the+politics+of+not+counting+caste+ +epw

8/23/2019 The+Politics+of+Not+Counting+Caste+ +EPW

1/4

PERSPECTIVE

Economic & Political Weekly EPW June 19, 2010 vol xlv no 25 39

Satish Deshpande ([email protected].

ac.in) is with the department of sociology at

Delhi University; Mary E John (maryj@cwds.

ac.in) is with the Centre for Womens

Development Studies, New Delhi.

The Politics ofNot Counting Caste

Satish Deshpande, Mary E John

In the debate on whether or not

to count caste in the 2011 Census,

there has been too little reection

on the implicit assumptions and

analogies about both the census

and caste that underpin the

positions that have been taken.

This article attempts to identifythe major models that have

been tacitly at work. Questioning

the view that the status quo is

benign or neutral, it argues that

not counting caste has defeated

the desire to transcend caste,

and suggests that caste

blindness be rejected in favour

of a fresh beginning.

An obvious and striking feature of

the debate on the proposed caste

census is that it concerns counter-

factuals what if scenarios rather than

actual facts. It is thus inevitable that both

opponents and proponents argue by analogy

and assertion, extrapolating from other

times and contexts to make their case. Far

from inevitable, however, is the careless

extravagance of much of the argument.

The treacherous terrain of counterfactuals

demands that we be cautious about the

weight we place on our speculative asser-

tions. It also requires us to be sensitive to

the inherent asymmetry of such debates.

Given that the Census does not count

caste today and that the debate is about

whether it should, the burden of proof is

unequally distributed. Those who want a

change in the status quo have to shoulder

a heavier load than those who are contentwith the way things are. This is just the

way that arguments about counterfactu-

als are structured, and it would be bad

faith for pro-changers (like us) to claim

underdog status.

However, an important but often invisi-

ble aspect of this structural asymmetry is

that the status quo escapes strict scrutiny.

Without such scrutiny, both sides tend to

calibrate their arguments as though the

status quo represented a neutral state, a

sort of zero-point, against which the pos-

sible negatives and positives are meas-

ured. But this need not be true the status

quo does not default to neutral just be-

cause we are debating counterfactuals. If

the present already has a strongly nega-

tive or positive value, the impact of the

future needs to be assessed differently. In

short, we must re-calibrate our present,

and to do that we need to also ask ques-

tions like: What kinds of damage has India

suffered because a caste census has notbeen held since Independence? What is

the politics ofnot counting caste?

While the main objective of this essay is

to ask questions of this sort, the following

sections revisit the debate on the caste

census in an attempt to identify the im-

plicit assumptions and tacit contrasts that

shape the most common arguments onview. As proponents of a caste census we

are certainly not neutral, but we hope to

show that it is not the presence of bias

whether our own or of the others that

makes for bad arguments, but rather the

absence of care.

Implicit Models of Caste

and Census

At the risk of oversimplifying, the most

common arguments against a caste cen-

sus are of two broad kinds those that in-

voke political-moral grounds, and those

that cite insurmountable practical-logistical

difculties. The most common pro-caste

census arguments tend to be mostly politi-

cal-moral, with some practical-logistical

counter-arguments against the claims of

opponents. Though they may often appear

so, pro and con arguments are not neces-

sarily mirror images of each other. How-

ever, both are based on implicit ideas

about the nature and role of caste andthe Census.

At least three distinct models of what

the Census is and ought to do are at work

behind these arguments. The rst sees the

Census as an extension or analogue of a

welare programme or social justice initia-

tive. In this view, the Census is an instru-

ment for rationalising such programmes,

making them more efcient and effective.

