Back to the Academic Staff College Home Page

Hierarchy and Difference:
An Introduction


From Dipankar Gupta (ed.)Social Stratification, 1991, Delhi: Oxford University Press, pp 1-21.

Reproduced by permission of the author. This etext is meant to be used only for academic and non-commercial purposes.

Scanned and proof read by Amman Madan at the Academic Staff College, JNU, 17/9/1998


The Idea of Stratification and the Caste System

Social stratification has a special place in the study of Indian society.
India has long been reckoned as the most stratified of all known socie-
ties in human history. The caste system with its myriad forms of super-
ordination and subordination, its many customs and taboos, is perhaps
most responsible for conferring on India this dubious honour. But this is
not all. Economically too India is highly stratified. Miserable slums
border expensive residential areas in city after city in India. The inde-
scribable poverty of the very poor has even led to a review of the limits
of physical endurance at pitifully low nutritional levels. This vast polar-
ity notwithstanding, India is also a significant economic power with a
sizeable bureaucracy and technically trained personnel. Add to this the
diversity of linguistic groups that make up our Indian nation state and
the fact of India being the most stratified society becomes near incontro-

India is also a very self-conscious society. There are endless debates
in India on what should be the path of development, and what internal
arrangements of power and wealth, of cultural status and economic
wherewithal, are best suited to propel the country into the modern,
industrial epoch. As a people Indians have been deeply involved in
moral and ethical questions regarding the caste system, cultural diver-
sity and economic inequality--all central issues of social stratification.
This is reflected in our Constitution which makes any discrimination


based on caste, language, religion or creed illegal. Clearly the found-
ers of independent India had pondered deeply over the cardinal fea-
tures of social stratification in our society.

Very often when we talk of social stratification in India we concen-
trate almost exclusively on the caste system. The uniqueness of this
outstanding institution has captivated sociologists and anthropologists
for generations. It would be hard to think of a sociologist working on
India who has not written or commented extensively on it. Quite natu-
rally, with all this literature, some of exceptional quality, discussions
on the caste system tend to subsume the entire field of social strati-

The Visibility Postulate
But social stratification includes a lot more. The fact that the caste
system is seen as an example par excellence of social stratification,
gives an indication of the specificity of the term and the range it can
include. The caste system, as it is understood widely, separates and
hierarchizes Hindus. However, it is not sufficient if this separation and
hierarchization are wholly internalized or intellectualized. It is only
when hierarchy and differences are externalized and socially demon-
strated that we can truly talk about social stratification. Rituals, dress,
tonsorial styles, marriage practices, and a host of other such phenomena
help in socially separating one caste from another. It is these phenom-
ena too that are appropriately valorized for the purposes of hierarchi-
cal ranking. It is for this reason, that when we talk of social stratification
we not only mean differentiation but differentiation that is made
socially visible. It is not just stratification but social stratification. In
other words there is a general acknowledgement within society of the
social markers that separate the population, and an awareness also of
the crucial criterion(sometimes a set of criteria) on which such
forms of differentiation are based (see Beteille 1977:4,9,40-1).
Social stratification then deals with the ways in which the human popu-
lation is socially differentiated, i.e. differentiated publicly and demon-
strably. The criterion for differentiation may be one but the social dis-
play of differentiation usually includes a host of factors. The principal
criterion on which the caste system is based is the principle of natural
Natural superiority in this case is not physical prowess or
intelligence, though these often work their way in, but the endowment
of bodily purity. It is a known fact that there is no unambiguous physi-
cal criterion by which individuals can be differentiated on the basis of


the extent of purity of their bodies. This is why it is essential that social
practices, occupations, life styles, rituals and taboos demonstratively
differentiate one caste from another for all to see.

Even in cases where there are clear biological differences such as
sex or race, these differences are not retained in their natural form
when we include them under the rubric of social stratification. Social
stratification is not satisfied with biological differences per se. These
biological differences must be socially amplified with respect to
dress, or food, or occupation, or residence, or mobility, or a combina-
tion of all these and more. Differences in race or sex become important
for social stratification because of the modalities by which the social
lives of people belonging to different sexes or races are socially sepa-
rated and distinguished. Those biological differences that are not thus
amplified upon become socially irrelevant and do not factor in the
reckoning of any system of social stratification.

That human kind everywhere demonstrates actively this propensity
to differentiate may seem a rather trivial and pointless failing (and
some orders or differentiations are arguably pointless), yet a closer
look will tell us that social stratification manifests itself in almost
every aspect of social life--even in the most intimate ones. The fam-
ily, the school, the office, the neighbourhood, all are marked deeply
by internal divisions of authority, wealth, or status; or language, culture
and customs. As a matter of fact, one might even say that order mid
coherence in a society, or in any of its aspects thereof, eventually rest on
its system of social stratification (ibid.: 17).

Naturally a lot depends on what aspect of social life we are inter-
ested in for our analytical pursuits. Need one be reminded that social
reality is diverse and no one factor can serve as a durable key to its many
secrets? Social stratification too is just one aspect of this multifaceted
social reality, but it is a factor that weighs rather heavily in politics, in
economics, and in moral considerations of right and wrong. Politicians
from the earliest of times have sought to transform or better existing
systems of unequal power distribution; economists--even before univer-
sity economists arrived--talked at length about the rich and the poor;
and of course, there have been renowned thinkers, philosophers and
men of religion, who have pondered over the questions of inequality
and social differences in order to ferret out their inner essence. It is for
this reason that a confrontation with the issues of social stratification
at the intellectual level is inescapable. Quite appropriately some of the
greatest sociologists have in the last century contributed significantly to
our understanding of social stratification (see Appendices).


