ArticlePDF Available
Contributions to Indian Sociology 54, 3 (2020): 388–408
SAGE Publications Los Angeles/London/New Delhi/Singapore/Washington DC/
DOI: 10.1177/0069966720945862
Comments: Axel Michaels’s article
Arjun Appadurai, Johannes Bronkhorst,
Veena Das, Pratap Bhanu Mehta, Jonathan Parry,
Thomas R. Trautmann and Ananya Vajpeyi
Axel Michaels’s essay on Louis Dumont’s call for a new relationship
between Indology and anthropology is detailed and thorough, though not
shocking to many of us who have tilled these fields for some time.
Dumont was a curiously influential figure, who stood at the crossroads
of many important streams of thought, including French structuralism,
British social anthropology, a certain sort of Indology and a typically
French tradition of hostility to empiricism which inclined him to privi-
lege the role of ideology in social life and to project that view onto his
areas of study, notably onto India. He also strived to mark himself off
from the great Claude Levi-Strauss, whose grand sweep included the
simple forest societies of Brazil, the mysteries of the human brain and
the abstractions of the cerebral savage.
Dumont was not in fact a close student of Indology, as Michaels
shows so well, and his idea of Indian ideology was inspired by his search
for a structuralist key to Indian social ontology (a term which was not in
use in his time). He found this key in the opposition between purity and
impurity, the foundational opposition of Indian ideology, in his view. His
sense of this opposition was derived from a tradition of ethnological
thought in France and Britain, whose exemplars were Marcel Mauss,
Mauss’s pupil Robert Hertz and the anthropologist Franz Steiner, who
wrote a remarkable work on Taboo, published in 1956. Dumont also
belonged to the French anti-Marxist tradition, dominated in his times by
Raymond Aron, and later by historians such as Francois Furet, who were
also against the dominant Marxism of leading French figures like Sartre.
Comments: Axel Michaels’s article / 389
Contributions to Indian Sociology 54, 3 (2020): 388–408
Thus, Dumont is best read as someone who used his idea of India not
to quarrel with Indologists or with most other anthropologists of India,
for whom he had a certain Olympian disregard, but with that tradition of
Western thought which he thought had been buried by its modern focus
on the individual as the bulwark of equality, liberty and value. India for
Dumont was the poster child of an alternative life-world, organised
around hierarchy rather than around equality. This is a tricky platform on
which to build a major manifesto, since equality and inequality are often
seen as two sides of the same coin. For Dumont, this linked pair was the
bane of all versions of Marxist thought and indeed of all forms of mod-
ern sociology. Instead of seeing hierarchy as the inverse of equality, both
modern thought and anthropology viewed caste through the lens of class,
race and stratification. For Dumont, hierarchy is simply the principle that
the whole is greater than the sum of its parts, a view he reported that he
owed to Raymond Apthorpe’s unpublished 1956 Oxford doctoral dis-
sertation, a now forgotten source. This view of hierarchy, as a principle
about wholes and parts, and not about rank or stratification, was trans-
posed by Dumont to the ‘encompassment’ of worldly power by cosmic
purity, thus making caste a manifestation of cosmological holism rather
than of empirical inequality. This rather dubious series of Dumontian
moves was justified less by his reliance on Indology and more by finding
elements of this logic in myriad village studies, themselves replete with
contradictions and gaps. Dumont believed that he had found just the
right all-purpose fix for these problems in his highly abstract view of
Indic ideology.
For Indian scholars like M.N. Srinivas, Dumont was far too distant
from the ground in Homo Hierarchicus, unlike in his previous studies in
rural South India. For others like Andre Beteille, he was interesting, but
ultimately too disconnected from the complexities of economy, power
and social change in India. For McKim Marriott and his many Chicago
proteges, he was insufficiently precise in his use of Indic cultural terms
to build a true ‘ethnosociology’ of India. To most Indologists working
anywhere, he was either irrelevant or ill-informed in the primary texts of
classical India. One important exception was the Dutch Indologist, Jan
Heesterman, whom Michaels also cites, who made Dumont’s style of
analysis livelier for Indologists and historians of ancient India than
Dumont ever did.
390 / Contributions to Indian Sociology 54, 3 (2020): 388–408
Where do things stand today with Dumont’s hope for a new synthesis
between Indology and the anthropology of India? The last few decades
have not been kind to Dumont’s main insight about hierarchy. Dalit
mobilisation, the incendiary growth of fascist Hindutva, the kleptocratic
domination of mercantile elites in India, the dissociation of Brahmin
prestige from many traditional sources of social legitimacy, all make
Dumont’s idea of a social world ‘encompassed’ by the principle of the
hierarchical whole, seem all too scholastic. But about one thing he could
claim to be prescient, and that is the persistence of caste in India, though
its brutality has shed the fig leaf of any cosmological framework, the
latter now wholly outsourced to fake ascetics and street corner kshatri-
yas. Axel Michaels has provoked us to ask once more about the signifi-
cance of the hierarchical whole in an India of deadly and disassociated
New York University, New York City ARJUN APPADURAI
United States of America
When Indologists prepare an edition (preferably a critical edition) or a
translation of a text, they can flatter themselves with the idea that their
work is done with the utmost rigour and that they do not need outside
help. Once the text is there, what follows is far less rigorous. What
Indologists do with texts depends to, a large extent, on their judgement
or prejudice, and no rigorous method is available any longer.
One of the questions that present themselves concerns the historical
reliability of the information contained in the texts. For a long time,
Indologists accepted that their texts provided reliable information about
early Indian society that could (almost) be taken at face value. They
accepted that early Indian society had been Brahmanical, and that deviat-
ing movements (primarily Buddhism and Jainism) could not but be
protest movements. Only slowly has it become clear that the texts concer-
ned are ideologically inspired and cannot be taken at face value where
historical information is concerned.