A second model of the Census is that of a

device or fxing identities, or creating

compulsory identities. In this view, the

Census will forever x the caste identity of

every citizen and thus comprehensively

sabotage the project of creating a univer-

sal Indian citizen. Less common than the

previous two but signicant in its own

right is a third model of the Census as a

gigantic research project designed to pro-

duce the truth about the categories it

counts. In this view, the Census is useful

only if it is accurate and truthful; other-

wise it is a waste or a liability.While these three are the most com-

monly employed models, there is also one


2/4

PERSPECTIVE

June 19, 2010 vol xlv no 25 EPW Economic & Political Weekly40

other model that has either been absent or

only eetingly present in the English me-

dia, but has received more play in the Hindi

press (as has the caste census debate more

generally). This fourth model of the cen-

sus, familiar to us from the literature on

nationalism, is that of a collective sel-por-

trait. Like maps, ags and other mnemon-ic devices, the Census is a representation

of the nation; it helps us to concretise an

abstractly imagined national community.

Seen from this angle a Census is not just

about social justice programmes, or xing

identity, or a source of knowledge it is a

collection of our collective identities.

Because it is available to all, it offers a

chance for all groups to look at themselves

in relation to others. Only the Census can

provide such an aggregated, comprehen-

sive picture of the collectivities that com-

prise the nation no individual or group

has the resources or the power to do this.

A comparable effort to list the different

models of caste that seem to be at work

would yield the following: The rst and

probably most common model is the one

in which caste actually means lower caste,

roughly in the same way that gender

comes to mean women and race comes

to mean non-whites. In this perspective,

caste is about the concerns and problemsof the lower castes, including especially

reservations, quotas and vote banks. An-

other model of caste heavily inuenced by

social anthropology sees it as a complex

meaning-giving institution of great impor-

tance in ordering everyday life. This per-

spective is concerned to emphasise the

complexity of caste and its irreducibility

to other social structures like class or eth-

nicity. A third perspective sees caste as a

web o distributional relations that deter-

mines the distribution of power, privilege

and material resources in conjunction

with class. From this perspective, caste is

necessarily relational the parts do not

make sense outside of the whole they t

into, although t need not imply harmo-

ny and is compatible with conict. Finally,

a fourth perspective on caste sees it as the

single most important obstacle to attaining

modernity. Caste is a peculiarly Indian

afiction, one that is based on clearly

un-modern values and it prevents us frombecoming fully modern and embracing

values of individualism and universality.

These sketchy and rather brief models

of census and caste are clearly too rudi-

mentary for too much weight to be placed

on them. They are not meant to be ex-

haustive it is surely possible to think of

others but only to indicate the variety of

available vantage points. They are all par-

tial, in the sense that the existence of onehardly precludes the others. They are also

partial in the sense that they are more or

less hospitable to different perspectives on

caste, and, more specically, to different

positions vis--vis a caste census. In other

words, models of caste and census com-

bine to structure arguments for or against

a caste census. One could also say that

they determine the possibilities and limits

of these arguments. It is not our conten-

tion that there is any neat or necessary

correspondence between these models;

rather, our main point is that, when ana-

lysing arguments in this debate, it is use-

ful to look for the implicit models that may

be animating them.

We now turn to the most common argu-

ments encountered in the caste census de-

bate. As proponents, we pay special atten-

tion to the arguments of opponents, sav-

ing a brief description of our prefered

pro argument for the conclusion.

Logistical Challenges

Perhaps the most common argument

against a caste census, one specially fa-

voured by academics, is that it simply can-

not be done. There are many strands to

this broad argument and they need to be

carefully sifted. The thickest strand com-

bines the research project model of the

Census with the complex-meaning-giving-

institution model of caste to argue that the

capabilities of the former are much too

meagre to capture the many-sided intrica-

cies of the latter.

One variant of this strand insists that

caste is too uid and polyvalent an identity:

the question, What is your caste? can

have more than one (sometimes several)

context-dependent answers. This usually

arises from the fact that, at the micro end,

caste usually subsumes within it other

distinct entities like sub-caste, sub-sub-

caste, etc, and at the macro end may itself

merge with other castes to form a largercaste-group (such as Maratha or Lingayat,

or broader still, Kshatriya for example).

I may answer the question differently de-

pending on whether I am seeking a bride

for my son, seeking a favour from some-

one, or deciding who to vote for. While

this is quite plausible, it is still difcult to

see why or how this poses a problem for

the Census. The typical Indian respondent

is not likely to be in any fundamental exis-tential doubt about her caste, for this is a

luxury available only to the upper caste

urban elite. As for the context-dependence

of caste names, this is not a major problem

precisely because the census enumerators

arrival at my door itself represents a very

specic context, and my response will

simply be whatever I believe to be appro-

priate for this context. The fact that I may

have responded differently in other con-

texts is irrelevant here.