Need for Conceptual Clarity
One must not forget that it is not at all the case that a society should
exhibit only one form of social stratification. In India, for instance, the
extant forms of social stratification are many. There is of course the
caste system, but even this 'extreme form of social stratification'
(Dumont 1988:3) coexists with occupational stratification, linguistic
stratification, sexual stratification and religious stratification (to name a
few). It is important for sociologists to remember that each of these
forms of stratification have their own axial principles. It would do us
no good if we were to be careless on this score. Any carelessness or
untidiness in this matter would lead to quite basic conceptual difficul-
ties. For instance, the oft asked query of whether caste is giving way to
class, is an outcome of conceptual fogginess. There is no reason to
believe that if there is caste there cannot be class, nor is it the case that
as one grows the other must wane. We should not forsake an ele-
mentary methodological tenet namely that a concept should be inde-
pendently defined.
Caste and class after all do not constitute a

The important point to bear in mind is that the various forms of social
stratification are analytically separate and separable. Empirically we
often find one form of stratification overlaid by another. Gender strati-
fication may correspond with economic stratification, class and
caste may demonstrate significant statistical correlation, or linguistic/
regional groups may show a great degree of co-variation with occupa-
tional stratification. This should not tempt us to conflate or submerge
one category of stratification with another. The co-variation between
two or more forms of stratification asks for a higher order of explana-
tion, and not the abandonment of one for the other, e.g. caste for class,
or class for caste.

The lines of social stratification in India are so deep and vari-
egated that their uniqueness often overwhelms the scholar. The temp-
tation to abandon 'theory' (or general laws?) when it comes to social
stratification in India is strong, but stronger still is the temptation to
construct an 'Indian theory' of the social phenomenon. This Indian the-
ory of social stratification, it is believed, would be faithful to the idio-
syncracies of the Indian situation and would fully flesh out the whole-
someness of caste, linguistic and religious diversity in this country.
Much as this 'territorial orientation' (Singh 1985 : 53) is attractive it has
not yet yielded anything which has risen above the capriciousness of the
author. There is good reason for this too. The advantage of theory is

not only that it attempts to explain and frame observations, but it also
performs the vital function of helping us communicate our varied expe-
riences. We must move from the particular to the general if we are
to share our experiences meaningfully. What is more, we get a deeper
insight into our own experiences as a result of such theoretical commu-
nication. Only a fool-hardy mountaineer would attempt to scale the
Everest without learning from the experiences of other mountaineers
who may have scaled other mountains. Theories presume concepts, and
concepts by their very nature allow us to group and categorize mani-
fold experiences. All theories, Indian and non-Indian, must utilize
concepts, and all concepts ought to satisfy certain basic logical prin-
ciples if they are to be of any theoretical use. The difficulty, it seems to
me, is that one is not always very careful about what the principal
concepts of the various theories of social stratification imply. This is
probably why there is some discomfort in certain quarters with respect
to the application of general theories of social stratification to the Indian

It is for this reason that it was felt that it would be best to devote the
following pages to a clarification of the concepts of 'hierarchy' and
difference' as they are central to all theories of social stratification (see
Appendices, I, II, III and Madan 1980). In addition, 'hierarchy' and
'difference' inform other commonly used concepts like caste, class
and status, as we shall soon see. Once we realize the importance of the
basic principles that govern these terms our usage of them will be
more sophisticated, and our understanding of terms like caste, status,
class, prestige, will become more amenable to rigorous theoretical
treatment. No claim is being made at this point of the relative merits
and demerits of the various general theories of social stratification,
such as those of Marx, Weber, functionalism, or culturology. What we
wish to underline is that a clarity of the key concepts, namely, hierar-
chy and difference, will help us in our individual theoretical drives.

Hierarchy and Differences : The Key Concepts
Before we settle down to a close scrutiny of the logical properties of the
concepts of hierarchy and difference, we should spend a little while in
carefully going over the more general term stratification itself, and
what it implies.

Stratification spontaneously signifies a multi-layered phenomenon,
much like the earth's crust (Beteille 1977: 129). The point to remember
in this connection is that the geological metaphor can be misleading in


the case of social stratification in so much as it might figuratively per-
suade one to believe that stratification always implies layers that are
vertically or hierarchically arranged. For a true understanding of strati-
fication we should be able to conceptually isolate it from hierarchy, as
the latter is but one of the manifestations of the former.

The various layers that stratification spontaneously signifies do not
imply unconditional differentiation. The differentiation is always on
the basis of a criterion, or a set of criteria. Stratification therefore im-
plies a common axis (or axes) that straddles the differences. Quite un-
like geology again, social stratification does not manifest itself readily
or 'naturally' to the naked eye. A deliberate act is required on the part
of the observer or analyst to unite certain kinds of differences in order
to construct a particular system of stratification. In discussing any sys-
tem of social stratification we acknowledge an overarching commonal-
ity (or similarity) which like a thread links the manifest differences to-
gether. Social stratification is not like distinguishing between cabbages
and kings: it does not group disparate entities without a clearly stated
criterion or a declared set of criteria.