In other cases, scholars refuse to take texts at face value, apparently
for no other reason than that their personal prejudices or inclinations do
Comments: Axel Michaels’s article / 391
Contributions to Indian Sociology 54, 3 (2020): 388–408
not like what they find there. It could be argued that the early Buddhist
texts have one common theme, endlessly repeated: there is a method
that leads to the end of suffering. I have not come across work by any
modern scholar who takes this claim at face value; no one even bothers
to reject it. It is apparently taken for granted that all those in their right
mind have no place for such claims. [Contrast this with the fact that a
passing reference by Plato to Atlantis has given rise to an extensive
literature; the idea of a lost continent has apparently more appeal to
modern readers.]
In order to be guided by something better than mere prejudice or
inclination, how should Indologists proceed? It all depends on the kind
of question to which they, or others, seek an answer. For the correct inter-
pretation of, say, a medical text, it may be useful to consult modern
medicine; for the correct interpretation of astronomical data, modern
astronomy may help. If one wishes to know where the authors of the
Ṛgveda (or their ancestors) came from, help is forthcoming from the
study of ancient DNA (e.g. Reich 2018). If one wishes to reject the cen-
tral claim of the early Buddhist canon, it seems appropriate to base such
a rejection on modern psychology (which, as a matter of fact, says little
about the issue). Only if one wishes to draw conclusions about ancient
Indian society, sociology might be of help. In other words, there is no
natural confluence between sociology and Indology. It all depends on the
use one wishes to make of the texts.1
Dumont is not a good example of how sociology and Indology could
collaborate. He adopted romantic notions from contemporary Indology
to attribute individuality to the renouncers of ancient India (Bronkhorst
1997, 2016: 241–256 (§ III.1)). This was not an enrichment of the socio-
logy of early India, nor did it help Indologists step beyond their preju-
dices (or dreams). Michaels enumerates other points (e.g. superiority of
Brahmins, the central role of purity) where Dumont got it wrong, partly
under the influence of Indology.
As an Indologist myself, I am open to the idea that sociology and related
disciplines (short: anthropology) may help Indologists to interrogate their
texts in more sophisticated and fruitful ways (as in the ‘Ethno-Indology’
proposed by Michaels). The study of surviving practices (sacrifices and so
1 Cf. Michaels: ‘The disciplines necessary to understand Indian society and culture must
be selected according to research problems and are therefore also more than just one or two’.
392 / Contributions to Indian Sociology 54, 3 (2020): 388–408
on) may also be helpful. The other way round, I have some doubts regard-
ing the way in which Indology may help anthropology. As the above-men-
tioned case of Dumont illustrates, the ‘results’ of Indological research may
express the prejudices or inclinations of their authors; if so, they are of no
use. And if Indologists use theories and approaches borrowed from anthro-
pology to interpret their texts, circularity may become hard to avoid.
Bronkhorst, Johannes. 1997. ‘Louis Dumont et les renonçants indiens’. Orientalia Suecana
45–46 (1996–1997): 9–12.
Bronkhorst, Johannes. 2016. How the Brahmins Won: From Alexander to the Guptas.
(Handbook of Oriental Studies 2/30). Leiden/Boston: Brill.
Reich, David. 2018. Who We Are and How We Got Here. Ancient DNA and the New Science
of the Human Past. New York: Pantheon.
University of Lausanne, Lausanne, JOHANNES BRONKHORST
I do not disagree with any of the criticisms made by Axel Michaels on
Louis Dumont’s and David Pocock’s proposal to consider the uniqueness
of Indian sociology through considering its relation to Indology.2 I am,
however, left a bit baffled by the fact that most of the criticisms he makes
have been made by many scholars, some of whom he cites, and some
whose views have been so completely absorbed in the discipline that
they do not stand out in the way they did in the 1960s and 1970s. If the
exercise that Michaels has undertaken is not to sound like flogging dead
horses, we need to perhaps go beyond strict disciplinary framings and
ask if there are any residues of that discussion on sociology and Indology
which have the power to still intrigue us.
2 Michaels rightly attributes this strand of the argument to Dumont and leaves out of the
discussion what Pocock’s contribution might have been to this formulation. I am not sure
what makes Michaels feel so certain that by sociology, Dumont meant ‘social anthropology’
because both in France and in India, the disciplinary divisions were not so clear-cut.
Comments: Axel Michaels’s article / 393
Contributions to Indian Sociology 54, 3 (2020): 388–408
The main criticisms that Michaels makes with regard to Dumont’s
project are unexceptionable. To be sure, Dumont pays very little atten-
tion to the colonial mediation of Sanskrit texts; moreover, he settles on a
narrow range of Sanskrit texts and largely excludes vernacular texts
from consideration. Finally, Dumont is not only inclined to frame his
discussion in terms of a sharp division between Western and Indian val-
ues ignoring their entangled histories but he also pays little attention to
the way in which Sanskrit texts were canonised and then delegitimised
from being ‘philosophical texts’ to either ‘religion’ or ‘mythology’ (see
Droit 1989; Marchignoli 2004). These issues go beyond what one might
characterise as orientalism or subjugated knowledges, but this is not the
place to engage these questions in further depth. Might we then ask:
After all is said and done, is there a provocation that remains?
