Ethnic Cleansing in Pakistan during Partition: A Preliminary
Statistical Analysis
Sridhar N.
Introduction
The partition of India into the two countries
of India and Pakistan in 1947 is a topic that
has been much studied and written about.[1]
In particular, there have been a number of scholarly studies, books
and films on the human tragedy that accompanied partition. A
very large number of people lost their lives in the orchestrated
violence that took place for sustained periods of times and an even
larger number of people were uprooted from their ancestral homelands
and forced to migrate to new and unfamiliar places.
Researchers, authors and filmmakers have explored a variety of issues
related to this tragic period of our history.
Of particular interest to us in this
article is the issue of ethnic cleansing of minorities –
specifically the Hindu and Sikh communities – in the newly
formed country of Pakistan. From the very beginning of the movement to create Pakistan, the proponents of this
movement espoused a blatantly communalist ideology – the two
nation theory. The ethnic cleansing of the minorities with
active participation of the rulers of the new state was but a logical
extension of the theory that said that Hindus and Muslims had
irreconcilable differences and could not live together.
The topic of ethnic cleansing in Pakistan has received a lot of
attention over the years.[2]
Yet, there seems to be a surprising paucity of hard data on the
extent of this ethnic cleansing except in the academic literature.[3]
There seems to be a wide diversity of numbers on how many people were
killed or forced to migrate from their homelands.[4]
[5]
Many authors state that approximately 0.5-2 million people were
killed and 10-17 million forced to migrate to either India or Pakistan.[6]
However, authors rarely give a breakup of these numbers by
country. Further, blame for ethnic cleansing is often equally
assigned to both countries.[7]
Often, no sources are quoted to substantiate these claims.
The need for this article flows from this
paucity of data on the partition of India and particularly in non-academic
articles easily accessible on the internet. The objectives of
this analysis would be modest but it would hopefully serve as a
source of credible data on the topic. The specific questions
that this statistical analysis would attempt to answer are:
- What was the extent
of ethnic cleansing of minorities in Pakistan
- What does this mean
in relative terms (i.e. proportions of populations) and in
absolute terms (i.e. the actual number of people)
- Were there regional
differences in the extent of ethnic cleansing of minorities in Pakistan?
- How did the
situation in this respect compare with that in India?
The main source of data for conducting
this analysis would be the census reports of undivided India in 1941 and the data
from the censuses of India and Pakistan in 1951. The
reason for using these data is that these are credible and cannot be
accused of bias.
Data Description
The data used in this analysis is the set
of population counts for the different communities in the two countries
– India and Pakistan - in the years 1941 and
1951. These numbers were obtained from the census report of
undivided India in 1941 and the two
censuses of India and Pakistan in 1951. The
census reports themselves are fairly easy to obtain, but the data involves
a complication. The 1941 census data is available only for the
undivided provinces of Punjab, Bengal and Assam and not for the units that became a part of India and Pakistan (the other provinces
did not pose a problem since entire provinces went to one or the
other country). Thus, district-level and even tehsil-level data
would have to be obtained and tabulated to separate the 1941 data for
these provinces for the two parts that respectively went to India and Pakistan. Fortunately, the
Census of Pakistan 1951 has already done this tabulation for the 1941
census data for the different religious communities in territories
comprising Pakistan then. The
corresponding numbers for India are then just the difference of the respective
numbers for undivided India and Pakistan.
The 1951 census in India as well as Pakistan were conducted in the
last three weeks of February of 1951. This, combined with the
fact that the 1941 census for undivided India took place at the same
time in the territories that were to later comprise India and Pakistan, allows us the
opportunity for comparisons without the need for statistical
adjustments. Thus, these data provide us a sound basis for
preliminary analysis.
The raw census data used in this study
are given in Appendix I.
Proportions of
Minorities
In order to investigate whether there was
ethnic cleansing at all in Pakistan or not, it would be useful to compare the
proportions of minorities in the territories that comprised Pakistan in the years 1941 and
1951. A drastic reduction in the proportions of minorities
gives prima facie evidence that there was ethnic
cleansing. It would not be conclusive and other pieces of
evidence would be required to show that it was an organized campaign
to drive out the minorities from the country. However, it would
be a good starting point.
Figures 1 and 2 show the relative
proportions of minorities in the territories that comprised West Pakistan and East Pakistan respectively in the
years 1941 and 1951. Figure 3 shows the relative proportions of
minorities in the whole of Pakistan in 1941 and 1951. We have aggregated all
the non-Muslim minorities together for this analysis. Some
patterns are clear from this analysis. West Pakistan shows prima facie
evidence for ethnic cleansing. The proportions of minorities
reduced from over 1/5th of the population in 1941 to a
negligible level in 1951. The pattern for East Pakistan is somewhat
different. While there was a substantial reduction in the
proportion of minorities between 1941 and 1951, it does not
demonstrate evidence for ethnic cleansing at least in this
period. We shall investigate these differences and
possible reasons for these differences in subsequent sections of this
article.