Commonality then exists as a pre-condition for all systems of strati-
fication. If only differentiation were to be emphasized then how would
systems of stratification emerge? How also could one justify the inclu-
sion of certain elements and not of others. Cabbages, kings, ships and
sealing wax do not after all make for any system. But when the popula-
tion is stratified, say on the criterion of income, then we have an uni-
form criterion which can bring together sweepers, managers, white col-
lar workers, and agricultural labourers into a single system of stratifica-
tion where monetary income is the regnant principle. Likewise when we
construct a social stratification of language groups the unifying basis is
language and it does not matter if the language speaker is a sweeper or a
college professor. Finally these sweepers, managers, white collar work-
ers, and professors can also constitute a system of stratification based on
the criterion of occupation. We are not really interested if these manag-
ers, sweepers, etc. are short or tall, married or unmarried. The only fac-
tor that interests us is that they all perform a manifest occupation. In
each case then there is a presumption of a commonality that systema-
tizes the differentiation of the various strata and binds the universe of
a particular form of stratification.

(a) Hierarchy implies the regular ordering of a phenomenon on a
continuous scale 'such that the elements of the whole are ranked in rela-
tion to the whole' (Dumont 1988: 66). Height, weight, income and even
power (once it has been quantified) can be arranged in a hierarchy. Tall


and short people can be arranged in a hierarchy of height. You cannot
position short or understand shortness unless you have a hierarchical
scale that tells you what is tall and tallness. Hierarchy is but one form
of social stratification and it certainly does not constitute the essence of
social stratification. Indeed this is just the mistake that the famous soci-
ologist Pitrim Sorokin made when he wrote :

Social stratification means the differentiation of a given population into hier-
super-posed classes. It is manifested in the existence of upper and
lower layers (the geological metaphor, D.G.). Its basis and very essence
consists in an unequaldistribution of rights and power
and influences among the members of a society (Sorokin 1961: 570; emphasis

Quite obviously for Sorokin, inequality and hierarchy were the stuff
of social stratification. The geological model of layers too is quite evi-
dent. The various layers are always arranged vertically. If, for instance,
we were to be discussing the stratification of power then those at the
top have more power than those below them and so on till we come to
the last layer that has the least power. The same can be said about wealth ·
and examples proliferate.

But not all systems of stratification are hierarchical. Some are, but
many are not. In the latter case 'difference' is valorized, and notions of
hierarchy may or may not surface.

(b) Differences rather than hierarchy are dominant in some stratifi-
catory systems. In other words, the constitutive elements of these differ-
ences are such that any attempt to see them hierarchically would do
offence to the logical property of these very elements. The layers in this
case are not arranged vertically or hierarchically, but horizontally or
even separately. Such an arrangement can be easily illustrated in the
case of language, religion or nationalities. It would be futile, and indeed
capricious, if an attempt was made to hierarchize languages or relig-
ions or nationalities. In these cases it does not matter at all if the sche-
matic representation of stratification places the different strata con-
tiguously or separately, as long as they are horizontally positioned.
India again is an appropriate place to demonstrate this variety of social
stratification. The various languages that are spoken in India speak
eloquently of an horizontal system of social stratification where differ-
ences are paramount. Secular India again provides an example of reli-
gious stratification where religions are not hierarchized or unequally
privileged in law, but have the freedom to exist separately in full
knowledge of their intrinsic differences.

A system of social stratification then implies differentiation among
one or more features in such a fashion that they can be grouped along
a common axis. But as stratification speaks not only of differentiation
but differentiation grouped along such axes, the factor that is common
indicates the nature of stratification. If it is language then we are deline-
ating a stratification of language; if it is income, then we are hier-
archizing a stratification of income; if it is religion then we are stratify-
ing the different religions.

Hierarchy is only one kind of stratification where the strata are ar-
ranged vertically. This is appropriate only when this vertical arrange-
ment is along a variable that can be measured on a continuous scale, as
in the case of numbers. One cannot measure the proletariat, or the capi-
talist, but one can measure income. Likewise one cannot measure lan
guages but one can measure the prestige accorded to a certain language
in a certain region. It is possible then to have a hierarchy of income or
ofprestige, such that in one case different income earners and, in
another, different language groups, can be placed along a continuous
hierarchical scale. The crucial fact in all this is that the differences in
prestige or in income should be either quantitative, or quantifiable (in
terms of more or less of a certain property).

Difference is salient when social stratification is understood in a
'qualitative' sense. According to this scheme, there are incommen-
surable entities or units, that constitute different systems of stratifica-
tion. In place of a continuous scale one encounters instead discrete
categories. Thus in a stratification of classes, for example, different
occupations may be listed without any scalar or hierarchical ranking;
likewise in the stratification of religious groups one might mention the
various religious denominations without imposing on them the uni-
formity of a scale based either on prestige, or on wealth, or on
rationality. Once this is clear then there is little reason to believe with
Sorokin that social stratification principally concerns itself with ine-
quality and hierarchy.

To sum up then, social stratification is the ordering of social differ-
ences with the help of a set of criteria or just a single criterion (which
is generally the case) which ties the differentiated strata into a sys-
Secondly, systems of social stratification just do not exist. They
emerge only after a deliberate act on the part of the observer or ana-
lyst to opt for that common criterion or criteria. Thirdly, because these
systems of social stratification are pivoted on mental constructions there
is often a good deal of heart burning, house burning, and even wife
burning on this account. Different people have different reckonings
of stratification, and when these systems do not match there is friction,
often fire.

Differences and Inequality
We have already said that social stratification implies differentia-
tion, but does this also mean that the strata thus differentiated are also
unequal? It is important to reiterate that there can be separate classes of
stratification, or strata, without there necessarily being any inequality
(whether of wealth, power or prestige) between them. To bear this in
mind is to guard against an oft adopted assumption that inequality per-
vades all forms of social differentiation. This then quite unthinkingly
leads one to hierarchize systems of social stratification which are essen-
tially horizontal. Unexamined prejudices thus find their way in aca-
demic exercises.