I think one interesting question that one could take up for further
inquiry is the question of why India remains such a thorn in the garden
of anthropological theory (Brandel et al. 2016). I am tempted to take up
three examples of canonical anthropological objects—gift, sacrifice and
gods—and ask if their treatment in Indic traditions might be taken as
models for thinking rather than treated as lower order ‘native’ concepts
that have to be lifted through some higher order concepts or super
concepts. Would such an exercise make these objects appear in a different
light from their assigned places in anthropological theory? For reasons
of space, let me just take the example of ‘gods’. Dumont and Pocock’s
(1959) thinking on gods in relation to sociology (or anthropology) of
religion was bold when he declared that, in India, the religion of gods is
secondary, the religion of caste is fundamental. However, it is striking
that Dumont did not actually take up the discussion on gods in any depth
from any texts in vernacular or Sanskrit despite the fact that such issues
were central to the Mīmāṃsā theory of ritual; to the later devotional
literature; or even to missionary translators of the Bible that had to
contend with the notions of god in the Indic traditions. Part of the reason
that such texts do not figure in Dumont’s discussion was that he had
already decided that the principle of purity and pollution in dharmaśāstras
was of major importance in the understanding of Indian hierarchy and
the caste system because this principle represented the domain of the
social. But texts on dharmaśāstras were brought into prominence
precisely because of the colonial concerns with eliciting a body of texts
that could be used to govern Hindu subjects and from which the ‘wily’
394 / Contributions to Indian Sociology 54, 3 (2020): 388–408
Brahmins could be excluded once they had produced these texts as
Michaels rightly points out. But had Dumont considered the Brahmins as
scholars, theoreticians of rituals whose concepts had the power to
challenge some of the formulations on these issues in social theory, he
might have encountered serious challenges to the dominant theories of
sacrifice that assume a transactional relationship with gods through
sacrificial offerings or to the ideas of gift and reciprocity.3 There is an
even deeper question here as to whether the anthropology of religion as
a field carves a space for itself through a conceptual furniture that claims
to be universal but is deeply embedded in a Christian understanding.
Successfully or not, there was at least a concern around such claims to
universality that Dumont tried to grapple with. One could contrast this
strand of Dumont’s thinking with that of other anthropologists, for
example, Evans-Pritchard and Godfrey Lienhardt, who assumed that
Christian concepts can be effortlessly retooled as secular universal
concepts (see Das 2020: Chapter 10, for a fuller exposition of this
argument). As Larsen (2014: 107) points out,
One immediately knows what kind of book Nuer Religion is: on the
very rst page Evans-Pritchard discusses the term kwoth, (‘spirit’) in
relation to the equivalents in what in traditional Catholic teaching are
the three sacred languages: the Latin spiritus, the Greek pneuma, and
the Hebrew ruah. In the preface, Evans-Pritchard asserted that Nuer
and Dinka religions ‘have features that bring to mind the Hebrews of
the Old Testament’ and therefore he deantly warned readers that the
Bible would be a recurring point of reference.
Having escaped this common trap of taking the common sense of
Christianity as the common sense of humanity, let us imagine that Dumont
might have been able take one more step and had turned to Sanskrit and
vernacular texts that actually engaged this question of the religion of gods
being secondary and treated them as companions to his thought. One
question he would have found that repeatedly comes up in the ritual
3 The contortions through which the paradox of marriage as the gift of a daughter for which
no return should be sought has been ‘tamed’ to save the centrality of the parallel cousin,
cross-cousin distinction in alliance theory, is one indication of how India presents, as I said,
a real thorn in the garden of anthropological taken-for-granted theoretical formulations.
Comments: Axel Michaels’s article / 395
Contributions to Indian Sociology 54, 3 (2020): 388–408
theory of Mīmāṃsā pertains to the ‘reality’ of gods—their presence or
absence outside the ritual context. The proponents of Mīmāṃsā ask, if
gods are simple creations of language and come into existence through
the act of invocation in the sacrificial arena, then do they have any
ontological reality outside the sacrificial arena? In practical terms, there
are such questions as to whether the offering in the sacrifice is primary of
whether the deity is primary? Could one deity (devatā) be substituted by
another? The great scholar Kumārila’s answer is that an object or a sign
can only be substituted by something similar.4 But how does one establish
similarity between two deities when the very omission of the word for a
specific devatā would means that the devatā disappears from the ritual
space. Similarity between two devatās, he thus concludes, is not possible
to establish, and it would follow that a deity cannot be substituted by
another deity for any specific ritual. As the Jesuit theologian and the
eminent Mīmāṃsā scholar, Francis Clooney, says, ‘It seems that in
Śabara’s view, there is no essence to devatā at least none that is relevant
to the sacrifice; there is only a web of grammatical and act-oriented
relations, whereby that which functions as a devata is established’
(Clooney 1997: 346). The questions that arise within this scholarship are
questions that take up theories of reference, signless signification,
projection and substitution,5 but what Dumont finds is caste and purity of
Brahmins. Could this be because, for Dumont, Brahmins were not real
scholars but only ritualists, nor did he attend to the fierce challenges to
Mīmāṃsā scholars that came from the Buddhists, or from the devotional
movements—one side claiming that their notion of absence as they
debated the nature of deities did not go far enough, the second that it went
too far. Why are these debates irrelevant for thinking of general theories
of gods or divinities, for the anthropology of religion?
Much as Michaels wants to distance himself from Dumont, I wonder
if in his insistence that missionaries, orientalists and administrators were
the main producers of knowledge on India he ends up at a point where
4 There are deep resonances between the philosophy of grammar, ritual and aesthetics as is
well known. The point about similarity draws from, or is resonant with, the theory of lopa
in Panini and other grammarians.
5 I take the turn to grammar here as generative of these questions not in the usual sense of
grammar as a set of rules governing correct speech but in the Wittgensteinian sense of telling
us what kind of an object something is (see Das 2020).