Figure 1: Comparison
of Ethnic Mix - 1941 and 1951
West Pakistan
|
Figure 2: Comparison
of Ethnic Mix - 1941 and 1951
East Pakistan
|
Figure 3: Comparison
of Ethnic Mix - 1941 and 1951
Pakistan (West Pakistan + East Pakistan)
|
The numbers in Figures 1 through 3 also
constitute a credible and indisputable source for the proportions of
minorities in 1941 and 1951 in Pakistan and in the two wings of the country. This,
in my view, is an important contribution of this article since
existing articles on the subject quote a wide variety of numbers for
the proportions of minorities, usually without backing them with any
credible source.
To investigate these numbers further, we
now compare the proportions of Muslims and non-Muslims in the
different provinces of West Pakistan (with East Pakistan constituting a province
by itself). The current configuration of provinces in
present-day Pakistan is used for the purpose
of this analysis. The four provinces (with the configuration in
1951 in parenthesis) are Punjab (Punjab + the princely state of Bahawalpur), Sind (Sind + the princely state of
Khairpur + Karachi federal area), NWFP[8]
(NWFP Settled Districts + NWFP States) and Baluchistan (Baluchistan Districts + Baluchistan States). Currently,
there is also an area called the Federally Administered Tribal Areas
(FATA). These areas were part of the ‘NWFP States’
in 1951 and hence are included in the NWFP numbers in our
analysis. Figures 4 through 7 present the analysis for these
four provinces respectively.
Figure 4: Comparison
of Ethnic Mix - 1941 and 1951
Punjab Province (Punjab + Bahawalpur)
|
Figure 5: Comparison
of Ethnic Mix - 1941 and 1951
Sind Province (Sind + Khairpur + Karachi Federal Area)
|
Figure 6: Comparison
of Ethnic Mix - 1941 and 1951
NWFP (NWFP Settled
Districts + NWFP States)
|
Figure 7: Comparison
of Ethnic Mix - 1941 and 1951
Baluchistan (Baluchistan
Districts + Baluchistan States)
|
The analysis of the minority proportions
for the provinces of erstwhile West Pakistan reveal that there is prima facie evidence
for ethnic cleansing in all the four provinces, though its degree
differs by province. NWFP had the worst degree of ethnic
cleansing – while its minority population was the smallest
amongst the four provinces in 1941, it managed to almost completely
rid itself of all minorities living there. Punjab came close to NWFP in
the scale of ethnic cleansing and in some senses it was worse since
it had the largest minority populations amongst the four provinces
– the minorities in Punjab constituted 71.28% of all minorities in Pakistan. Baluchistan also reduced its
minority population from about 8.5% to about 1.5% of its
population. The best situation was in Sind, where the minorities
continued to constitute about 8.5% of the population in 1951.
However, there is evidence for ethnic cleansing even in this province
since it had a high 28.6% minority proportion in 1941. Also, a study
that ends in 1951 does not tell the complete story since ethnic
cleansing in Sind continued after 1951.
To summarize the discussion in this
section, that there is evidence for ethnic cleansing of minorities
and particularly of Hindus and Sikhs in Pakistan and particularly in West Pakistan. This seems to
have happened in all four provinces of West Pakistan, though its
degree was the greater in NWFP, Punjab and Baluchistan than in Sind.
Absolute Population
Numbers for Ethnic Cleansing
The analysis till now was based on
proportions of minority populations in Pakistan in 1941 and
1951. However, this does not give us a picture of how many
people are likely to have been killed or forced to migrate from
Pakistan to India during partition. The main problem that we
run up against when trying to estimate these numbers is that we do
not have any numbers for 1947 – immediately before the killings
and migration began. There were special censuses conducted in
India in 1948 and 1949 but none in Pakistan. Additionally, the
1948 and 1949 numbers only give a picture of the situation after
partition. We do not have numbers for the situation before
partition. Thus, we have no way of separating the numbers for
killings and migration from those for natural growth of population of
the various communities.
Another way to analyze this is to derive
estimates for the populations of each ethnic community in Pakistan in
1951 under the hypothetical scenario that there had been no ethnic
cleansing. Comparing this hypothetical population with the
actual populations in 1951 would give us estimates of the gap.