A social differentiation that separates without implying inequality is
not always easy to appreciate. This is why an awareness of one's preju-
dices as well as those of others is so essential to the study of social
stratification. Humankind, unfortunately, has not yet developed to a
stage where we can all indulge in and celebrate our differences. Differ-
ences in language, religion, race or sex are differences that in them-
selves do not contain the property of inequality. This may however not
be the popular understanding of these differences.

In the eyes of most people religions, languages, sexes, nationalities
are all hierarchized--though it would be difficult to get an unambiguous
statement of the criteria on the basis of which these hierarchies are
constructed. In fact, a worthwhile question for a sociologist is to ask:
Why is it that people tend to hierarchize horizontal differentiations
whose logical property is equality?

Caste and class both bring to our minds inequality and hierarchy. And
yet only certain operationalizations of these terms justify the implica-
tion of inequality. One can in fact talk of the various castes, both
rural and urban, without directly implying inequality of caste, wealth
or status. As a matter of fact, A.M. Shah in a recent study on the Vanias
and Rajputs of Gujarat has treated each caste as a separate entity
without making any statement on the nature of hierarchy that might
pertain between them (Shah 1988: 3-29). Indeed on many occasions
attempts to hierarchize different castes are fraught with ambiguous
and contradictory postulates. Where can one place the Jat farmer of
west UP? If one adopts the varna system then he would be placed
quite low in the hierarchy, perhaps even deserving the contempt that


the Manusmriti accords to the Sudra. But try calling any of these proud
and prosperous Jats Sudras to their face, and immediately another hier-
archy will become readily visible. The proud Jat bows to no one, not
even to the Brahmin (see Pettigrew pp. 163-75 in this volume). As a
matter of fact the Brahmin is a butt of ridicule in all of Jat land, espe-
cially in Punjab and west UP. And his is not the only caste which is
made fun of. Jat opprobrium fails abundantly on other castes as well:
the unctious and oily merchants, the lowly Chamars, Nais, and
Valmikis - in fact on everyone who is not a Jat.

Likewise one need not imply at all that the white collar worker is
inferior to the capitalist entrepreneur, unless we are talking of power
within the organization or work place, or wealth and material posses-
sions. But once we move away from these attributes and go on to detail
and discuss the differences in life styles, aesthetic tastes, world views
etc., between the white collar worker and the capitalist entrepreneur we
quickly realize that it would be futile, indeed tendentious, if any attempt
were made to hierarchize these differential traits. A little careful reflec-
tion therefore will tell us that certain kinds of strata which we quite
uncritically assume to be concepts denoting a hierarchical system of
stratification may not be just that alone. Such concepts as the white
collar or even middle class may have a life quite distinct from their
popular placement in a hierarchical scale.

It is therefore of the utmost importance to situate concepts in their ap-
propriate contexts. There are at least two common terms used in the so-
ciology of social stratification which sometimes emphasize hierarchy
and sometimes difference. These two concepts are class and caste. It
would be useful to spend some time in working out the varying dimen-
sions of hierarchy and difference in the application of these terms. This
we believe will give us useful insights into the variety of ways in
which caste and class manifest themselves.

Hierarchy and Difference in Caste
Caste has resisted definition quite successfully precisely because its
two dimensions, namely, hierarchy and difference, deflect any single
unifying definitional probe. After a long deliberation Leach (1960: 2-3)
settled more or less for J.H. Hutton's descriptive statement of the caste
system where endogamy, pollution, occupational differentiation and
hierarchy, with the Brahmins at the top, are the important diacritical
features of the phenomenon. Bougle too essentially described the caste
system though he believed he was defining it. Yet Bougle perhaps more


than any one else clearly emphasized the two aspects of hierarchy and
difference in it. According to him hierarchy, repulsion and hereditary
specialization are the three important characteristics of the caste system
(Bougle 1971: 9). Today we know that both in traditional India, as well
as in modem India (of course) castes cannot be linked to occupations
except in exceptional cases. To be fair Bougle noticed this fact too
(ibid.: 18-19). But the reason we commend Bougle is because he gave
hierarchy and repulsion equal importance in his description of the caste

According to Bougle the spirit of the caste system is determined in
an important way by the mutual repulsion that exists between castes. In
other words Bougle is here emphasizing the differences that exist be-
tween different castes. Repulsion, Bougle argued, manifested itself in
endogamy, commensal restriction, and even contact (ibid.: 9). For this
reason different castes stayed as discrete entities, 'atomized', 'op-
posed', and even 'isolated'. The methodologically relevant point is
that Bougle did not see any problem in the coexistence of hierarchy
with repulsion. The logic of the situation should have however led him
to allow for differences in hierarchies as well, but he stops well short of
that and provides for only one hierarchy (presumably the Brahmin vari-
ety); and this is where he is logically at fault.