In fact, one can go further to argue that

the Census should in fact ask about syno-

nyms for caste names where they exist

and are relevant. Thus, in addition to a

question such as What is the name of

your caste? there could be a follow-up

question like Is your caste known by any

other names in your locality? Finally, as

an additional aid to future tabulation and

grouping, one could also ask for the family

surname, if any. (Assuming that this is a

household level question asked of thehead of the household alone.)

It must not be forgotten that since 2001

the Census has had access to technology

far superior to that of earlier editions. In

particular there is now character recogni-

tion software that enables scanning and

digitisation of hand-written forms, very

inexpensive digital storage media, reliable

methods of data retrieval, and, above all,

the possibility of retaining raw data down

to the unit level, so that downstream pro-

cedures for grouping and consolidation

are reversible in a costless manner. What

this means in practice is that synonyms

are not a problem, nor is the gathering of

additional information, which is con-

strained only by the time available to enu-

merators to canvas each schedule. Unlike

what is often implied by opponents, syno-

nyms are certain to be centripetal rather

than centrifugal, that is, they will be tight-

ly concentrated around a central core

rather than diverging greatly from it. It isvery unlikely that persons with some local

knowledge (like the local schoolteacher


3/4

PERSPECTIVE

Economic & Political Weekly EPW June 19, 2010 vol xlv no 25 41

who is the most common enumerator) will

be unable to recognise the family resem-

blance common to synonyms.

Given that the admittedly difcult job

of collation, grouping and consolidation

need only be done later and post-digitisa-

tion, signicant economies of scale can be

harnessed. For instance, Census 2001 setup two special task forces in the ofce

of the Registrar General of India to over-

see this process for the religion and the

caste (SC and ST) data. The same can be

done for caste data, and this task force

can access local knowledge and compare

the data with past records in order to

suggest the most prudent modes of con-

solidation. Should errors be pointed out,

digital data bases ensure seamless and

costless corrections.

At a different level, while the Census

ought to be accountable for the reliability

and validity of its data, the standards of

this accountability must be determined

with reference to its role and function. It

would be absurd to make the Census ac-

countable in the same way and to the

same extent that an individual ethnogra-

pher is held accountable for her data. For

each produces truths that the other can-

not, nor does one falsify the other.

Large Numbers?

Another puzzling argument presents the

very large number of castes that will be

returned as a self-evidently decisive ob-

jection. For example, the 2001 Census

enumerated a total of 1,234 castes in the

SC category and 698 tribes in the ST cate-

gory. Incidentally, between them, the

1971, 1981 and 1991 Censuses returned a

total of 1,700 religion names, which were

analysed by a working group set up for

this purpose. Do these numbers make the

SC-ST or the religion data useless? In a

country of 1.2 billion people that is 12 fol-

lowed by eight zeroes large numbers are

likely to be the rule rather than the excep-

tion. As such, they do not signify anything

in and of themselves.

An implied argument could be that such

large numbers are hard to handle in the

sense of comparing, contrasting and so

on. The problem here is that the starting

point is itself articial. The fact that thereare 1,234 SCs in India has no meaning at

the all-India level because the schedules

are territorial, that is to say, SCs and STs

are recognised only in a specic geo-

graphical context. In fact, a useful gener-

al principle for a caste census is to insist

on a conservative strategy for aggrega-

tion. The district or even the sub-district

or taluka level could be set as the default

threshold, with aggregation beyond thispoint having to be clearly justied. The

availability of disaggregated data will

enable informed debate and act as a

built-in antidote to the misuse of data.

This also makes the Census the only

possible source for this level of disaggre-

gation, since no other data source, not

even the National Sample Surveys, can

provide meaningful sample sizes at the

district level.

The main point, however, is not simply

to prove that counting caste is logistically

and practically feasible, which it clearly is.