This estimation needs to be done using extrapolations of 1941
populations. We run into another difficulty in this
extrapolation. We have no way of knowing the growth rates of
different communities in Pakistan because the growth rate is confounded with
movements of large masses of people in both directions.
Therefore, we compute the growth rate of each community for undivided
India as a whole. For
1951, we add the populations of each community in both India and Pakistan and compare them with
their respective populations in 1941 to obtain the growth rates for
each community for undivided India. We make the assumption that the growth
rates were the same in the two countries in order to project the
population for each community in 1951 in both the countries.
A more sophisticated analysis could be
done if we had data on populations of different communities for Pakistan and India in the censuses before
1941 and those after 1951. This would have allowed us to use
time-series methods or other statistical techniques to more
accurately estimate the populations of the various communities in the
two countries under the above-mentioned hypothetical scenario.
This involves a large amount of data collection (district level data
would need to be aggregated for every census before 1941). That
is a topic for future research. However, in the absence of
these data, the assumption of equal growth rates in the two countries
is reasonable.
Figures 8 through 10 show the results of
this analysis for West Pakistan, East Pakistan and Pakistan. The two numbers compared in each of
these graphs are the hypothetical population of non-Muslims in 1951
if there were no ethnic cleansing and the actual numbers according to
the 1951 census. The difference between these two numbers would
constitute those who were killed or forced to migrate to India. A limitation of
this analysis is that it cannot separate out the numbers of those
killed but can only look at the sum total of these two. Also,
it must be borne in mind that these numbers critically depend on the
assumption on growth rates. Yet, they give an approximate
picture for the extent of ethnic cleansing.
Figure 8: Comparison
of Hypothetical vs. Actual Minority Populations in 1951
West Pakistan
|
Figure 9: Comparison
of Hypothetical vs. Actual Minority Populations in 1951
East Pakistan
|
Figure 10: Comparison
of Hypothetical vs. Actual Minority Populations in 1951
Pakistan (West Pakistan + East Pakistan)
|
The key number of importance for this
analysis is the difference between the hypothetical and actual
populations of non-Muslims in 1951. As seen in figures 8
through 10, this difference is about 7.76 million for West Pakistan, 6.93 million for East Pakistan and about 14.69 million
for Pakistan as a whole. This
number is considerably different from the statistics that have been
used till now. The 1951 census for India reported 4.699 million
displaced people from West Pakistan and 2.549 million people from
East Pakistan and a further 0.047 million who were displaced people
from an unspecified location (i.e. they did not indicate whether they
came from East Pakistan or from West Pakistan).[9]
These numbers for displaced people are vastly different from our
numbers of the difference between hypothetical and actual populations
in 1941 . Some potential reasons for this difference come to
mind. One number that the difference between hypothetical and
actual populations includes but the census count of displaced people
does not include is the number of people who were killed in the
violence during partition. In addition, there were reports of
conversions of large number of minorities (particularly Hindus and
Sikhs) to Islam,[10]
[11]
though there are no reliable estimates for how many conversions took
place. Thus, the difference could represent the numbers killed
or converted. The numbers in our analysis constitute relatively
reliable estimates for these numbers. If we take these numbers at
face value, it would seem that about 7.46 million people were either
killed or converted into Islam in Pakistan as a whole during
partition. The numbers for West Pakistan and East Pakistan are
respectively 4.69 million and 2.54 million. Table 1 depicts
these numbers in detail. Since the typical numbers for those
killed in Pakistan during partition range from half a million to a
million people, these numbers might suggest that a very large number
of people were forced to convert to Islam due to the circumstances
they found themselves in. While there has been anecdotal
evidence[12]
as well as popular writings[13]
for this, we are not aware of attempts to quantify this. It
must be noted however, further research is required to rule out
alternative explanations for the difference between our estimated
numbers and the number of displaced persons reported in the Census of
1951.
Table 1: Displaced
Persons and Our Estimates for Ethnic Cleansing: Pakistan
Place from where the
refugees originated
|
Number of Displaced
Persons from Pakistan to India (million)
(Source: Census of
India 1951)
|
Difference between
Hypothetical and Actual Populations
(million)
|
Difference : Persons
unaccounted for
(million)
|
A
|
B
|
B-A
|
West Pakistan
|
4.69
|
7.76
|
3.06
|
East Pakistan
|
2.54
|
6.93
|
4.39
|
Pakistan
|
7.28
|
14.69
|
7.41
|
Note: The numbers for West Pakistan +
East Pakistan in the Census 1951 numbers (in column A) do not add up
exactly to the numbers for the whole of Pakistan since there were
0.047 million displaced people for whom the place of origin is
unspecified.