In most popular renditions of the caste system, hierarchy alone is
emphasized and that too from the Brahmin point of view. The Puru-
legend whereby the Brahmins are said to have come from the
head of the primeval being and the Sudras from the feet is too well
known to bear repetition here. But what is generally not equally well
known is that there are as many such legends, or origin tales, as there
are castes. These origin tales, or jati puranas, justify different hierar-
chies and the Brahmin is not always at the top. As a matter of fact, there
are castes that find even the Brahmin defiling (Dumont 1988: 59). The
presence of such multiple hierarchies is in consonance with the reality
that there are also varying models of emulation which castes employ for
purposes of upward mobility. The Brahmin model is one such model but
there are other non-Brahmin models, such as those of the Kshatriya,
Rajput, Maratha, and even the Vania, and they are all equally persua-
sive to their adherents. It would be incorrect to consider these other
non-Brahmin models to be less worthy of attention because each in-
stance of model adoption implies the conscious and deliberate rejection
of the other available models. The reason for this variation in templates
of emulation lies in the notion of 'difference' and its obverse 'equality'.
Nobody, no matter what their caste may be, would ever accept that


they are made of impure substances, or that the substances in them are
less pure than those of another. Caste legends of Doms, Chamars, Chasa
Dhoba, Kahars, all proclaim exalted origins (see Risley 1891) which of
course the Brahminical texts vehemently deny. Yet each of these tales
captures independently the essence of 'difference' between castes
and are therefore logically of equal status. Gerald Berreman writes of an
incident that portrays this tension over hierarchy accurately. During
the course of his fieldwork he once related to some of his low caste
respondents the orthodox hierarchy according to which the Brahmin
was unequivocally on top, After listening carefully to Berreman these
low caste respondents laughed, and one of them said, 'You have
been talking with Brahmins' (Berreman pp. 84-92 in this volume).

If this is the case then any attempt to study the caste system in terms
of a single clearly ranked hierarchy would obviously run into great
difficulties. If we look at castes closely then we find that each main-
tains its own traditions and customs zealously and clearly distinguishes
itself from others in its universe. Often this has been understood as a
kind of 'caste patriotism'(1). But to make matters difficult, caste is not
just a separation between different castes. Each case of separation
and valorization of differences is accompanied by a unique hierar-
chical ordering of castes. It is another matter that there are disagree-
ments over this hierarchy, and that not all hierarchies can be socially
enforced on a single scale, yet castes as such are never quite rested
even after they have repulsed one another. Differentiation involves
a cathective judgement regarding the elements of bodily purity and
impurity, and this quite spontaneously suggests different, yet specific
attempts to construct hierarchical rankings of castes. It should however
be underlined that such hierarchies are idiosyncratic and equally valid.
We are thus forced to disagree with Dumont in so much as he posits

(1) This caste patriotism differs from self-identification on racial grounds by the fact
that as one goes down to the finer and lower order divisions within a caste till one finally
comes to the level of the endogamous jati, one continually finds a proliferation of
differences all the way. Thus the maintenance of the first order division such as of the
Vania, or the Brahmin, hardly exhausts the plentitude of other distinguishing markers
that exist at lower levels. Thus one cannot say that endogamous jatis are segmented
groups of the first order caste rankings. These jatis are not simply smaller segments of
the bigger whole (see Dumont 1988 : 42).
Race identity, on the other hand, moves in the opposite direction. The passion
with which one identifies first order divisions like 'white' or 'black' is far greater than if
the white and blacks were seen at lower and more disaggregated levels. The significance
of being a white is far greater in South Africa than being a white of Dutch, German or
English origin (for other views on this subject see Dumont 1988 : 247-66).


a single hierarchy for understanding and explaining the caste system.
In this true hierarchy the Brahmins are unanimously at the top and the
Untouchables without dissension are at the bottom. The hierarchy is
therefore a ritual hierarchy (and that is why it is 'true', says Dumont)
which is dependent upon a state of mind and is not influenced by secu-
lar forces of economics and politics (see Dumont 1988: 19, 34, 66;
Madan, 1970: 1-13). In our opinion such unanimity over the Brahminic
hierarchy does not really exist; and the reason for this lack of consen-
sus is remarkably simple. As castes are different and separate it is but a
logical corollary that they should also hierarchize differently and sepa-

Notwithstanding our disagreement with Dumont it is incontrovert-
ible however that Dumont introduced, perhaps for the first time in
sociology and social anthropology, a technical understanding of the
concept of hierarchy. True hierarchies, Dumont clarifies, 'are ranked
in relation to the whole' (Dumont 1988: 66), with the added proviso
that 'that which encompasses is more important than that which is en-
compased' (Dumont 1988: 76). Thus it is clear that hierarchies suggest
an overall unity such that the differentiated strata within the hierarchy
are encompassed by the defining criterion of the system. This is why he
advised, contra Bougle, that to understand the caste system it was all
important to grasp the principle of the true hierarchy and not wander
among differences (Dumont 1988: xlviii, 43). It is perhaps because
Dumont did not pay attention to the active principle of differences
in the caste system that the Hindu caste order is presented as one with-
out internal tension and dynamism, and the Hindu person as an arche-
typal representative of the species, Home Hierarchicus (see also
Desai 1988 :49).

One might, at this stage, ask the question whether hierarchy comes
before differences or differences before hierarchy (see Desai ibid.:
42)? There are good grounds to dodge the question but it would be
more forthright to suggest that the existence of different hierarchies
encourages one to take the position that differences dominate the ar
ticulation of a hierarchy in the caste system. Those hierarchies that
are socially enforced on a general scale do not subsume the number

(2) Gerald Berreman's works are most instructive on this account. In recent times
Berreman has time and again emphasized the fact that different castes have different
evaluations of the caste hierarchy. It is unfortunate that Berreman's views are not as
well known in India as Dumont's ideas are (see Berreman 1963 : 214-15, 222-3; 1979 :
that exist in an intraverted form in the more closeted observances
and beliefs of the subjugated castes. The ideological motivation to San-
skritize does not appear only when the hitherto subordinated castes
have either money or power to fancifully conjure another hierarchy.
The other hierarchy is always there waiting for a propitious moment to
extravert itself generally over the entire society.

