Scientific Empire and Imperial Science: Colonialism and Irrigation Technology in the Indus

Basin
Author(s): David Gilmartin
Source: The Journal of Asian Studies, Vol. 53, No. 4 (Nov., 1994), pp. 1127-1149
Published by: Association for Asian Studies
Stable URL: https://www.jstor.org/stable/2059236
Accessed: 25-05-2020 10:42 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Association for Asian Studies is collaborating with JSTOR to digitize, preserve and extend
access to The Journal of Asian Studies

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
 Sclentl'fic Empire
and Imperial Science:
Colonialism and Irrigation
Technology in the Indus Basin

DAVID GILMARTIN

.N THE EYES OF MANY COLONIAL ADMINISTRATORS in the n

advance of science and the advance of colonial rule went hand in hand. Science
helped to secure colonial rule, to justify European domination over other peoples,
and to transform production for an expanding world economy (Adas 1989). The
history of irrigation in India, where the British built large new irrigation works to
increase colonial revenues and expand commercial production, provides a dramatic
illustration of this.
But science played contradictory roles. The imperatives of "imperial science,"
the worldwide application of science to the productive control of nature, were closely
associated with the expansion of colonialism and commercial agriculture. But the
transformative operation of "imperial science" was constrained in the colonial context
by the presence of another powerful scientific "discourse" of power, the "science of
empire," or imperial administration. Indian administration hardly rested entirely
on science. But in the nineteenth century a scientific administrative "discourse"
influenced significantly the forms of control that came to dominate the imperial
Raj-and that shaped Indian irrigation.1
Bernard Cohn, among others, has written extensively on the development of a
"colonial sociology," a structure of knowledge about India, shaped by social science,
that strongly molded structures of British political control. Seeing India as a fabric
of local communities held simultaneously within a structure of intellectual and
administrative control, the British used census surveys and classifications of castes,
tribes, languages, and religions to help to lay the foundations for their power (Cohn
1987). Clive Dewey has discussed the late nineteenth-century development of
evolutionary social theory among Punjab administrators as defining for them "the
dazzling vision of a 'science of government'" (Dewey 1991:28).

David Gilmartin is Professor of History at North Carolina State University.

'The use of the word "discourse" here is influenced by Foucault, though not used in a
rigorous sense. The term is used to suggest that "imperial science" and "scientific empire"
were not the exclusive preserves of competing colonial classes or groups, whether engineers
or civil administrators, but were two coherent, but distinctive, modes of scientific thinking,
which were often in tension within the thinking of the same individuals.

The Journal of Asian Studies 53, no. 4 (November 1994): 1127-1149.

(? 1994 by the Association for Asian Studies, Inc.

1127

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
1128 DAVID GILMARTIN

Most critical to the "science of empire," however, was the fact that it defined
not simply a structure of domination, but also a distinctly colonial political system
linking the colonial state and indigenous elites together in a common political order.
The key to this order-and the "science of empire"-lay in the scientific appropriation
of what might be called "local knowledge." Indigenous "local knowledge" (and the
local power relations and forms of knowing that it embodied) of necessity occupied
a critical place in the scientific structuring of colonial administration. Though
"objectification" lay at the heart of colonial social science, the complete objectification
of the ruled was impossible within the discourse of "scientific empire." The object
of study in India was not simply a "society" of individuals controlled, in the words
of Foucault, through "small techniques of notation, of registration, of constituting
files, or arranging facts in columns and tables" (Foucault 1977:190), but an alien
world of "communities" and "cultures." To know these, local information and local
subjects were critical. Many historians have thus stressed the important interactive
role played by indigenous informants, particularly colonial elites, in defining the
scientific discourse that shaped the basic "colonial sociology" underlying British
colonial power (e.g., Amin 1989, Freitag 1991). Power was hardly equal in this
system, but the "science of empire" linked the state and Indian elites together in
a common political structure.2
"Imperial science," with its command over environmentally transformative
technologies such as irrigation, operated in India within this political framework.
"Imperial science" suggested a colonized world that became, in many respects, a
great laboratory in which the natural world was not only catalogued, studied, and
observed, but also technologically manipulated in the name of commercial
transformations on a great scale. But "imperial science"-however important in
justifying colonial rule for Europeans-did not itself define a common political
discourse in which the colonizer and the colonized were linked. The knowledge that
produced these transformations was a universal, yet bounded field of expertise in
which "local knowledge" (though potentially useful to the scientist) had no formal
place. Critically, "imperial science" (and technology) was thus not, as a distinctive
discourse, grounded in a context-specific political field in which the rulers and the
ruled shared.
Indians were not, of course, excluded wholly from the transformative realm of
"imperial science." In fact, the discourse of physical science and technology in India
was joined, as time went on, by an increasing number of middle-class Indians educated
in new universities and technical schools. But it remained under the British a realm
of limited access. If there was a potential political bond linking "imperial science"
to the indigenous population, it was the promise of improved living conditions that
the control of nature offered. But the "science of empire" itself discouraged such a
political bond, however sympathetic administrators were to increasing production
and improving living standards. "Scientific empire" was predicated not on a society
of maximizing individuals in potential alliance with the state, but on a political
order shaped by state classification and analysis of inescapably alien cultures and
**3
communities.