An interesting analysis would be to
calculate the numbers of displaced people and those that we
conjecture were killed or converted as proportions of the
hypothetical population of minorities in Pakistan as a whole and its
two wings. This would give a fairly accurate picture of what
proportion of people were ethnically cleansed. It is a cleaner
measure than the comparisons of proportions of minorities in the
population in 1941 and 1951 (depicted in Figures 1 through 7) because
that is sensitive to the number of incoming (Muslim) migrants into
Pakistan. These proportions are shown in Table 2.
Table 2: Displaced
Persons and Our Estimates for Ethnic Cleansing: Pakistan
Proportions of
Hypothetical Populations in 1951
Place from where the
refugees originated
|
Number of Displaced
Persons from Pakistan to India
|
Difference between
Hypothetical and Actual Populations
|
Difference : Persons
unaccounted for
|
(% of total minority
population)
|
(% of total minority
population)
|
(% of total minority
population)
|
West Pakistan
|
53.79%
|
88.93%
|
35.15%
|
East Pakistan
|
15.27%
|
41.67%
|
26.41%
|
Pakistan
|
28.70%
|
57.92%
|
29.23%
|
The numbers in table 2 show that almost
89% of the minorities in West Pakistan were ethnically cleansed, i.e.
killed, converted or driven out of the country. Almost 54% of
the minority population ended up as refugees in India, while a very
high 35% of the minority population is simply unaccounted for.
These are people who were likely killed or converted into
Islam. The situation was different in East Pakistan, but there
are also points of similarity. While only 15% of the minority
population of East Pakistan ended up as refugees in India by 1951, as
much as 26% of the population is unaccounted for – i.e. killed
or converted. While the number of refugees as a proportion of
minority population differs substantially from those in West
Pakistan, the numbers unaccounted for is in the same ballpark.
This contrast requires further analysis to understand some of the
factors responsible for it. We shall attempt to speculate on
some potential factors in subsequent sections. In sum, this
analysis suggests that one in three people belonging to the minority
community in Pakistan was either killed or forced to convert to
Islam. We are not aware of any study that discusses this
shocking suggestion and backs it with evidence. It definitely
requires far greater attention from the academic community than has
been given till now.
Situation in India
Partition caused violence on both sides
of the newly created border in 1947. There were riots on both
sides and migrants streamed into both countries. However,
authors have often assigned equal blame to both sides in terms of
ethnic cleansing.[14]
It is not our case that there was no violence against Muslims
in India in the days preceding and after partition. However, we
shall demonstrate using an analysis similar to the one done for
Pakistan that the situation in India was drastically different from
that in Pakistan. While there was violence and while a large
number of people migrated across the borders, there was no ethnic
cleansing of the kind witnessed in Pakistan and particularly West
Pakistan. Please note that unless we specifically refer to
‘undivided India’ when we refer to India, we imply those
territories that went on to constitute India in 1947.
In the case of India, our analysis would
focus on Muslims, since they were the main aminority migrating to
Pakistan and it can be safely assumed that there were no non-Muslim
migrants to Pakistan. Figure 11 depicts the proportion of
Muslims and non-Muslims in the censuses of 1941 and 1951 in the
territory that went on to constitute India in 1947. The
proportion of Muslims in the population of India declined from 13.24%
to 9.70%. While this shows a decline, it is nowhere close to
the numbers we saw earlier for Pakistan. It may be useful to
recall that in West Pakistan, the proportion of minorities fell from
20.91% to 3.36% in the same period.
Figure 11: Comparison of
Ethnic Mix - 1941 and 1951
India
|
Similar to our analysis of hypothetical
populations in 1951 in Pakistan, we project the Muslim population of
India from in 1941 to estimate the Muslim population in 1951 if there
were no partition, with its associated violence and migrations of
people. The key number of interest, as before is the difference
between the hypothetical population and actual population of Muslims
in 1951 in India. This analysis is shown in Figure 12.
Figure 12: Comparison
of Hypothetical vs. Actual Muslim Populations in 1951
India
|
The difference between the hypothetical
and actual Muslim population of India in 1950 is about 7.66
million. This number represents the total number of Muslims who
migrated to Pakistan or were killed in violence. Unlike in the
case of Pakistan, there are no reports of forced conversion of people
from Islam into other religions in India during partition, either in
contemporaneous publications of those times or in subsequent
writings, including by Pakistani authors.
It would be useful to see, as we did in
the case of our analysis of Pakistan, how our estimated difference
between hypothetical and actual population of Muslims in India
compares with official statistics on displaced people in
Pakistan. The Census of Pakistan 1951 reports a total Muhajir
(refugee from India) population of 7.23 million.[15]
This number is remarkably close to our estimated difference between
hypothetical and actual populations of about 7.66 million.