Hierarchy and Difference in Class
The importance of distinguishing between hierarchy and difference can
be exemplified with reference to the concept of class as well. Like caste,
the concept of class finds its way into a large number of theoretical for-
mulations of social stratification. Not always is it made clear whether it
is being used in a hierarchical sense or in the sense of a horizontally
differentiated and separate stratum. Most often any mention of class
stratification presumes a hierarchical ordering though the concept is not
logically limited to such operationalizations alone.

Class refers to a system of stratification that is economic in charac-
ter. We are all familiar with terms like upper class, middle class and
lower class; or, rich, middle and poor farmer. Sometimes these terms
can be increased depending upon how fine one would like the catego-
ries to be. Therefore, it is often the case that one separates the upper
middle class from the middle class or the lower middle class, and so
on. There is no analytical problem in adding to the numbers of strata
thus, because they are all being read off a hierarchical scale. Therefore
we can have a class category depending upon the criterion of land, or
one depending on the variable of money, or one on marketable yield, or
one on disposable income. The important thing is that all of these crite-
ria are convertible directly into money and that is why in class stratifi-
cations money or wealth is always central.

In spite of the matter appearing so simple one must exercise a num-
ber of precautions when using these terms. First, it ought to be real-
ized that the cut-off points on a hierarchical scale which signify strata
like upper class, upper middle class, middle class, lower middle class,
lower class, and so on are essentially arbitrary. At what point the lower
middle class becomes a lower class depends on considerations not
imminent in the hierarchy. That is why it is important to remember that
cut-off points on the hierarchy are justified on the basis of cohort fac-
tors which do not figure in the hierarchy itself, but are employed by
the analyst to justify the demarcations for the purposes of a specific
analysis (see also Singh 1977: 21 2). For each analysis therefore the
cohort factors justifying the demarcations in the hierarchy will differ.
For this reason agrarian classifications which use such strata, as 0-5
acres, 6-10 acres, 11-15 acres, 16 acres and above, have to be revali-
dated with every fresh analysis. The problem in many cases is that
these classifications are often seen as absolute in themselves, thus
committing the analyst to elementary errors of reification. In one area a
person with 5 acres may be an impoverished and marginal peasant, but
elsewhere a farmer with 5 acres may be a prosperous member of the

This leads us to the second point of caution while employing strata
that eponymously signify a hierarchical scale. The middle peasant, or
the middle class, refers quite obviously to a stratum which is in the
middle of the hierarchy of land and wealth respectively. But the man-
ner in which these terms have been used and have gained salience urges
us to a construct a much fuller picture than the flat one-dimensional
one that is read off a hierarchical scale. The understanding of the term
middle peasant has attained a certain analytical status because its first
approximation as one belonging in the middle of the land hierarchy
has been abundantly superimposed by a host of other characteristics
which are well outside the scope of the criterion that defined the land
hierarchy. Some of the factors that give the concept of the middle
peasant its analytical leverage in contemporary literature are ideologi-
cal innocence, thriftiness, the employment of family labour, negligible
interactions with the market, production for consumption, and so
forth. The middle class too is often conceptualized in a similar
manner. The attribute of cultural pretension, or the propensity for
urban occupations are not features of the income hierarchy on which the
concept of the middle class may receive its initial validation. Income
or land ownership then become only one of the many characteristics by
which such classes are understood. But as many of their other features
cannot be merged into a single hierarchy, these strata gain much of
their salience from their other attributes, namely, those that signify
'difference'. Even when certain features can be hierarchized, such as
the employment of family labour, in the case of the middle peasant, the
change in quantity on the continuum is regarded as so significant that
when the peasant employs hired labour in the main he undergoes a
qualitative transformation. In keeping with this qualitative transforma-
tion other attributes peculiar to the rich peasantry, such as entrepre-
neurship, urban preferences, ideological aggression, and so on
become critical cohort factors--some of these are not amenable to
hierarchization, and those which are call out to other hierarchies based
on different criteria. One should then, with a little care, be able to dis-
tinguish between pure hierarchical strata, and those which are epony-
mously so but depend in addition on attributes of difference.

While we have so far discussed hierarchical strata based on a single
hierarchy, the same principal obtains even when a composite index
is made up of different variables which have been quantitatively opera-
tionalized. In other words the aim of fashioning such composite indices
is to arrive at one hierarchial measure. In the formulation of the indices
of Socio-Economic Status (SES), education, occupation, prestige, in-
come were first hierarchized and then merged together. Thus though
each strata in the hierarchy have a variety of attributes they are visual-
ized as being causally linked. For instance, Yogendra Singh and B.
Kuppuswamy write:

Education has been considered to be a deciding factor of one's occupation,
occupation an important intervening variable in the translation of educational
advantage into income advantage, and the income a positive factor in de-
ciding one's social prestige which, in turn, influences the educational level
of the succeeding generation and possibly of the same also (quoted in Singh
1977: 21).

The differences are here merged and united reinforcing thus the
single criterion hierarchies. Qualitative differences between the
different variables that go into the making of SES, like composite indi-
ces, are deliberately sublated in order that these indices be quantifiable
and obey the principle of hierarchy.