Rather, it is that the Census is not an in-

strument for producing the truth about

caste. With variables like caste, religion,

or language the Census effectively records

what respondents wish to be recorded,

subject to local consistency checks, and

this is how it should be.

The fears of falsication of census data

are often exaggerated and not thought

through carefully. Some commentatorseven go so far as to suggest that people

would simply choose what caste to re-

turn themselves as, based on expectations

of personal gain. This is a rather curious

view because the caste that you name in

the Census entitles the individual re-

spondent to nothing at all. Benets such

as reservations, etc, depend upon a caste

certicate, something which is entirely

unrelated to the Census. Given that the

Census is counting what people would like

to be counted, its main purpose is precisely

to perform this act of aggregation that no

individual or group can do. In this sense,

the Census is indeed about that elusive

level of reality called the social, which

emerges into view only through aggrega-

tion. Though comprised only of individu-

als, this aggregate turns out to be much

larger in social signicance than the sum

of its parts. Given all this, it is somewhat

beside the point to worry about false in-

formation; indeed, it is hard to gure outwhat false and true might mean here.

In this sense, and with respect to variables

like caste or religion, the Census is some-

thing like a mandatory opinion poll.

Political Objections

This brings us to the political objections to

a caste census, the lead argument in

which is the claim that to count caste is to

return to the colonial era of divide andrule. Even within the framework of this

argument it is clear that the colonial census

did not only divide but provided powerful

support for nationalism and the idea of

India. On the other hand, one needs to be

clear about where one is standing when

talking of divide and rule. From the per-

spective of the vast majority of the Indian

population subjected to centuries of elite

rule, the existence of a higher power ena-

bled an otherwise improbable revolt

against native authority. It is only from the

perspective of an elite that sees itself as

the owner of the nation that any and all

subaltern claims for power sharing will

seem divisive. Finally, the contemporary

context needs to be stressed. What could

divisiveness mean in the 21st century

when there is simply no future in seces-

sionism and the only agenda is power

sharing within existing state jurisdictions

or boundaries? In short, this is a very dif-

ferent time from the mid-20th centurywhen new nations were being born every-

where. It is not just those who are labelled

as divisive who have vested interests; calls

for unity are seldom innocent and are

often anything but altruistic.

Other arguments stress the ways in

which a caste census will turn into a mess

of political manoeuvring, as everyone

attempts to choose the caste identity that

provides maximum advantage. Here again

it is necessary to reiterate the point that it

is not the task of the Census to capture the

pure truth of caste apart from what peo-

ple say they are. The Census cannot stand

outside of the politics of its time. But what

is often forgotten is that the politics of

caste identities is not innitely uid and

malleable, whether at the level of the indi-

vidual or the group. For most Indians, caste

is an interrelational identity embedded in

the politics of everyday life. It is only from

the perspective of those who do not live

such an identity, that it can somehow beturned into something purely instrumental

and volitional. At most, attempts will be


4/4

PERSPECTIVE

June 19, 2010 vol xlv no 25 EPW Economic & Political Weekly42

made by already active social movements

and groups to try and use the Census to

their advantage, but the Census will not be

the creator of such movements. If, indeed,

large numbers of people are claiming to be

another caste, then perhaps it is time that

the Census record this. This is also why it is

necessary to ask people in more than oneway about their local caste identity.

The politics of counting religion might

be instructive about the possible effects of

counting caste. In 1995 when the 1991

Census population tables based on reli-

gion were rst made public, right wing

Hindu groups (and scholars) tried to use

this data to prove that Muslims would

soon outnumber Hindus. These were the

years of the rise of the Hindu right in all its

ferocity. The subject of the rate of growth

of Muslims relative to others has since g-

ured in many contexts, political and aca-

demic alike, and has seen different argu-

ments being made, including demograph-

ic evidence that, at every socio-economic

level, Muslim families on average have

one child more than their Hindu counter-

parts. A decade later, the same Census

data has been put to very dif ferent use by

the much cited report on the Economic

and Social Status of Muslims in India,

popularly known as the Sachar Commit-tee Report. This report has helped engi-

neer a shift in the public perception of

Muslims as a pampered minority to that

of a community that in social, educational

and economic terms is one of the most dis-

enfranchised in the country. This, too, is

politics, if clearly for the better.