Compared to about 7.41 million unaccounted for people from the
minority communities in Pakistan, there are about 0.43 million
unaccounted for Muslim people in India in 1951. The fact that
the numbers based on our estimate is close to the actual number of
refugees as per the Census of Pakistan gives us confidence on the
robustness of our methodology. These numbers are in Table 3.
When we look at these numbers as
proportions of the hypothetical Muslim population in India in 1951,
this contrast becomes even starker. Table 3 again has these
numbers in terms of proportions of the hypothetical population.
The number of Muslims who migrated from India to Pakistan or got
killed in violence was 17.95% of the total hypothetical Muslim
population of India in 1951 as per our estimate (i.e. hypothetical
minus actual Muslim population in India in 1951). Recall that
in West Pakistan, 88.93% of the hypothetical minority population was
found to have been killed or forced to migrate by an identical
analysis. This number was 57.92% for the whole of
Pakistan. The proportion of Muslims in India who migrated to
Pakistan or were killed is lower not just than the corresponding
numbers for West Pakistan but also those for East Pakistan
(41.67%). Thus, while there is evidence for ethnic
cleansing of minorities in West Pakistan in particular, there is no
evidence for the same in India.
Table 3: Displaced
Persons and Difference between Hypothetical and Actual Populations:
India
|
Number of Displaced
Persons from India to Pakistan
(Source: Census of
Pakistan 1951)
|
Difference between Hypothetical
and Actual Populations
|
Difference :
Persons unaccounted
for
|
A
|
B
|
B-A
|
Actual Numbers
(million)
|
7.23 million
|
7.66 million
|
0.43 million
|
Proportion of
Hypothetical Muslim Pop. in 1951 (%)
|
16.95%
|
17.95%
|
1.00%
|
Even more interesting is the result on
the number of people unaccounted for, i.e. the difference between our
estimate of those who migrated or were killed and the official census
number for migrants. The number of such unaccounted for Muslims
in India is 0.43 million or 1.00% of the hypothetical Muslim
population in India in 1951. Thus, while a large number of
Muslims can be believed to have been killed in violence in India
during partition, they constituted 1% of the total population.
This is a large number, but is not even close to being comparable to
the corresponding number in Pakistan – a shocking 35.15% in
West Pakistan and 29.23% in Pakistan as a whole.
Summary of Results
The following points summarize the
results of the analysis we have conducted till now.
- There was a drastic
drop in minority population as a proportion of the total
population in Pakistan after partition. In West Pakistan, the
proportion of minorities came down from almost 21% of the
population to about 3% between 1941 and 1951. Even in East
Pakistan, this number came down from about 29.5% to about
23%. The drop in minority populations was high in all four
constituent provinces of West Pakistan, though it was highest in
Punjab and NWFP and lowest in Sind.
- Almost 15 million
people belonging to the minority communities were either killed
or converted to Islam or forced to migrate to India during
partition. This number is substantially higher than usual
estimates.
- Of these 15
million, about 7.28 million migrated to India and the rest are
unaccounted for. Presumably, they were either killed or
converted into Muslims.
- As much as 89% of
the minority population in West Pakistan was affected by ethnic
cleansing, i.e. killed, converted into Islam or forced to
migrate to India. Even in East Pakistan, 41% of the
minority population was ethnically cleansed.
- In India, by
comparison, the total proportion of Muslims who migrated to
Pakistan or were killed in violence was less than 18% of the
total Muslim population. It is a high proportion but
certainly no comparison with those for Pakistan. The total
number of migrants to Pakistan was 7.23 million by the year 1951
and the number of people unaccounted for is relatively small at
0.43 million.
- An important fact
that is not often acknowledged in western and even Indian
writings on the subject of partition is that at partition,
Pakistan had a higher proportion of non-Muslims than India had
Muslims. Yet, the leadership and people of India
chose a secular path for the country and enacted a constitution
that gave equal rights to all citizens irrespective of
faith. On the other hand, Pakistan, which had a higher
proportion of minorities made it a policy to drive them out of
their ancestral lands and worked towards a constitution that denied
equal rights to non-Muslims.[16]
Today, Pakistan has virtually no religious minorities (i.e.
non-Muslims) left while India has a larger proportion of Muslims
than it had in 1951.
Why were the situations
so different in the two countries?
It has been established by this analysis
that the situations in the two countries were vastly different.