Hierarchy and Differences in Order and Conflict
Rarely do social classes present themselves simply as clusters around a
continuous hierarchy. Life styles, beliefs, family size, etc., come in to
characterize, almost uniquely, strata which, in the first quick look, may
be considered to belong to the continuous hierarchy alone. Distinguish-
ing between classes on the hierarchy from those that may only be remi-
niscent of it has other advantages too. It predisposes us, for instance, to
anticipate the different analytical consequences that follow when one
uses them especially with reference to conflict, continuity, order and
change. If hierarchy alone is emphasized then there is little scope for
allowing for change, conflict and dissension. In a hierarchy, as we
know with the help of Dumont, 'that which encompasses is more
important than that which is encompassed' (Dumont 1988: 76). Classes


understood simply in terms of their hierarchical placement cannot be
utilized analytically for the study of change or class conflict. The prin-
ciple of the true hierarchy, namely, that of encompassment undermines
the potentialities of conflict if it does, not negate them altogether. Hi-
erarchy with its principle of encompassment signifies order and con-
formity. When one makes a hierarchy of wealth, or of power, or of
prestige, then in each case continuity, conformity, order and objectifi-
able acquiescence to the hierarchy are valorized. Even multiple SES
indices conform to this logical rule. While they help us make syn-
chronic comparison, they are 'essentially static' (Singh 1988 : 23-4).

In order to understand the dimensions of conflict within the frame-
work of social stratification it is essential to realize that conflict and
tension can only be examined with the aid of concepts which do not
owe complete allegiance to a hierarchical order but which have signifi-
cant diacritical features of their own. While the manager, the superin-
tendent, the white collar worker, and the dirty white collar worker, may
be placed in a hierarchy within an organization, yet, if one is to
understand organization tension and conflict then, these very classes
must step out of their one dimensional profile in the hierarchy and
assume a more qualitatively rounded presence. This is a logical require-
ment. As the hierarchy emphasizes unity and conformity, therefore
any attempt to go beyond this level will necessitate an absorption of
characteristics outside the criterion of the hierarchy or, in other
words, attention must be paid to the multiple features that spell differ-
ences. With differences comes the notion of equality. Thus though the
hierarchy may spell out unambiguously the inequality within the sys-
tem we are still within an interiority whose sovereignty cannot be
undermined without bringing in 'differences' from without. Thus
though a manager remains a-manager, a worker remains a worker, and
a bobbin boy remains a bobbin boy, yet in an industrial dispute an alter-
native dimension comes into play. Now there is scope to, and room
enough for, protest, agitation, or strike, for the hierarchically subordi-
nated seek equality at other levels through the medium of differences.
Political commitment, world views, aesthetic tastes and ethical values,
are some of the differences that come into focus that separate the
working class from the managers. For any agitation to take place in a
hierarchically ordered organization it is an unconditional necessity
that the dimension of 'differences' become salient.

In the case of castes, too, if there is a single true hierarchy (as Dumont
posits) then that logically forecloses the possibility of conflict within
the system. This is because the caste hierarchy, like all hierarchies,


inheres in the relation between that which encompasses and that which
is encompassed. Caste conflicts and caste mobility occur because there
are full-fledged differences between castes. Because of these differ-
ences, as we said before, alternative hierarchies, which are logically of
equal status, arise. And as Dumont said correctly in another context,
conflict arises only among equals (Dumont 1972). But Dumont's con-
text was restricted to the rivalry for supremacy in village factions and
caste panchayats. Dumont allows for conflict in these limited areas such
as the caste panchayat because the members of a caste panchayat, or
of a dominant caste, belong to the same caste and hence are equals.
This led to 'plurality' of power (Dumont 1988 : 164, 182-3). Once
again, we believe, Dumont is logically correct but is empirically too re-
strictive. His mistake in this case is that he is restricting the play of
equality far too strictly. Conflicts arise on a far more general scale in
caste societies because of the existence of multiple caste hierarchies,
which are all separate and 'equal' and support their positions through
their own caste ideologies.

While such a position is initially perhaps a little difficult to accept
with reference to caste and class, a fidelity to the logical requirements
of the terms, hierarchy and difference, help us see the matter somewhat
freed from our reigning prejudices. As we all, researchers and respon-
dents alike, live in stratified societies, prejudices of one kind or the
other are bound to exist even within the most self-conscious amongst us.
This is why it is useful to look at the logical requirements of the key
concepts of social stratification, namely those of hierarchy and differ-
ence, and then examine how these concepts imbrue the more empiri-
cally determined concepts like caste and class.

Conclusion : Hierarchy and Difference in Weber and Marx
Hierarchy and difference not only add to our understanding of con-
cepts like caste and class but also help us to get a deeper reading of
the various received theories of social stratification. Not always
have the authors, or the exegetists of these theories, spelt out the logi-
cal implications of their concepts such that it would further our appre-
ciation of the basic principles of social stratification.

Weber's formulation of the three axes of stratification, namely, class,
status, and party, has many interesting possibilities from our point of
view (see Appendix I). Of the three, 'status' received far greater at-
tention, for Weber was always keen to delineate the alternative ways
by which men gave meaning to their different life styles. But for Weber
each of these axes revolved around a single variable. Class was deter-
mined by reward in the market place;(3) status centred around the con-
cept of social prestige; and the crucial variable behind the party was
power. As can easily be seen, rewards, prestige and power can be
hierarchized and measured along univariate axes and this is probably
why Weber despaired that all changes were only superficial. The
only change that he foresaw with great trepidation and heightened
distaste was the further consolidation of the principle of hierarchy in
bureaucracies that dominated every aspect of society. For this reason
he argued for the persistence of the democratic system for when it
decides among various demagogues there is a fleeting moment during
which bureaucracy is temporarily checked. Even such ephemeral cor-
rectives were welcomed by Weber as the scene was otherwise too
bleak. The fact that Weber saw no alternatives to the present may ex-
plain why he should be the leading figure with a large number of
generally conservative behaviourial scientists of today. This is no small
irony for unlike many later Weberians the master himself never glori-
fied the present but reviled instead the 'iron cage' in which modem
European societies were trapped (see Weber 1948:120-5; see also
Loewenstein 1966: 24-5).