2For a good discussion of the critical role of agrarian elites within the colonial p
structure, see Yang 1989. As he makes clear, the existence of a common political system
did not suggest an order free of conflict.
3The notion of a political alliance between maximizing water users and state science as
the foundation for a program of irrigation development has been developed by Donald Worster
in discussing the western United States. The "drive to make the bleakest, most sterile desert

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
 SCIENTIFIC EMPIRE AND IMPERIAL SCIENCE 1129

The role of "science" in its various forms, was thus, in the colonial context,
politically problematic. By examining some important aspects of the development
of irrigation, which was a vital key to the expansion of agricultural production in
the subcontinent, this article will analyze the impact of the interlocking, yet
conflicting, discourses of "scientific empire" and "imperial science" on the colonial
politics of technological and agricultural transformation.

Irrigation and the Imperatives

of Scientific Empire

Perhaps no technical innovation had more potentially transformative effects in

colonial India than the extension of irrigation. Numerous writings by early colonial
administrators (and modern experts) attest to the critical importance of water
availability and reliability as determining constraints on Indian agriculture. And
nowhere was this more true than in the vast Indus Basin region in northwestern
India (and what is today Pakistan), which will be the primary focus of this article.
Although we lack a continuous history of Indus Basin agriculture, the region
is one of the oldest areas of agricultural production in the Indian subcontinent. The
annual floods of the Indus and its tributaries have shaped the development of agriculture
in the region for at least four millennia. In an area of generally sparse and unreliable
rainfall, in which pastoralism had long provided the predominant livelihood of much
of the population, technical innovations that increased water availability were critical
in the region's agricultural history. The introduction and diffusion of the Persian
wheel between the thirteenth and sixteenth centuries, for example, allowed animal
power to be harnessed effectively to wells, and this encouraged, as Irfan Habib has
argued, large-scale migration and agricultural settlement in the central Punjab (Habib
1970:149-55; see also Singh 1985). Similarly, cycles of inundation canal construction
strongly influenced localized patterns of agriculture, opening areas away from the
direct action of river floods to regular irrigation when the rivers filled the canals
in the hot spring and summer months.
We know relatively little about historical patterns of canal construction before
the modern period, but what little we do know suggests that canal-building was
not simply a matter of technical innovation but was almost always connected as
well with the political imperatives of rule. This was evident in the expansion of
inundation canal construction-canals requiring relatively little in terms of permanent
headworks-in the period immediately preceding the coming of the British. Though
some Mughal canals had existed earlier, the regional Indus Basin states that sought
to consolidate their local power in the wake of the decline of the Mughal Empire-
the Sikhs, the Nawabs of Bahawalpur, the Amirs of Sind, and others-turned widely
to inundation canal construction in the late eighteenth and early nineteenth centuries
to provide an agricultural base for their control of local and regional elites. These
elites were sometimes warrior kinsmen of the ruler, sometimes pastoral or tribal

produce more and more of everything," Worster writes, provided "an ideology shared
wholeheartedly by agriculturists and water bureaucrats," thus creating a political bond uniting
these "potentially rival centers of power into a formidable alliance" (Worster 1985:52-53).
Though this alliance produced contradictions and serious environmental consequences, it
defined popular political foundations for the operation of an ideology of "imperial science."
"'Scientific empire" provided no room for such an alliance.

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
1130 DAVID GILMARTIN

, ' J The Indus River System

r(and current national boundaries)

NPAKISTAN cW ,I

'.>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~do
INDIA

. 11~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~1
N

0 250 500 750 1 000 km

Map 1. The Indus River system and current national boundaries.

chiefs, sometimes religious figures, and sometimes, as during the early nineteenth-
century rule of Sawan Mal in Multan, monied men of commercial castes whose
support was critical to the state. But construction of canals depended usually on
the ability of the state to mobilize local elites and their followers in canal digging,
while creating new "communities" of sharers in canal water and in the yearly obligations
of canal silt clearance and maintenance that kept canals flowing during the summer
months.4 This allowed not only the localized production of valuable commercial
crops (such as indigo), but also the definition of a structure of power linking the
state and local elites together.

