This latter tradition i s distinct from the kingly model because the power to command is not
The "armies" are long gone, of course, but they indicate that weavers were to some extent
beyond the control of the dominant, land-based powers. The traditional occupation of
weavers i s not just weaving, but also textile trade. As traders and merchants, weavers move
among villages and towns even today, as Beck (1972:266) notes. Consequently, they are
neither economically confined to a single village nor fully dependent o n dominant
agricultural castes. They are nondominant in the agricultural village, but they have a
degree of separateness or social marginality; they are not so readily dominated as a conse-
quence. Just as the local tradition i s marginal t o the Brahmanic one, the KaikkooLar i s
marginal to the kingly power of agrarian castes.
This economic marginality, seen from the merchants' point of view, stems from their
distinct locus of economic power, the market network. This economically based distinc-
tiveness i s important for explaining the appearance of KaikkooLar untouchable status in
mid-19th-century Tinnevelly (Frykenberg 1977). The KaikkooLars were met with violence
when they attempted t o take a funeral procession through a Vellala street, as though they
were untouchables attempting to enter an area reserved for clean castes. Yet I think that
the assessment of their untouchable status i s wrong for several reasons. First, KaikkooLars
are not presently considered or remembered as untouchables, Second, their historical role
in the important Srirangam temple in ancient times in Tiruchirapalli disallows untouch-
ability. Third, KaikkooLar caste organization i s complex, regional in extent, and very old,
apparently dating back at least to Vijayanagar times (by contrast, evidence we have of un-
touchable caste organization is strikingly local, e.g., Moffatt 1979:57).' Fourth, at least in
some regions, members of neither the right-hand nor the left-hand were permitted to enter
the streets allotted to the other with their marriage, funeral, or festival processions (Ap-
padurai 1974:239); consequently, exclusion alone cannot be taken as a sign of untouchable
status Rather, exclusion may be an outgrowth of the chronic antagonism that existed be-
tween the two sections, of which the KaikkooLars and Vellalas were opposing members.
Marginality offers a better explanation for KaikkooLar exclusion. KaikkooLars are out-
side the landowner-dominated caste hierarchy, yet they demand rights equal to those of
high status (Frykenberg 1977:6). Vellalas responded by attempting domination, a standard
high-caste approach to maintaining control. However, because KaikkooLars are socially
marginal to the economic and social relationships normally used by landowners to
dominate other castes, the Vellalas resorted to exclusion, enforced by physical violence.
Moffatt (1979:72) notes that even untouchables can treat Sengunthars with contempt,
despite their relatively high status, because of their nondominance and their nonagricul-
tural occupation. Marginality, therefore, cuts both ways: the KaikkooLars have few
economic dependents and, consequently, little power over anyone; but, they are also large-
ly free of the economic control of landowners. Marginality more easily explains the Tin-
nevelly situation than does the assumption that KaikkooLars were untouchables. Kaik-
kooLars as marginals, however, are structurally similar t o untouchables who, as structural
inferiors, occupy a kind of marginality (see Turner 1969:125ff.; 1974:237), since marginality
can only be maintained through the use of withdrawal, exclusion, and force.
Appadurai (1974). in his analysis of the left-handlright-hand division of castes, has of-
notes
Acknowledgments. Fieldwork on which this paper is based was conducted from September 1978 t o
August 1979, funded by the American Institute of India Studies. The author wishes t o thank Donald E.
Brown and David Brokensha for their helpful readings of an earlier version of this paper.
’ M y elderly informant, a KaikkooLar, clearly does not know all of the right-hand castes and men-
tions only the main ones; his l i s t of left-hand castes i s also vague with respect to Chettiars, since there
are several kinds. It i s worth noting that the division of barbers into two sections continues t o some ex-
tent in the Erode region. The two groups still distinguish between themselves.
* The error in Heesterman’s logic i s that while Brahmans are ritually ranked high, they are none-
theless involved in interdependent economic relationships as priests. Consequently, they remain
dependent on landowners.
This battle i s annually dramatized in Ammapet‘s streets by KaikkooLars, following Deepavali.
’ It has been suggested that j a t i mobility among artisan-merchant castes is easier precisely because
of their independence from specific patrons. An untouchable past (see Pocock 1962:85) might not be
incompatible with their present status were it not for their participation in the Srirangam temple and
their extensive caste organization. ”Dangerously ambiguous” i s a better assessment of their status,
since they are so unlike landowner-dominated untouchables.
references cited
Appadurai. Arjun
1974 Right and Left Hand Castes in South India. Indian Economic and Social History Review
111216-259.
Arokiaswami, M .
1956 The Konku Country. Madras. University of Madras.
Beck, Brenda E. G.
1970 The Right-Left Division of South Indian Society Journal of Asian Studies 29:779-798.
1972 Peasant Society in Konku. Vancouver: University of British Columbia Press
Buchanan, Francis
1807 A Journey from Madras through the Countries of Mysore, CaQara, and Malabar. 3 vols.
London: W. Bulmer and Co.
Census of India
1961 Temples of Madras State. Coimbatore and Salem. Vol. 9, Part 11-D. Government publica-
tion.