In Weber's understanding of status group there lies a great poten-
tiality for emphasizing differences. But as we just mentioned, Weber
himself chose to unite these divergences in the hierarchy of prestige.
For this reason caste was seen by him as a case of closed status groups
(Weber 1958 : 39) and differences within the caste system were thus
unfortunately sublated. In his understanding of caste perhaps Weber
was not to blame, for such has been the power of popular conception on
caste, that even today hierarchy is often emphasized over all else in the
understanding of this unique Indian system of stratification.

This is not to suggest that Weber saw no 'difference' at all. But in
Weber's understanding of difference, such as in his typology of world
religions, the different religions are portrayed as unique totalities quite
independent of one another. For instance, the world affirming religions
exist quite independently of the world abnegating religions. This is
quite different from the manner in which different classes are dialecti-
cally related in Marx's works. Change impulses exist in Marx because,
with the exception of Marx's distinction between literate and pre-literate
societies, differences are integrally related and not uniquely isolated.

(3) Dumont rightly observes that for Weber 'all buyers and all sellers are as such iden-
tical (Dumont 1988: 105; see also Parkin 1982:93).
To refine then what has been said earlier, it is not just 'differences',
but linked (or related) 'differences' that allow for an appreciation of the
forces of social change and for a more dynamic frame of reference.
Marx too accepted the primacy of hierarchy in the caste system, for
which reason he was compelled to suspend the logic of historical
materialism when it came to India. Marx has numerous passages to
this effect where he talks of an unchanging India trapped timelessly by
superstition and caste dogma till such time as it was shaken by British
colonialism (see in particular Marx and Engels 1959: 15-18, 31, 34).
This tradition in Marxist thinking still remains a vital strain with many
(see for a good review treatment, Hindess and Hirst 1975).
Quite in contrast to his description of India, Marx saw great poten-
tialities for change in all class societies. The Manifesto of the Communist
quite clearly states on the first page that class societies are coter-
minous with literate societies where the first distinctions were made
between manual and mental labour (Marx and Engels1969: 108-9). In-
dia then should have also been examined as a class society replete with
the potentialities of historical materialism. Even so, Marx, persuaded by
Orientalist literature, quite uncharacteristically chose to view pre-Brit-
ish India in the main as a society still outside the process of history.
To return to the Manifesto of the Communist Party we see two types
of classes, and not just two classes, as it is popularly believed. The first
type of classes are the so-called social classes, like freeman, journey-
man, apprentice, and guildmaster. The second type of classes are the
analytical classes, such as the bourgeois and the proletariat, whose dia-
lectically contradictory relationship defines and constrains specific
social epochs and also shapes social change. But the social classes of the
first type are descriptive classes and historical change does not hinge on
them. In the Middle Ages there existed 'subordinate gradations' (ibid.:
109) beginning at the top with the feudal lord, who was followed by
vassals, guildmaster, journeyman, apprentices and serfs. But such 'gra-
dations' (hierarchy?) were of little use to Marx in his formulation of the
laws of motion in society. In order to get the laws of motion, the contra-
dictions (differences in their extreme form) between the determinate
classes in society (or classes of the second type) were of critical impor-
tance and as such had to be unearthed for each specific social forma-
tion. In feudal societies, Marx contended, the basic classes in contra-
diction were the classes of the feudal lord and serf; and in capitalist
society the contradiction was between the bourgeois class and the
proletariat (ibid.). All this was packed in the first two pages of the
Manifesto of the Communist Party.

The fact that Marx spent very little time on hierarchical gradations
led him to undermine the aspects of order, continuity and stability in
class societies. But this again was a logical denoument for he under-
stood classes in terms of contradictions, i.e. in terms of extreme mutual
'differences', such that the interest of the two determinate classes in
opposition would always remain irreconcilable. There is just no ques-
tion of the encompassing and the encompassed being applicable here.
These irreconcilable differences can only be overcome by a qualitative
transformation of society as a whole. Wesolowski is thus wide off the
mark when he attributes to Marx a hierarchical understanding of these
very basic classes. According to Wesolowski (l969: 128),
The bourgeoisie enjoy a higher income, a higher level of education and higher
prestige. Theworkers have a low income, a low level of income and low
prestige. The petit-bourgeoisie have an intermediate income, enjoy me-
dium prestige and their level of education is higher than that of the bourgeoi-

To mistake Marx's clear postulation of class contradiction as a spe-
cies of strata continuum again demonstrates that the sociologists of
social stratification quite uncritically tend to assume that all forms of
stratification must necessarily be hierarchical in character. This was
the mistake that we mentioned elsewhere in this paper that Pitrim So-
rokin committed and is repeated again as we just saw by Wesolowski
(see also Fried 1967:52). Perhaps a conscious awareness of the logi-
cal properties of hierarchy and difference will pre-empt such errors in
the future and allow for a more systematic exposition of the basic prin-
ciples that underpin the sources of continuity and change in diverse
systems of social stratification.
 Back to the Academic Staff College Home Page