4The system of chher labor on the inundation canals of southwestern Punjab and Sind
illustrated this well. Though the state usually helped to organize canal silt clearance, canal
sharers were responsible for annually clearing the canals of silt (which was critical to their
functioning) in return for rights to take water. Unpaid canal laborers provided by irrigators
were known as chhers. Though increasing control of the environment proved critical to state
policy, it was achieved less by direct state control than by the state's ability to define and
manipulate "communities" of canal sharers, frequently dominated by privileged elites, in
order to mobilize the investment and labor necessary for expanding cultivation.

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
 SCIENTIFIC EMPIRE AND IMPERIAL SCIENCE 1131

Indus Basin Canals z4 I7 IfUBC

) Suliur ~~~~~ Perennial Canal Systems

| .1 Em~~~LC Lower Chenab: 1892 l

l 4 i E3~~~~~U Lower Jelum: 1902l

l ~ ~ ~ ~ ~Based on B.LC.C Johnson, Pakistan (1979)

co 1900 h0o,100 200 300km

Map 2. Indus Basin canals, c. 1900.

The coming of British rule in some ways transformed these patterns of canal
construction, as the British brought new technologies and new conceptions of state
power to the Indus Basin region. In the years after the British fastened their control
on Sind and the Punjab through military conquest in the 1830s and 1840s, British
irrigation policy generally in India stressed the importance of returns to state revenue
and protection against famine as .guiding considerations for irrigation development
(Whitcombe 1982; Stone 1984). But colonial canal-building in the Indus Basin in
the second half of the nineteenth century also took on a distinctive political and
ideological significance. Though the British sought initially to expand and maintain
existing inundation canal networks in the region, colonial officials had begun to
call as early as 1849, the year of Punjab's annexation, for "science" to "fertilize the

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
1132 DAVID GILMARTIN

bar."5 Subsequently, an emerging class of professional irrigation

construct the works of "bold and magnificent conception" (in the words of the
Punjab Irrigation Manual) that were to transform the Indus landscape. Beginning
in the 1880s, the British built a series of perennial, all-season canals, their flow
controlled by permanent weirs built across the region's rivers, that opened millions
of acres of arid land for new agricultural settlement. Few undertakings so caught
the imperial imagination and expressed the politically legitimizing power of
technological transformation as "the grand edifice" of these canals (Public Works
Department, Irrigation Branch, Punjab 1943:section 3.50). Facilitating the production
of large agricultural surpluses (particularly wheat) for export, these canals offered
not just a return to state revenues; they were triumphs of "imperial science."
But the building of these canals related also to the political imperatives of state-
building in the Indus Basin region. For the British, as much as for earlier Indus
Basin states, the link between canal building, agricultural settlement, and political
control was central to the construction of state power. This had been evident in the
earliest days after British military conquest of the region. The first major canal
project undertaken after Punjab's annexation was the (Upper) Bari Doab Canal (UBDC)
project in the old Sikh heartland of Gurdaspur, Amritsar, and Lahore Districts.
Though conceived as the expansion of an older Mughal canal, the work was pressed
forward because, in the words of a later engineer, "the immediate construction of
the canal was regarded as almost a matter of political necessity to provide employment
for the disbanded Sikh soldiers, who, having their homes in the centre of this tract,
would otherwise have had little encouragement to turn to agriculture" (Frost 1904).
The political imperative of agricultural settlement was all the more compelling
in the subsequent decades, as the British sought to stabilize their authority in what
was essentially a frontier region. In Sind, for example, the British viewed irrigation
policy as critical to controlling the upper Sind frontier, where, since the days of
John Jacob in the 1840s, canal-building was seen as a "civilizing" lever "to induce
the roving predatory Baluch tribes . . . to take to peaceful agricultural pursuits"
(Shoubridge 1898, see also, Lambrick 1975:268-70). Similar concerns motivated
the early British encouragement of Baluch canal investment along the Dera Ghazi
Khan frontier further north. With agricultural settlement a key to British policy,
"the good faith of the Baloch chiefs," in the words of Sir Robert Sandeman, came
largely to depend on local British-supported canal construction and administration
(Sandeman 1870). Nor was this link between canals and political control confined
to the trans-Indus frontier. In arid western Punjab particularly, the British sought
to establish their power also along a series of "internal frontiers," where pastoral
communities were prominent. Settlement of pastoralists, in fact, emerged as a major
element in the establishment of British power in the Punjab, particularly after the
1857 Mutiny, when John Lawrence sought to define a stable peasantry as the social
bedrock for the colonial state. As one local officer wrote of parts of the Multan bar
in 1849, "the people . . . are predatory herdsmen, little engaged in agriculture,
and without extensive means of irrigation. . . . To give them the means of cultivating
would be the most efficient aid to the Magistrate" (James 1849). British policy,
in fact, encouraged canal investment by pastoral chiefs and other notables (often
through large land grants) as a vital element in political stability throughout the
area (e.g., Perkins 1882). Such men were viewed as critical intermediaries who