While there was violence on both sides, there is no evidence of
ethnic cleansing of Muslims in India. But there is at least prima
facie evidence for ethnic cleansing of Hindus and Sikhs in
Pakistan, particularly in West Pakistan. While an analysis of
the comprehensive set of factors responsible this difference is
beyond the scope of this article, it would be useful to look at the
vastly different positions of the leadership on the two sides.
Both in words and in actions, the
leadership in India was attempting to put out the flames of hatred
and violence that had spread as a result of partition. On the
other hand, the leadership of the Muslim League in Pakistan was doing
all it could to incite violence and force the minorities out.
This was in line with its ideology whose central premise – the
two nation theory – was that Hindus and Muslims could not live
together. Therefore, it was but a consequence of this theory
that minorities had to be forced out of Pakistan.
Selected quotes from the leadership of
the Muslim League, given in Table 4, demonstrate how this was an
organized campaign with the Government of Pakistan completely on
board.
Table 4: Selected quotes
of Muslim League leaders espousing violence
Leader
|
Quote
|
Mohammed Ali Jinnah
|
“Direct Action
by Muslims will lead to one hundred times more destruction than the
Direct Action of the Hindus.”[17]
|
Raja of Mehmoodabad
|
“It is the dictatorship
of the Koranic laws that we want —and that we will have
— but not through non-violence and Gandhian truth.”[18]
|
Sardar Shaukat Hayat
Khan
|
“The Punjab
Muslims do not believe in non-violence and should not, therefore,
be given cause for grievance because once the Muslim lion is
infuriated it would become difficult to subdue him.”[19]
|
The Muslim League formed a militia force
by the name of Muslim National Guards to implement the plan to kill,
convert and force out the minorities, particularly Hindus and Sikhs.[20]
Nothing was done by the leadership to inspire confidence amongst the
minorities that their lives and properties would be protected in
Pakistan. To the contrary, it was made adequately clear that
they would not have an equal status with the Muslim majority
population. The Objectives Resolution was passed by the
Constituent Assembly of Pakistan on March 12, 1949 and it made amply
clear that Muslims would have a higher status in the country than the
minorities.[21]
The contrast with the situation in India
cannot be starker. The architect of India’s independence,
Mahatma Gandhi, spent the last several months of his life trying to
douse the fires set throughout the subcontinent. In fact, even
as the country was celebrating the dawn of independence on
August 15, 1947, he was far away in Noakhali, walking from village to
village, trying to fight violence through non-violence and appealing
to the innate goodness of people. He succeeded to a great
extent in his efforts. In Calcutta, he single-handedly managed
to douse the fires of violence through a fast unto death. It
must be noted that Calcutta was where the violence began, when the
Muslim League Premier of Bengal, Hussain Shaheed Suhrawardy let loose
an orgy of violence that killed thousands, primarily Hindus.
After his fast unto death succeeded in ending violence there, the
then Governor General of India, Lord Mountbatten called him a one-man
army. In the last few days before his assassination, he managed
to quell violence in Delhi, again through a fast unto death.
Not only Mahatma Gandhi, but the
Government of India and the leadership of the Congress Party were
trying their best to contain the violence in the territory of
India. The Constituent Assembly of India, in sharp contrast to
its counterpart in Pakistan, played a major role in instilling confidence
amongst the minorities. It adopted a constitution that provided
full constitutional rights to all people, whichever community they
may belong to. As a result of this, 82% of the Muslim
population residing in India at that time continued to live in the country.
Only 18% thought it fit to migrate. While not denying that many
Muslims migrated because of the communal riots, the patterns of this
migration (most of those who migrated were professionals) also shows
that a substantial proportion of them went to Pakistan in order to
take advantage of the opportunities the new country offered them in
the Government Services.
In sum, it would not be unreasonable to
conclude that the huge difference in the proportion of people who
were killed or forced to migrate in the two countries was primarily
due to the fact that there was an organized campaign to spread terror
amongst minorities in Pakistan while it was quite the opposite in
India. It is no wonder therefore, that while Pakistan has
virtually no minorities today, India has a larger Muslim population
than Pakistan.
The regional differences between West
Pakistan and East Pakistan is a topic that deserves further
study. However, some potential reasons why the project of
ethnic cleansing did not succeed to the same extent in East Pakistan
as in West Pakistan include the calming influence of Gandhi during
his long stay at Noakhali and Calcutta and a higher minority
proportion to begin with (with geographical pockets where they
constituted a very large proportion of the population). The
latter reason also holds in Sind, where some areas like the
Tharparkar district had large concentrations of minority
populations. Perhaps, this provided them greater protection
than in areas where they were in smaller numbers.
Conclusions
We conducted this analysis with the
objective of finding hard numbers for the extent of ethnic cleansing
in Pakistan during partition. Hopefully, this study would serve
as a source of reference for those who want to know these
facts. It would also hopefully motivate further research in
this area using credible data sources and a rigorous methodology.