396 / Contributions to Indian Sociology 54, 3 (2020): 388–408
the knowledge interests of Indians and their investment in producing
knowledge on their own society is simply bypassed. What about the
work of genealogists, essayists who wrote in the vernacular, or the
Pandits who wrote in Sanskrit engaging the kinds of issues I outlined
above? And why do Indian missionaries or Indians employed within the
colonial administration not count as producers of knowledge? Finally,
we might remind ourselves that Europe was not the only ‘other’ that
India encountered. In fact, defining anthropology as the knowledge of
the ‘other’ might itself be the problem. To counter this view, one has to
step back and look at the vibrant discussions across regions and languages
with Farsi and Arabic on one side and Tibetan, Chinese and Japanese on
the other.6 India is not just a story in a chapter written by Europe about
itself. The point is not that one turns to Indian texts to find terms that will
make anthropology or sociology authentically ‘Indian’ but that our very
picture of thinking needs a bit of a wash. I am grateful to Michaels for
reengaging these compelling issues.
Brandel, Andrew, Veena Das, & Shalini Rennderia. 2016. ‘Locations and locutions:
Unravelling the concept of “world anthropology”.’ In Post-Western Sociology: From
China to Europe, edited by Laurence Roulleau-Berger and Li Pielin, 88–106. London
and New York: Routledge.
Bühler, Georg. 2013. Detailed Report of a Tour in Search of Sanskrit MSS: Made in
Kashmir, Rajputana & Central India. Srinagar: Jaykay Books.
6 I take just one example recalling the work of Pandit Iśvara Kaul, who was well versed
in Sanskrit and Persian in addition to Kashmiri, and, toward the end of his life (he died in
1893), he was the head of the translation department created by Raja Ranviira Simha and later
head astrologer in the court of his successor, Maharana Pratap Singh. The Pandits played a
crucial role in the compilation of Grierson’s Linguistic Survey of India, and, as many recent
texts on linguistic philosophy testify, the scholarly depth of their work cannot be reduced
to simple research assistance. In his fascinating book on Grierson and his collaborators,
Javed Majeed says that if Grierson spoke through the Pandits, the Pandits also spoke
through Grierson’s voice (Majeed 2018) That, much remains to be done to bring a deeper
understanding of the unpublished Sanskrit, Persian and vernacular manuscripts is obvious
(Bühler 2013) but new questions are emerging across the disciplines of Indology, history
and anthropology that hold great promise for opening up these issues with fresh eyes (see
for instance, Sengupta and Ali 2011).
Comments: Axel Michaels’s article / 397
Contributions to Indian Sociology 54, 3 (2020): 388–408
Clooney, Francis X. 1997. ‘What’s a God? The Quest for the Right Understanding of
Devata in Brahmanical Ritual Theory (Mimamsa).’ International Journal of Hindu
Studies 1 (2): 337–85.
Das, Veena. 2020. Textures of the Ordinary. Doing Anthropology After Wittgenstein. New
York: Fordham University Press.
Dumont, Louis, and David Pocock. 1959. ‘Pure and impure.’ Contributions to Indian
Sociology 3 (1): 9–39.
Larsen, Timothy. 2014. The Slain God: Anthropologists and the Christian Faith. Oxford:
Oxford University Press.
Majeed, Javed. 2018. Colonialism and Knowledge in Grierson’s Linguistic Survey of India.
New York: Routledge.
Marchignoli, Saverio. 2004. ‘Canonizing an Indian Text: A.W. Schlegel, W.von Humboldt,
Hegel, and the Bhagvadgītā.’ In Sanskrit and ‘Orientalism’: Indology and Comparative
Linguistics in Germany,1750–1958, edited by Douglas T. McGetchin, Peter K.J. Park,
and Damodar SarDesai, 245–70. Delhi: Manohar Books.
Sengupta, Indra, and Daud Ali, ed. 2011. Knowledge Production, Pedagogy, and
Institutions in Colonial India. New York: Palgrave Macmillan.
Droit, Roger-Pol. 1989. L’oubli de l’Inde : une amnésie philosophique. Paris: Presses
Universitaires de France.
Johns Hopkins University, Baltimore, MD, USA VEENA DAS
Axel Michaels’s characteristically clear, erudite and historically informed
essay is right to conclude that the sociology of India is ‘not at the conflu-
ence of anthropology and classical Indology but at the point of the con-
fluence of many different disciplines’. But, even at the risk of being a
little unkind, for those outside of these disciplinary debates, this conclu-
sion might remind one of Sherlock Holmes’s quip to Dr Watson: ‘Your
grasp of the obvious amazes me’. So the question is why were so many
smart and erudite people engaging in a non-debate for so long.
It seems to me that the debate over the relationship between Indology
and sociology was more about social theory than it was about studying
Indian society. Axel Michaels hints at this in his reference to Dumont’s
relationship to Durkheim. Dumont’s project was less about Indology. His
was more an attempt to position himself in a tradition of French social
theory that was concerned with parsimony and moral meaning, rather
than complexity and behaviour. Just as for Tocqueville, equality or the
‘democratic social state’ becomes the organising key to study the ‘West’,
398 / Contributions to Indian Sociology 54, 3 (2020): 388–408
for Dumont, hierarchy becomes the organising key. They are of course
aware of complexity, but these act as the organising principles which
define the horizons within which agents in these societies act. And like
Tocqueville, Dumont was more interested in the kind of possibility this
form of social organising held out for human subjectivity. What kinds of
human subjects does it create? The ‘renouncer’ for example.
It is not surprising that this project privileged the confluence of
Indology and sociology. Dumont on the one hand needed to show that
India had legitimising principles of social organisation that were cap-
tured in thought (Indology) and he needed to show that these did in fact
work their way through social organisation (sociology). History, eco-
nomics or even politics was less of interest to Dumont, in part because of
an underlying assumption that some form of agency—the idea that
agents make and remake their own social world that is subject to
change—is a necessary condition for these disciplines to emerge and
take hold. In a way what had made the confluence of sociology and
Indology attractive was that both (at the time) shared an underlying
assumption that social agents can be read of a script as it were. Bringing
politics and history throws the whole project sky high, because the
emphasis shifts to contingency and change rather than an organising
principle and meaning.