5The bar was the high land in the doabs between Punjab's rivers, outside the action of
river floods (James 1849).

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
 SCIENTIFIC EMPIRE AND IMPERIAL SCIENCE 1133

could order relations between the state and a developing social base of settled peasant
communities.
The extension of irrigation thus had critical political implications for the British.
Agricultural settlement played a constitutive role in the structuring of the emerging
colonial commercial and political order in the Indus Basin. But it also helped to
define the contours of a "scientific" administration. On one level, agricultural
settlement was important to the British not only for its role in the establishment
of order, but perhaps even more critically, for its role in encouraging a general
"moral" transformation that would draw Indians into the developing colonial legal
structure. "Rude races first learn civilization by becoming possessed of property,"
wrote Richard Temple in the 1850s. "Take a wild wanderer of the Bar, give him
some land to squat upon and call his own, and he forthwith becomes a wiser and
better man" (Temple 1852). British state power was, in fact, tied to a legal structure
in which the definition of property rights in land was central. But agricultural
settlement was important in British eyes not only for fixing Indians in a world of
private property, but also for locating them within a scientific political discourse.
Critical for the structure of British rule was the fact that by tying Indians to the
land (and by legally defining their relationship to the land through the institution
of private property), the British tied them also into a social scientific discourse in
which the language of "community" and "custom" defined a local social order associated
with the colonial state. This discourse lay at the heart of "scientific empire."
The constitutive administrative act for the establishment of this administrative
order was the British revenue settlement. The periodic settlement and revision of
the land revenue demand was, as many historians have noted, at the center of colonial
finances and administration. But the revenue settlement created also a nexus of
community and land that linked local society to the colonial state. At the heart of
every land revenue settlement was an effort to define the position of each individual
both with respect to a local, social "community" and to the land. British officials
thus recorded carefully the genealogies ("pedigrees," as they were sometimes called)
and local "customs" of the settled agriculturists of the Punjab, which were usually
recorded in the village papers at settlement (most importantly in the village "record
of rights"). To record the genealogical tables was, in the words of one settlement
officer, to know the history of a village (Saunders 1873:75-76). Customary "shares"
in village assets (such as commons) and individual property "rights" in land (and
water) were further ordered by a large (and speculative) literature on the history
and social evolution of local forms of "community" organization in India that both
drew on and, at the same time, ordered the mass of "local knowledge" collected
in British Settlement and Assessment Reports (Dewey 1991:22-30). To be "settled"
on the land was thus to be "settled" not only in a physical sense, but also within
a framework of local relationships ("pedigrees" and "customs") shaped simultaneously
by an indigenous language of kinship and "custom" and by the legal and administrative
language of the colonial state.
Powerful local men were also vitally implicated in this process. Like their
predecessors, the British used the process of settlement to tie powerful local elites
to the state. But this involved more than just grants of land. The creation of the
revenue system drew local elites into a "scientific" discourse of "privilege," couched
(usually) in the local idioms of customary authority, genealogy, and descent.
"Privileged" descent (whether based on religious ancestry, descent from a village
founder or "tribal" chief, or membership in a dominant biradari) was a vital ordering
force in local culture (even if it served, at times, as an idiom that "normalized"