This study found strong evidence for
ethnic cleansing in Pakistan and particularly in West Pakistan.
It also compared these numbers with corresponding numbers for India
and found that there is no case for apportioning blame to both
countries. There was no evidence for a corresponding ethnic
cleansing in India. West Pakistan had a higher minority population
(21%) than India (13%) in 1941 but by 1951, Pakistan had virtually no
minorities left (3%) while India continued to be home to a
substantial Muslim minority (9%). This analysis stops at 1951
but a further analysis would reveal that while the Muslim minority in
India has increased over the years (reaching 12% in 1991), it has
continually declined in both West Pakistan and East Pakistan.
This analysis is however, beyond the scope of this article.
The study reveals that about 15 million
people belonging to the minority communities in Pakistan were
ethnically cleansed during partition. These numbers are
substantially higher than existing numbers quoted in most articles
and books. While our estimates are not foolproof, they would
suggest that many of the established numbers might need to be
revisited. It would be important to note however, that these
numbers are in terms of 1951 population – hence accounting for
growth (birth and mortality), a more accurate number could be arrived
at for the entire period (in other words a discounted value could be
arrived at to get a better estimate of the actual numbers ethnically
cleansed).
An interesting, albeit chilling finding
of this analysis is that there are almost 7.5 million people
belonging to the Hindu and Sikh minority populations of Pakistan that
are unaccounted for over and above those who migrated to India.
Some of them would have been killed and others converted from their
faiths to Islam. This fact seems to have received inadequate attention
amongst researchers till now and more research may perhaps reveal
much more than this limited analysis could reveal.
At this stage, it would be in order to
list some of the limitations of the study presented in this
article. First, it is based entirely on census data.
Therefore, while it may have revealed interesting facts, it cannot go
into the underlying causes for these facts. Second, the
extrapolation of the hypothetical populations of minorities in 1951
(if there were no partition) is relatively crude as we had to depend
only on data from 1941 and 1951. A more sophisticated
analysis could have been conducted if more data were available.
Third, our estimates depend on the assumption of equal growth rate of
population for various communities between the two countries.
If we had time-series data available from before 1941 and after 1951,
we could have arrived at more accurate growth rate numbers.
While some of the numbers would change with more accurate growth
rates, our results are likely to remain directionally the same.
Additionally, the fact that the number of migrants to Pakistan as per
the census is almost equal to our estimate gives us confidence on the
robustness of this result. Fourth, some of our conclusions
(e.g. the number killed/converted)are sensitive to the assumptions of
equal growth we have made. These numbers can therefore be taken
as first-cut estimates rather than conclusive facts.
Future work would attempt to analyze some
of the issues raised by this study in more depth. It would also
attempt to conduct a more rigorous analysis using additional data
from censuses starting from 1901. In addition, it would attempt
to look beyond 1951 to analyze continued ethnic cleansing in Pakistan
even after 1951. This is of particular relevance in the context
of East Pakistan and Sind, where the extent of ethnic cleansing till
1951 was somewhat lower than in other provinces of Pakistan.
Appendix I
Raw Data Used in this
Study
Sources: Census of India,
1941, Census of India 1951 and Census of Pakistan 1951
Table A1: Population
Counts for Religious Communities - Pakistan and India : 1941
Province/State
|
Population Count in
1941 (‘000)
|
Muslims
|
Hindus
|
Others
|
Total
|
Baluchistan Districts
|
439
|
45
|
18
|
502
|
Baluchistan States
|
346
|
10
|
|
356
|
Present Day
Balochistan Province
|
785
|
55
|
18
|
858
|
NWFP Settled Districts
|
2789
|
180
|
69
|
3038
|
NWFP States
|
2378
|
|
|
2378
|
Present Day
NWFP Province
|
5167
|
180
|
69
|
5416
|
Punjab
|
11744
|
2082
|
1861
|
15687
|
Bahawalpur
|
1099
|
174
|
68
|
1341
|
Present Day
Punjab Province
|
12843
|
2256
|
1929
|
17028
|
Sind
|
2999
|
1035
|
65
|
4099
|
Khairpur
|
254
|
50
|
2
|
306
|
Karachi Federal Area
|
153
|
186
|
26
|
365
|
Present Day
Sind Province
|
3406
|
1271
|
93
|
4770
|
West Pakistan
|
22201
|
3762
|
2109
|
28072
|
East Bengal
|
29577
|
11918
|
430
|
41925
|
Pakistan
|
51778
|
15680
|
2539
|
69997
|
India
|
42222
|
254320
|
22461
|
319003
|
Undivided India
|
94000
|
270000
|
25000
|
389000
|
Note: The Census of Pakistan 1951 is not
consistent about reporting religious communities other than Hindus
and Muslims. In some cases, they have been aggregated as
‘others’ and in other cases, they have not been
aggregated. In order to maintain comparibility, we have
aggregated individual community into ‘others’ wherever
they have been separately reported. This does not affect our
analysis, which focuses on Muslim and non-Muslim populations in the
two countries.