The issue is deeper. Broadly speaking, one can make a rough distinction
between those who are looking to ‘theorise’ society and those who are in
some ways are giving ‘complex descriptions’. Dumont was very much in
the former camp. These are not hard and fast distinctions; observation
often presupposes theory, and theory has a lot of empirical conditionality
built into it. But, in practice, these sensibilities lead to the production of
different kinds of work. For theorists, the aim is parsimony: the search for
an organising principle that is the key to unlocking a society’s secrets.
This principle will have to encompass both social organisation and how a
society thinks of itself. For others, the organising principle is complexity:
the search for a myriad of mechanisms that sustain society. For the
theorist, the search is for a principle that can anticipate the direction of
change. For others, the emphasis is on contingency that will complicate
any attempt to come up with a parsimonious explanation that can persist
over time. The idea is not to search for a master principle that can be
threaded through different aspects of a society, but the quest for more
Comments: Axel Michaels’s article / 399
Contributions to Indian Sociology 54, 3 (2020): 388–408
complex descriptions. The second broad fissure is between those who are
interested in the question: What is the normative significance of a
particular form of social organisation and experience? What conceptions
of the self or the moral life does it give rise to? And those who think these
questions are left to philosophers.
Any complex society marked by a long history will have produced
conceptions of social order, ideologies, normative frameworks, philo-
sophical debates, and many of these will have to be accessed through
classical texts. Texts are attempts to shape reality, not just to represent it.
These texts will suppress knowledge as much as they articulate it. Any
alert hermeneutics will understand this and decode these texts in various
ways. Then there is a question of what is the relationship between these
texts and the social identities of and actions of the people whom sociolo-
gists are studying. Surely this is an open question. Human action can be
studied in many frames: the incentives that drive behaviour, the uncon-
scious motives that propel us to act; the web of symbolic meanings that
provide the horizons within which we act, the relations of power that trap
us, the discursive forms that structure action in different ways and so
forth. At a more collective level, the same thing applies: For any given
institution or human practice, you could decode its symbolic meaning,
you could look at the latent processes that sustain it that lie outside the
realm of the agent’s consciousness, you could look at the incentives that
sustain a particular equilibrium, or the actions of the powerful that keep a
practice in place, you could study a society’s self- image and its norma-
tive debates to see how agents in might or might not act.
These are all commonplace questions in the study of any society.
Different disciplines privilege some questions over others and have a
built-up comparative advantage. An economist might notice incentives
underlying social process; a psychoanalyst might spot unconscious
symbolic identifications, just as a political philosopher might look at the
forms of justification advanced when people make claims on each other.
If you think this way, you will be sympathetic of Axel Michaels’s call for
a multifaceted confluence.
On the other hand, if you are an ambitious social theorist, you might
say this pluralism and complexity is too easy; it is an intellectual
abdication. Saying to Dumont, you need history and politics is like
saying to Marx that he needs religious studies to complete his work. It is
to misunderstand the nature of the project. The debate between those
400 / Contributions to Indian Sociology 54, 3 (2020): 388–408
looking for a master key to enter into the workings of society, and those
looking at dozens of different small windows to break in will continue.
Ashoka University, Sonipat, India PRATAP BHANU MEHTA
Professor Michaels’s critique puzzles me—not least because I wonder
why we still bother with Dumont’s ancient manifesto if it was as anaemic
as he implies. Lately, India had been ‘a kind of anthropological and socio-
logical backwater’ (Dumont 1966). Contributions (O.S.) probably did lit-
tle to make it mainstream, though crucially it laid the foundations for the
1966 publication of Homo Hierarchicus which did a great deal. When I
returned from PhD fieldwork at the end of the 1960s, even my Africanist
peers felt obliged to form an opinion of it and there was a new ‘buzz’
about the study of India. For the handful of British anthropologists of my
generation who did research there, Dumont seemed to provide powerful
purchase on our data and a fresh way of synthesising the field, and his
theory provoked passionate debate for more than two decades. Of why
anybody cared one gets little sense from what Michaels says.
I would unsurprisingly contextualise Dumont’s plea that the sociology
of India locate itself at the confluence between sociology and Indology
differently. Over the past 30 years, the towering figures in British social
anthropology had been Malinowski and Radcliffe-Brown. Their writings
were resolutely ahistorical. Much later, Adrian Mayer—a former student
of Malinowski—recalled that his first reaction to Dumont’s editorial was
that this was just what he and his SOAS colleagues like Freddy Bailey
‘were trying to get away from … writing about kinship on the basis of
Manu, about politics on the basis of Kautilya. These people were trying to
drag us back’.7 Evans-Pritchard (1962), however, had already committed
apostasy with his 1950 Marett Lecture in which he broke with Radcliffe-
Brown’s view of anthropology as a ‘natural science of society’ and
squarely aligned it with the humanities and especially history. The disci-
pline in Britain was then heavily dominated by Africa, but Evans-Pritchard
foresaw that this reorientation would become increasingly urgent as
7 C. Fuller, J. Parry and E. Simpson, unpublished interview with Adrian Mayer, July 2009.
Comments: Axel Michaels’s article / 401
Contributions to Indian Sociology 54, 3 (2020): 388–408
anthropologists took up the study of communities enclosed in the great
historical civilizations; and he ended his lecture by predicting that there
anthropology will become ‘the counterpart to Oriental Studies’ (for
which, on Srinivas’s account,8 he had a somewhat snobbish regard).
Dumont joined the Oxford Institute in 1951 and was much influenced by
Evans-Pritchard. In 1955, he presented a version of the Contributions
manifesto as his inaugural lecture in Paris.