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
1134 DAVID GILMARTIN

other forms of local power relations). Its significance under colonial rule lay in the
fact that, within the social scientific discourse of imperial power, it linked the
struggle over local access to resources (status, land, and water) to the discourse of
"community" and power that bound settled "communities" to the legal and
administrative structure of the colonial state. It legitimized local power holders in
the eyes of the state, even as it gave them leverage, as the chief purveyors of "local
knowledge," within the state structure. The fact that British administrators and
local notables sometimes attached radically different cultural meanings to "custom"
and "descent" in no way changed this basic fact.
In "settling" the countryside of the Indus Basin into a social science discourse,
the British thus laid the foundations for a structure of "scientific empire" in which
attachment to the land and local "community" played a critical part. It is no surprise,
therefore, that in the Indus Basin the history of "scientific empire" and the history
of irrigation came to be closely intertwined. On one level, of course, this was because
of the close connection in this arid region between the local control of water and
settlement on the land. But on another level, it reflected the fact that control of
irrigation was a hinge between the power of the local "community" and that of the
state.
Nowhere was this clearer than in the colonial debates about the nature of the
labor systems used to maintain canals. Before the advent of perennial, all-season
canals, virtually all inundation canals required annual, large-scale mobilization of
labor for silt clearance during the dry winter months. Systems of labor mobilization
at the time of annexation varied, but on many canals silt clearance depended on
unpaid chher labor provided by the irrigators during the winter clearance season.
Some British officials attacked this inherited system as a form of coercive or "statute"
labor that was incompatible with the guarantees of private property that structured
the British colonial state, and it was for this reason that the system was abolished
in Sind in the 185Os (Jacob 1855). But others stressed the importance of chher labor
not as a form of state coercion, but as evidence of the importance of "custom" and
"community" in canal organization. As J. B. Lyall (later Punjab Lieutenant-Governor)
put it in 1882: The system of chher labor should be supported, "not only on grounds
of economy of management, but also on the ground that it tends to preserve and
promote self-government. The system is solidly founded on custom, and suits the
habits and circumstances of the people concerned" (Lyall 1882).
In fact, debates about chher continued as the Indus Basin canal system expanded
in the late nineteenth century, with some arguing that it hampered effective state
management of canals. The very fact that the system was grounded in "custom,"
some argued, tended to legitimize the exercise of local "privilege" in the management
of irrigation water, even though this often undercut effective water management.
"That the system, at least as worked under the British administration, leads to
oppression on the part of the headmen of villages, and that it is unsatisfactory to
the Engineers, whose object is to maintain the canals in an efficient state," one
leading engineer wrote, "are facts not denied by anyone who has had experience in
the matter" (Crofton 1875). But for others, reliance on (and manipulation of) local
irrigation "customs" fit into the pattern of "scientific empire" through which the
British managed local affairs generally in the Indus Basin-not by rejecting a local
discourse of "custom" and "privilege," but by incorporating it into a larger social
scientific and administrative discourse.6 Though the chher system was finally abolished

6See, for example, the British self-perception of their mission in the words of P. J.

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
 SCIENTIFIC EMPIRE AND IMPERIAL SCIENCE 1135

in Punjab at the turn of the century, in many districts the British continued to
record customary shares in irrigation-like other village customs-in the village
"record of rights." Perhaps more striking, in some districts in which canals were
particularly important politically, such as Dera Ghazi Khan and Peshawar Districts,
the British went further and maintained special registers of irrigation "customs"
(haquq-i abpashi = "rights of irrigation"), which defined the broader relations of
"communities" and powerful local men within canal networks. Control of irrigation,
in these circumstances, was powerfully integrated into the structure of knowledge
and authority by which the colonial state ruled.

The Challenge of Imperial Science

But expanding British irrigation works nevertheless created tension as the

imperatives of "scientific" imperial administration met the countervailing discourse
of technological transformation. Increasingly, in the late nineteenth century, irrigation
became a subject not just for settlement officers and public works officials, but for
the professional engineers who increasingly staffed the provincial Irrigation
Departments.7 Provincial irrigation engineers emphasized new concepts of irrigating
"efficiency" in ways that pushed them toward projects of larger and larger scale,
capable of integrating local irrigation works into larger systems. New technologies
of perennial canal construction allowed for the opening of large areas of arid "wasteland"
for permanent agricultural settlement, thus facilitating the great expansion of the
Indus Basin's population and the expansion as well of winter (rabi) crops, such as
wheat, that became by the end of the nineteenth century commercial export staples.
Large irrigation networks produced the new "canal colonies" of Punjab and Sind,
in which large tracts were opened not only to settling pastoralists, but also to
immigrant agricultural colonists transferred from the "congested" districts of central
Punjab (Ali 1988:8-61). Driven by a growing preoccupation with water-use efficiency
(and with the social engineering required to achieve this), late nineteenth-century
engineers thus sought to assimilate irrigation works as fully as feasible-and including,
where possible, the old inundation canals-into the increasingly large, technically
integrated networks of water delivery that could support such transformations.
No absolute line can be drawn, of course, between the attitudes of these new
engineers and those of the revenue administrators who shaped the discourse of "scientific
empire" in colonial India. The increasing engineering concern with efficiency and
with the technical integration of larger and larger canal networks led in practice to
the bureaucratization of canal administration in ways that paralleled the
bureaucratization of the civil administration. The Canal and Drainage Act (VIII of
1873) provided the basic structure of canal administration in northern India. Irrigation
engineers were ranked in provincial hierarchies of authority from the Chief Engineer
down through Superintending Engineers to the Executive Engineers who controlled
circles of canal sections. At the hinge of local canal administration, Executive Engineers
concerned themselves not just with problems of technical design and overall efficiency