Table A2: Population
Counts for Religious Communities - Pakistan and India : 1951
Province/State
|
Population Count in
1951 (‘000)
|
Muslims
|
Hindus
|
Others
|
Total
|
Baluchistan Districts
|
594
|
4
|
4
|
602
|
Baluchistan States
|
543
|
9
|
|
552
|
Present Day
Balochistan Province
|
1137
|
13
|
4
|
1154
|
NWFP Settled Districts
|
3217
|
2
|
4
|
3223
|
NWFP States
|
2642
|
|
|
2642
|
Present Day
NWFP Province
|
5859
|
2
|
4
|
5865
|
Punjab
|
13511
|
20
|
401
|
13932
|
Bahawalpur
|
1808
|
13
|
2
|
1823
|
Present Day
Punjab Province
|
15319
|
33
|
403
|
15755
|
Sind
|
4149
|
454
|
3
|
4606
|
Khairpur
|
309
|
10
|
|
319
|
Karachi Federal Area
|
968
|
16
|
23
|
1007
|
Present Day
Sind Province
|
5426
|
480
|
26
|
5932
|
West Pakistan
|
27741
|
528
|
437
|
28706
|
East Bengal
|
32227
|
9239
|
466
|
41932
|
Pakistan
|
59968
|
9767
|
903
|
70638
|
India
|
35000
|
304000
|
22000
|
361000
|
Undivided India
|
94968
|
313767
|
22903
|
431638
|
Endnotes
[1]
A select bibliography of material on the partition of India in 1947
is available at the following link http://www.sscnet.ucla.edu/southasia/History/Independent/partition_bibliography.html
[2]
Two books that have concentrated on this topic include Now It Can
be Told by A.N. Bali (Akashvani Prakashan, Jullundur City; year
of publication unknown) and Muslim League Attack on Sikhs and
Hindus in the Punjab: 1947 by Gurbachan Singh Talib (S.G.P.C.
Amritsar, 1950).
[3] See for instance Pravin M. Visaria (1969),
“Migration between India and Pakistan 1951-61”, Demography,
Vol. 6, No. 3, pages 323-334.
[4] Ahmed, Ishtiaq (2002), “The 1947
partition of India: A Paradigm for pathological politics in India and
Pakistan,” Asian Ethnicity, Vol. 3, No.1, pages 9-28. (http://www.sasnet.lu.se/partition.doc)
[5] Muthuswamy, Moorthy (1999), “Demanding
Land from Pakistan”, available at the following link http://www.saveindia.com/crimepak.htm
[6]
See for instance, The Sole Spokesman: Jinnah, the Muslim
League and the Demand for Pakistan by Ayesha Jalal (Cambridge
University Press, 1985).
[7]
see, for instance http://news.bbc.co.uk/hi/english/static/in_depth/south_asia/2002/india_pakistan/timeline/1947.stm
[8]
NWFP is an acronym for North West Frontier Province
[9]
These numbers are quoted in Visaria (1969), op. cit.
[10]
See Talib op. cit.
[11] see also Urvashi Butalia (1998), The Other
Side of Silence: Voices from the Partition of India, Penguin
India, for some personal accounts relating to conversion of people
after partition.
[12]
See for instance, Butalia (1998), op. cit.
[13]
See for instance Cracking India: A Novel (Ice Candy Man) by
Bapsi Sidhwa (William Heinemann, 1991), on which the film 1947
Earth was based.
[14]
See endnote 7
above for instance or Andrew Whitehead’s “Refugees from
Partition”, in the India Disasters Report (http://www.indiadisasters.org/idrpdf/Refugees/Refugees-Partition%20refugees.PDF).
[15]
Visaria (1969), op. cit.
[16]
The Objectives Resolution of the Pakistan Constituent Assembly set
the parameters for the constitution in explicitly Islamic
terms. The full text of the resolution is available at the
following link: http://www.pakistani.org/pakistan/constitution/annex_objres.html
[17]
Talib (1950), op. cit.
[18]
Ahmed (2002), op. cit.
[19]
Talib (1950), op. cit.
[20]
Talib, ibid.
[21]
See endnote 16
|