Until then, according to Michaels, ‘the anthropology of India took
hardly any notice of Indological publications…. (and) simply did not
need Indology’. Really? From 1924 to 1959, Bombay’s sociology depart-
ment was headed by G.S. Ghurye whose training was in Indology, and
whose ‘writings reflect an abiding engagement with the classical texts’
and are characterised by a ‘rather uncritical acceptance of the textual
view of Indian society’ (Momin 2013; cf. Dirks 2013). K.M. Kapadia,
Ghurye’s student and successor, also had a first degree in Sanskrit, wrote
a first book on Hindu Kinship (1947) based largely on Indological
sources, and a second, Marriage and the Family in India (1955), that
drew on them extensively. Early in her Kinship Organisation in India
(1953: 22), Iravati Karve—another Ghurye student and classically-
trained scholar—recorded how she would turn to textual sources to
understand her field observations. Strangely, Michaels singles out
Srinivas as a partial exception to his generalisation, though of Ghurye’s
former students he was the one who most stressed his preference for the
‘field’ over the ‘book’ view of Indian society. Nevertheless, he would
much later supervise the PhD thesis of Veena Das (published in 1977),
who analysed a corpus of puranic myths, and who among much else
showed how in them the Brahman is represented as a renouncer in rela-
tion to the householder, and a householder in relation to the renouncer—
an analysis that shed considerable light on contemporary ethnographic
reality in a major centre of Hindu orthopraxy (Parry 1994).
Michaels finds it revealing that Dumont’s editorial referred to hardly
any Indological work, and, that out of roughly 400 references cited in
Homo Hierarchicus, only 11 were authored by Indologists. But is that the
point? What he properly emphasises is Dumont’s unity of India postulate,
a unity claimed to be more than a merely cultural one of the sort found
among neighbouring African tribes. It is ‘the traditional higher, Sanskritic,
8 C. Fuller, 1999. An interview with M.N. Srinivas. Anthropology Today 15 (5): 410.
402 / Contributions to Indian Sociology 54, 3 (2020): 388–408
civilisation’ that not only demonstrates but creates it, and ‘the existence of
castes from one end of the country to the other, and nowhere else’ is proof
of it (Dumont 1957: 9). Does Michaels deny that there is a qualitative dif-
ference from the unity of African tribes? Unless by reference to an over-
arching literate culture, how would he account for it? What he crucially
misses is Dumont’s insistence that ‘the notion of “structure” is indispen-
sable’, and that the unity of which he speaks is above all a matter of the
relations that order the ideological world at both textual and ethnographic
levels. It is not in the number of references cited, but in the identification
of a few common structuring principles that the confluence between soci-
ology and Indology should be sought and that is where the unity of India
lies. No longer one micro-study after another, those of us influenced by
him thought we could now see how our own observations fitted into a
wider structural pattern that embraced both literary and popular levels.
Perhaps we were deluded, but it will not do to dismiss an intellectual
strategy without reference to its fruits. Had Dumont turned his back on
textual knowledge and Indology, he could hardly have come up with
several of his most suggestive hypotheses, for example, his conclusion
that the world-renouncer was the most important source of ideological
innovation in traditional Indian society (Dumont 1965, 1970). Nor with-
out recourse to the varna scheme of the texts could he have arrived at his
theory of the distinctive relationship between status and power in Indian
ideology, from which several other radical propositions are derived: that,
for example, there is also a very distinctive character to the kind of in-
equality around which Indian society is organised; that the ideological
devaluation of power goes along with the historical instability of the
Indian state. Not that Michaels ignores the status/power issue, though he
writes—as is now conventional—as though it had been definitively
resolved in favour of Dumont’s critics. In my view, it has not; works
commonly seen as the most damning indictment of his position are in
crucial respects unconvincing, and what commentators have consistently
failed to appreciate is how this argument fits into a much broader com-
parative enquiry into the way in which religion, politics and economics
are inter-related in a range of different societies—a project he inherited
from Mauss (Parry 1998). What matters here is not whether Dumont was
right or wrong, but rather the importance and productiveness of the ques-
tions he addressed and whether he could have raised them at all without
regard to India’s literate tradition. It is true that a discipline stuck in
the groove of the same old questions may die of inertia, but one that
Comments: Axel Michaels’s article / 403
Contributions to Indian Sociology 54, 3 (2020): 388–408
abandons those that really matter when the wind changes may deserve to
die of frivolity.
Das, Veena. 1977. Structure and Cognition: Aspects of Hindu Caste and Ritual. Delhi:
Oxford University Press.
Dirks, N. 2013. ‘G.S. Ghurye and the Politics of Sociological Knowledge.’ Sociological
Bulletin 62 (2): 239–53.
Dumont, L. 1957. ‘For a sociology of India.’ Contributions to Indian Sociology 1: 7–22.
———. 1965. ‘The Functional Equivalents of the Individual in Caste Society.’
Contributions to Indian Sociology 8: 85–99.
———. 1966. ‘A Fundamental Problem in the Sociology of Caste.’ Contributions to
Indian Sociology 9 (1): 17–32.
———. 1970. ‘World Renunciation in Indian Religions.’ In Religion, Politics and History
in India: Collected Papers in Indian sociology, edited by L. Dumont, 33–60. Paris:
Evans-Pritchard, E.E. 1962. ‘Social Anthropology: Past and Present.’ In Essays in Social
Anthropology, edited by E.E. Evans-Pritchard, 13 – 28. London: Faber and Faber.
Kapadia, K.M. 1947. Hindu Kinship: An Important Chapter in Hindu Social History.
Bombay: The Popular Book Depot.
———. 1955. Marriage and Family in India. Calcutta: Oxford University Press.
Karve, I. 1953. Kinship Organization in India. Bombay: Asia Publishing House.
Momin, A.R. 2013. ‘Revisiting the Legacy of the Bombay School of Sociology.’