Fagan, a high Punjab official: The administration of inundation canals under the British
was "a very signal instance," he said, "of that great central and fundamental problem of
Indian administration, the adaptation of indigenous forms to modern requirements" (Fagan
in Duthy 1919:48).
70n professional engineering education in India, see Headrick 1988:316-20.

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
1136 DAVID GILMARTIN

in canal operation, but also with such matters as managing the large numbers of
indigenous canal subordinates involved in assessing canal rates, settling local differences
over distribution, and noting unauthorized irrigation and violations in canal rules
(Nasir 1988:226-31). As cogs in an increasingly integrated system, they operated
in a structure of bureaucratic control little different from most other colonial officials.
But in certain respects, the outlooks of these engineers and those of revenue
administrators tended to diverge significantly, particularly with respect to the meaning
of administrative and scientific "expertise." Central to the meaning of "expertise"
for most colonial administrators was the collection and processing of "local knowledge"
within the framework of "scientific empire." For engineers, however, "expertise"
rested also on full participation in a universal, increasingly mathematical discourse
of "irrigation science," shared with engineers in Egypt, France, Australia, the United
States, and the rest of the world. Viewed in mathematical terms, the hydraulics of
irrigation channels and the mechanics of dam construction were the same whether
applied in California or the Indus Basin. From this perspective, "local knowledge"
counted for little.
Debate over the place of "local knowledge" in colonial irrigation development
thus suggested the increasing divergences in outlook among state officials concerned
with irrigation systems. To many colonial officials, the science of controlling irrigation
water, like much else in the colonial world, preeminently required "local knowledge"
to be effective. Many administrators resented the engineering emphasis on mathematical
calculation without reference to the "local views" of the people (which were best
interpreted, of course, by administrators with long "local" experience). "The natives
of these districts are excellent practical water engineers," one British official declared
(Merk 1908). They "have known every inch of ground and every turn of the river
and canal since childhood," wrote another (Gladstone 1879). But many professional
engineers rejected these arguments as self-serving and inappropriate. "The Deputy
Commissioner thinks that the zamindars [landownersl know more about some things
than 'we European theorists do,'" replied one engineer. But engineering science
was in reality "a blending of theory and practice" (Bellasis 1907). Certainly, an
awareness of local circumstances was critical to the effective application of universal
irrigation principles. But this did not mean relying on (or "scientifically" recording
and interpreting) the local "prejudices" of the people. Without careful attention to
hydraulic principles and mathematical measurements of slope, attention to local
irrigating conditions would, by itself, mean nothing.
In fact, the increasing concern with technical "efficiency" meant that the roles
of local cultivators in the effective "use" of irrigation water were matters of considerable
concern to engineers. Local community, as even many engineers realized, could be
critical to the successful application of technology, for it provided a form of organization
to structure the local cooperation often necessary to make "efficient" irrigation systems
work. But for engineers, "local community" held very different meanings from those
it held for the colonial rulers. For most colonial administrators, "local communities"
were shaped by the discourse of "custom" and "genealogy" that supported the political
structure of the state. But for engineers, effective local cooperation was shaped not
by "custom" and "privilege," not by the social scientific categorization of local
people, but by the organization of people for productive "efficiency." While strongly
identifying themselves with the power of the imperial state, Indian engineers thus
operated, ironically, within a discourse that tended in critical respects to challenge
the political foundations on which the power of the "scientific" colonial state in
India had been built.