Breaking with Caste Blindness

We come now to the models and argu-

ments that we ourselves favour. Our cen-

tral point is that not counting caste has

been one of independent Indias biggest

mistakes. Perhaps this mistake could not

be avoided as it only became visible retro-

spectively. But at least for the last two dec-

ades, the damage wrought by this policy

of caste-blindness has been plainly visi-

ble for anyone who cares to see it. Most

arguments against a caste census treat the

implicit contrast with the status quo as

though it, were neutral it is not. In fact,

arguments against a caste census need toask not only if we would be worse off with

it, but also if we are better off without it.

The model of the Census we are partial to

is that of a collective self-portrait, which,

along with a model of caste that empha-

sises its role as a distributional axis, yields

the argument outlined below.

The most important reason to ask for a

caste census is because it offers the op-

portunity to break with the model ofcaste blindness that the Indian state and

mainstream polity has followed since in-

dependence. Starting from the premise

that caste was to be singled out for aboli-

tion, the notion of caste blindness com-

bined the formal abolition of caste in the

Constitution with what amounted to a

ban on public discussion of caste. The STs

and SCs were treated as a regrettable ex-

ception to the rule of caste blindness. How-

ever, the formal abolition of caste was not

accompanied by serious attempts to abol-

ish its substantive privileges and disprivi-

leges. This half-hearted caste blindness

enouraged the perpetuation and deepen-

ing of caste inequalities under a supposedly

casteless Constitution. Half a century of

this perspective brought us to Mandal, or

the point where a large plurality of lower

castes could no longer ignore the contra-

diction between their political entitle-

ments and their actual share in the nation.

The ip side of rising caste conscious-ness among the lower castes was the fos-

tering of the upper caste belief that they

had left caste behind and were now caste-

less. For the most privileged sections of

the upper castes this was true in a certain

sense because three generations of caste

blindness had allowed them to fully

encash their caste advantages. They were

now in a situation where they no longer

needed to invoke caste explicitly, having

acquired all the other resources that

guaranteed them the legitimate advan-

tages of inherited wealth, expensive edu-

cation and abundant connections among

their own kind. It is these groups of upper

castes who are the most vociferous

advocates of caste blindness today. It is

they who believe that the Census is mainly

about and for the lower castes and their

squabbling about quotas. The collective

portrait model of the Census insists that

everyones caste be counted and that the

upper castes be denied the anonymity thatthey have enjoyed under caste blindness.

In this sense, a caste census can mark the

end of a remarkably unsuccessful phase of

Indias attempt to transcend caste and in-

augurate a fresh initiative.

In our view, therefore, arguments in fa-

vour of a caste census should begin by de-

tailing the harm that its absence has done

and is doing to our ambitions of annihilat-

ing caste. Today we are in a situationwhere caste is exclusively associated with

Indias lower castes, specially dalits and

adivasis, who must prove their member-

ship as SCs and STs by giving identiable

caste and tribe names. We have been ar-

guing that this situation must change in

favour of a fuller, more inclusive picture,

where everyone must answer the question

of their caste. This is by no means an en-

dorsement of an unequal system. Saying

one has no caste is an option for those

who wish to exercise it. However, it is in

the interests of a democratic politics that

the false image of caste as the exception

suffered only by its most discriminated

citizens give way to a situation where we

are willing to recognise that caste is rela-

tional. There is no caste disprivilege with-

out a corresponding privilege accruing to

some other caste. It is this mutual connect-

edness that we have denied so strenuously

throughout our post-colonial history.

It is time to recognise that contrary toKabirs famousprem gali (or alley of love

and devotion) which admits only one at a

time, the alley of caste requires that we

enter and leave together. A universal caste

census for all Indians will be a tting ac-

knowledgement of the undeniable fact

that, when it comes to caste and its injus-

tices, our complicity and our redemption

are both inescapably mutual.

the+politics+of+not+counting+caste+ +epw

Documents