Sociological Bulletin 62 (2): 204–16.
Parry, J. 1994. Death in Banaras. Cambridge: Cambridge University Press.
———. 1998. ‘Mauss, Dumont, and the Distinction Between Status and Power.’ In Marcel
Mauss: A Centenary Tribute, edited by W. James and N.J. Allen, 151–72. Oxford:
Berghahn Books.
London School of Economics and JONATHAN PARRY
Political Science, London, United Kingdom
I vividly remember how stimulating it was, as a student, binge-reading
early issues of Contributions to Indian Sociology for Dumont’s new
ideas. Those ideas—so contrary to the received wisdom, so arrestingly
and oddly put, so strenuously argued—did not lead to a conversion expe-
rience, but acted on me as a provocation to formulate yet newer ideas and
argue them rigorously and in depth.
404 / Contributions to Indian Sociology 54, 3 (2020): 388–408
This was especially so for Dumont’s many contributions to kinship
and marriage in South Asia. In the bibliography of my Dravidian Kinship
(1981) I listed eight:
The Dravidian Kinship Terminology as an Expression of Marriage
Une Sous-Caste de l’Inde du Sud: organization sociale et religion des
Pramalai Kallar (A Sub-caste of South India: Social Organization
and Religion of the Pramalai Kallar) (1957)
Hierarchy and Marriage Alliance in South Indian Kinship (1957)
Les marriages nayar comme faits indiens (Nayar Marriages as Indian
Phenomona) (1961)
Marriage in India: The Present State of the Question (Parts I–III,
Le vocabulaire de parenté dans l’Inde du nord (Kinship Terminology
in North India) (1962)
Introduction à deux theories d’anthropologie sociale: groups de lia-
tion et alliance de marriage (Introduction to Two Theories of Social
Anthropology: Descent Groups and Marriage Alliance) (1971)
Dravidien et Kariera: l’alliance de marriage dans l’Inde du sud, et
en Australie (Dravidian and Kaiera: Marriage Alliance in South
India and Australia) (1975)
This amazing body of work seemed to me a step towards a more compre-
hensive framing of the study of Dravidian kinship, to which Dumont had
contributed greatly through his fieldwork, and to which he brought an
analysis informed by alliance theory.
A more comprehensive framing, I thought, would be historical in
character, using the evidence of ancient texts. Dumont’s postulate
showed a way to carry analysis further than he had taken it. For while
there is rather little Indology in Dumont’s own writings, in spite of his
postulate, as Axel Michaels notes, there are in fact abundant ancient
writings bearing upon Dravidian kinship, its theory (Dharmaśāstra) and
practice (South Indian royal inscriptions and their genealogies, Pali
chronicles of Sri Lanka). These could be used to improve the under-
standing of Dravidian kinship by combining its deep history with field-
work studies in the historical present. It was the project of my book, and
I am glad of this occasion to acknowledge anew its debt to Dumont.
Comments: Axel Michaels’s article / 405
Contributions to Indian Sociology 54, 3 (2020): 388–408
There is much to like in the analysis of Michaels. I am doubtful, how-
ever, of the argument that there are ‘almost no points of contact between
the two disciplines’. In the early 19th century, philology and ethnology
were twinned in England, in the sections of the British Association for
the Advancement of Science (also its American counterpart), as well as
the many editions of James Cowles Pritchard’s Races of Man, but, by
mid-century as ethnology was narrowing into race science, Max Müller
called urgently for a divorce. The tendency of the time was toward a
split, not toward a convergence.
The end-point of the paper, however, is entirely to my liking: ‘India,
then, is as much in Europe as Europe is in India.’ Against holism, and
toward intermingling.
Here is an example of India in Europe, one that is hidden in plain
Assuming that Sanskrit is a language like any other, and that it has a
history and is not eternal, Sir William Jones and others concluded that its
similarity to Latin, Greek and certain other languages was a sign of being
co-descendants of a common ancestral language, ‘now perhaps lost’.
This idea of what we now call the Indo-European family of languages
was an entirely new understanding of ancient history, and of the relation
of India and Europe to one another, one that contested prevailing beliefs
both in Europe and in India. It was driven by the imperative in the Europe
of that time to recover the ancient history of relations among peoples
through the historical relations of their languages.
But in carrying out this project, European scholars learning languages
of ancient India absorbed from their Sanskrit and Pali and Prakrit and
Tamil teachers the language-science of ancient India that had developed
so precociously and reached such an advanced state of refinement in
phonology (prātiśākhya) and grammar (vyākaraṇa). Phonology, especially,
was a sharp new tool essential for scientific comparison of languages, and
it was folded into the emergent new science of linguistics, which became,
and remains, an international standard in continuous use. This is so even
though its practitioners rarely acknowledge that their science did not
come about by virgin birth from the mind of Europe but has a mixed and
highly fruitful parentage of the very different philologies of Europeans
and Indians.
So, this mixture thesis of Axel Michaels is wholly good and fruitful.
And his paper is another example of the continuing capacity of Dumont
to provoke the creation of new ideas.
406 / Contributions to Indian Sociology 54, 3 (2020): 388–408
University of Michigan, Ann Arbor THOMAS R. TRAUTMANN
Michigan, United States of America
Axel Michaels’s reconsideration of Louis Dumont’s founding principles
for a desired confluence of sociology and Indology in the form of ‘Indian
sociology’ is indeed appropriately placed in this journal, Contributions
to Indian Sociology. He invites us to look at the histories of both disci-
plines in Europe and India, their separate trajectories over time, as well
as the ways in which they could and could not be merged to yield a third
area like the sociology of India. He also considers their respective place-
ment with regard to proximate disciplines such as Sanskrit philology,
cultural anthropology and the history of religions; all used separately or
in conversation with one another for the study of Indian social and cul-
tural life, whether past or present.