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
 SCIENTIFIC EMPIRE AND IMPERIAL SCIENCE 1137

Conflict between imperial notions of irrigation "customs" and "rights," on the

one hand, and the technical imperatives of canal development under engineering
auspices on the other, thus marked irrigation development as the roles of professional
engineers expanded-both on the new perennial canals and on older inundation
canals as well. One example illustrates this, the new engineering technology of silt
control. Control of silt had long been critical for the operation of irrigation in the
Indus Basin and was vital to effective irrigation. But silt problems took on special
significance in the late nineteenth century because they were closely connected to
the distribution of water within irrigation networks and to the place of local
"communities"-and of "custom" and "privilege" -in the structure of canal operation.
Engineering control of silt drew engineers directly into the scientific discourse of
local "community. "
As one of the heavier silt-bearing river systems in the world, the Indus system
was one whose irrigation development had long had to contend with pervasive problems
of annual silt deposits. Heavy silting by rivers at flood stage made the annual silt
clearance of inundation canals an essential operation if canals were to flow from year
to year. It was, in fact, the problem of silt clearance that had given shape to many
of the predominant community institutions of local canal management in the pre-
British period, including the system of chher labor, which had traditionally linked
canal clearance with shares in water distribution. While many engineers were
uncomfortable with and critical of this system from an early date, they had considerab
difficulties in devising effective alternate means for dealing with silt clearance. The
early abolition of "statute" labor in Sind, for example, produced a financial crisis
when canal officials discovered in 1870 that about a quarter of all Sind revenue was
going to pay for yearly canal clearances by hired labor and that, even then, inadequate
clearances were responsible for continuing canal deterioration (Merewether 1870;
Fife 1871). Control of silt was thus a major early concern to canal engineers, and
it was no accident that breakthroughs in silt management represented one of the
major contributions of Indus Basin engineering in this period to the development
of international irrigation science as a whole.
The achievement of British Indian engineers in the late nineteenth and early
twentieth centuries was not so much to eliminate silt within the canal system (which
would await the mid-twentieth-century construction of large storage dams), but
rather to define a new, "scientific" way of looking at the problem that would justify
excluding local "communities" from a role in the administration of major canal
channels. The major scientific breakthrough in silt control came with the work of
R. G. Kennedy, later Chief Engineer of the Punjab, on the (Upper) Bari Doab
Canal in the 1880s and 1890s. Kennedy was the first to derive a reliable scientific
formula for the slope and water velocity that could be maintained in an unlined
canal channel without causing silting or scouring. This, in effect, allowed engineers
to define mathematically what were known as "regime channels," or "canals in
which silting balances scouring during continued periods of operation" (Houk
1956:100-2; see also Sharma 1948:157-85). In fact, the mathematical definition
of "regime channels" by no means guaranteed that the optimal conditions for such
a channel could be fully reached or permanently maintained. But the definition of
the concept suggested that large-scale annual silt clearances should be viewed not
as a normal part of canal operation, but as evidence of flaws in canal design. Annual
silt clearances, in such a view, only perpetuated the problems that caused heavy

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
1138 DAVID GILMARTIN

silting. As one engineer put it in 1916: "Mr. Kennedy's work was to explain
scientifically that silt clearances were a hopeless task."8
A new "scientific" view of canal operation thus transformed the role of indigenous
labor in canal maintenance. Increasingly, the dominant engineering view in the
early twentieth century came to be that gradual changes in canals should be met
by periodic remodeling that adapted outlets to the developing "regime" of the channel.
Regular mobilization of labor for clearance was replaced by constant scientific
monitoring, management, and remodeling. As one engineer commented: "An
irrigation system in its parts comprises a very delicate machine, and these several
parts constantly require adjustment and overhauling; to deprive the machine of these
adjustments can only spell immediate loss of efficiency and in a very short time
disaster" (Gee 1914). In the case of silt control, therefore, labor mobilized by
"communities" could be replaced, to a very large extent, by more sophisticated
techniques of scientific adaptation to a changing environment. The management of
"community" (through the use of social science) was replaced by the metaphor of
a machine.
This development had critical political implications because it was closely associated
with engineering efforts to transform also the extremely sensitive political question
of the distribution of water-an area in which notions of local "community" had
long played critical roles. With the passage of the Canal and Drainage Act of 1873,
the limitation of the number of outlets from each channel and the careful fixing of
the area to be irrigated from each had already begun to emerge in the name of
efficiency as central engineering concerns.9 But control over outlets, like silt control,
had critical implications for the role of local "communities" of irrigators within the
structure of the irrigating system. Just as the new science of silt control dictated
the freeing of irrigation management from reliance on "community" mobilization
of labor on main channels and distributaries, increasing departmental control over
outlets suggested the concomitant need to restrict local irrigating communities to
the irrigating units defined by each outlet. "Communities," in such a view, needed
to be removed from a world defined by "genealogy," "custom," and "privilege,"
which bore no necessary relationship to the structure of irrigation networks, and to
be encapsulated instead within the technical world of irrigation channels-and within
a scientific discourse of "efficient" control over nature.
Evidence of this concern to remake local "community" within a technical irrigating
system emerged from British efforts to develop and expand the operation of an Indus
Basin water distribution system known as warabandi, based on the fixing of timed
turns for water (waris). Though it was not a new system, British irrigation engineers
sought to adapt the warabandi system, as a form of cooperative local water distribution,
to "fit" into the scientific structure of a larger, developing irrigation network. Timed
water turns had long existed on jointly owned wells (and on some small canals) in
much of the Indus Basin region (Elphinstone 1860:43; Prinsep 1865:104; Steedman
1882:77). Co-sharers in the building of a new well often divided the water available
according to 3-hour (or sometimes 6-hour) pahrs, which were distributed among
the users of the water on a rotational schedule. Though sharers in wells were sometimes