Michaels critiques Dumont’s analysis of Indian society, although this
critique is now long established, having been launched by Arjun
Appadurai in his article ‘Is Homo Hierarchicus?’ (1986), elaborated by
Ronald Inden in his book Imagining India (1990), and continued through
the post-orientalist scholarship of the late Bernard Cohn, as well as his
student Nicholas Dirks, culminating in his Castes of Mind (2001).
(Appadurai too, before Dirks, was Cohn’s student at the University of
Chicago). Thirty years of subaltern studies also played a major role in
dismantling the Brahmin-centric Dumontian view, with its singular focus
on questions of purity and pollution, and its inability to bridge the gap,
whether historically or ethnographically, between social theory and
social experience.
Towards the end of his paper, Michaels suggests that his own method
of ‘ethno-Indology’ as well as his ‘transcultural approach’ offer the nec-
essary corrective to Dumont and his school. But these alternatives are
not fleshed out, at least not in this paper, so that it is not clear how he
would like us to proceed, and what is to become of ‘Indian sociology’
after more than six decades of its career, at least somewhat successful, in
post-colonial India and elsewhere. Further, while the combined effects of
the ideas of Michel Foucault, Edward Said and Ranajit Guha are implic-
itly or explicitly accounted for by Michaels in his sketch of the rise and
Comments: Axel Michaels’s article / 407
Contributions to Indian Sociology 54, 3 (2020): 388–408
fall of the Dumontian study of India, what is missing in this quick survey
is a proper acknowledgement of the seismic shift in Indology created by
the work of the Sanskritist Sheldon Pollock and his method of critical
New methods of Indological practice, developed by Pollock and
now widely used by his students and colleagues in Sanskrit and other
Indic languages of pre-modernity, attempt to place texts in the histori-
cal contexts of their production, reception and circulation as well as in
inter-textual contexts (that are often trans-historical over long periods
of time); continually calibrate between classical and vernacular lan-
guages, as texts traverse different sociocultural terrains on the backs of
multiple languages through translations, commentaries and other vec-
tors of modular reproduction and reinvention detailed by Pollock and
his associates; and finally, closely read texts, against the grain if neces-
sary, to excavate as much of the ‘reality’ of human experience as can
be reconstructed through an awareness of the highly codified rules of
language, genre, political power and social ideology—much of it
Brahminical, but some of it not—that structure Sanskrit texts in their
formal properties.
Michaels could also have paid closer attention to the interesting
entanglements of sociology/social and cultural anthropology/Indology
in the 20th-century Maharashtra in particular. There, starting with giants
such as R.G. Bhandarkar, P.V. Kane and V.S. Sukthankar on the one hand
and G.S. Ghurye, D.D. Kosambi and Iravati Karve on the other, we
arrive at Gunther Sontheimer, R.C. Dhere and Anne Feldhaus, who
achieve a unique confluence of ‘desk’ and ‘field’ in their study of that
region in western India, encompassing texts in many languages, oral
traditions, castes, religious communities and popular culture. Nothing
comparable exists for other parts of the country.
But even with all these advances in both sociology and Indology, a
fundamental gap remains, and this I have found in my own work on the
śūdra over the last 20 years. As Michaels correctly points out, where we
founder is on the question of caste. The view from the varṇa is well-nigh
impossible to reconcile with the view from jāti; in turn, these ‘traditional’
categories of social stratification are radically mediated through colonial
categories of caste and census; further complications arise at the time of
the making of the Indian Constitution in the late 1940s, followed by ever
proliferating indices of social status, economic opportunity and identity
politics after the Market–Mandal–Mandir upheaval of the early 1990s.
408 / Contributions to Indian Sociology 54, 3 (2020): 388–408
The Hinduism and Brahminism of old Indology have become irrelevant
in the face of Hindutva and massive caste mobilisation in 21st-century
With the capture of the Indian state by the Hindu nationalist BJP, and
the ascension to power of Narendra Modi in 2014, discourses of caste are
under direct and indirect pressure, partially from the Brahminical bias of
the RSS, and partially from the overall assault of the Hindu Right on the
Constitutional structure of fundamental rights, egalitarian citizenship
and affirmative action that has been in place since Independence. If the
Nehruvian secular state failed to mitigate caste-based inequality for one
set of reasons (despite its good intentions), the reactionary Hindu Rashtra
comes with its own agenda to re-entrench caste as never before, as it
aggressively pursues majoritarian policies that polarise Indian society
along the twin axes of caste and religion.
In my own work, I have found B.R. Ambedkar to exemplify the very
problem that Michaels sets out to describe. Ambedkar has both an
Indological and a political understanding of caste, simultaneously: He
approaches it both ideologically and anthropologically; attacks it both at
the level of theory and the level of lived experience; explores as much
India’s historical past as his colonial and nationalist present. He was
unable to reconcile many contradictions, even with the unique role that
he alone could play in moulding the text of the Constitution, abolishing
untouchability, and initiating legal reform in the Hindu community as the
first law minister in Nehru’s cabinet.
His final attempt at the ‘annihilation of caste’ and finding a way out
of Hinduism is his ‘conversion’ to Buddhism in the last weeks of his life,
in late 1956. In that decision, to found and embrace Navayana, a new
Buddhist way, Ambedkar the sociologist, Ambedkar the philologist,
Ambedkar the law-maker and Ambedkar the political leader all bow
before what he hopes will be the greater salvific power of religion to
truly deliver freedom, equality, justice and fraternity to Indian society.
His dream remains, alas, as yet substantially unrealised. On the contrary,
under the regime of the Hindu Right, it has receded even further on the
horizon of historical transformation.
Centre for the Study of Developing Societies ANANYA VAJPEYI
New Delhi, India