8"For 30 years," he continued, "the Irrigation Department has given up silt clearances
in distributaries" (Ward 1917).
9This was connected as well with efforts to design gauges or modules capable of controlling
and measuring water at each outlet. R. G. Kennedy himself devised a gauge for this purpose;
nevertheless problems with the design of modules that could withstand fluctuations in canal
levels and the tampering of irrigators hampered engineers for decades.

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
 SCIENTIFIC EMPIRE AND IMPERIAL SCIENCE 1139

connected by kinship ties, this form of distribution was one whose essential features
were shaped not by genealogy but by common access to a single, controlled, and
self-contained source of water. For engineers, the significance of the system lay in
the fact that it offered a clear, quantifiable system of water distribution that could
be adapted to each outlet. Its emphasis on the controlling importance of measured
time perhaps might be seen also in Foucaultian terms as an attempt to impose
individual "discipline" within an irrigating "machine." But far more important in
the discourse of engineering science was the fact that it offered a system of local
cooperation that could be effectively encapsulated, not within a discourse of social
science, but within a larger, technically integrated irrigating system. ?
Though the spread of this system occurred only gradually, control of channels,
fixation of outlets, and the encapsulation of such local "communities" of distribution
thus tended to go hand in hand. And they marked an ongoing program of scientific
transformation that affected not only the new, perennial canals of the region, but
also the old inundation canals of the Punjab. With the abolition of the chher system
on many Punjab inundation canals at the turn of the century, the Irrigation Department
launched a broad series of reforms focusing largely on the amalgamation and remodeling
of channels and on reductions in numbers of watercourses and outlets-a process
known as chakbandi (or the definition of fixed areas, or chaks, assigned to each outlet)-
all intended to give engineers greater control, not only over the "management" of
silt in channels, but also over the distribution of water.
Not surprisingly, these reforms produced a wave of protests. Protests in 1907
in the new canal colonies linked opposition to new Irrigation Department rules with
other, broader political grievances (Barrier 1967), but in the older inundation canal
areas, petitions of irrigators focused directly on the implications of the Irrigation
Department's new chakbandis. When the Chief Engineer of Punjab toured Multan,
Muzaffargarh, and Dera Ghazi Khan Districts in 1909, he was deluged with petitions
focusing on loss of access to irrigation water contingent on new chakbandi arrangements.
Irrigators objected to being forced to take water from new channels, to being
amalgamated onto watercourses with other irrigators (with whom they often had
no previous connections), and to being denied the customary right to control their
own supply by putting stop-dams into irrigation channels (Gordon 1909). Many
new irrigation procedures were in fact distasteful to zamindars for precisely the same
reason that irrigation engineers favored them-because they attempted to assimilate
irrigation management to a single system controlled by engineers. As the Multan
Deputy Commissioner wrote: the zamindar "objects to be linked up on a large
system as under this he is entirely at the mercy of the department officials, he can
do nothing to supplement a bad supply, nor has he information in time to adjust
his cultivation to the supply of water available" (Clarke 1908).
But such objections also reflected the general attack in Irrigation Department
reforms on the whole structure of power and "privilege" that had long shaped popular

l'Evidence on the spread of warabandis is somewhat fragmentary but suggests that the
imposition of individual "discipline" on irrigators was not at the heart of engineering concerns.
The legal foundations for the enforcement of warabandis were laid by the 1873 Canal Act.
But warabandis were not rigidly enforced in Punjab on Irrigation Department canals in the
late 19th and early 20th centuries, largely because of administrative difficulties in doing
so. Generally, the Department only intervened to record and enforce a warabandi on an
outlet if it was directly appealed to by the irrigators, and instead allowed irrigators to make
their own distribution arrangements (kachha warabandis), confined within each outlet. It was
estimated that by 1939, less than half the outlets on the perennial canals of the Punjab had
departmentally framed warabandis (Public Works Department, Irrigation, Punjab 1940).

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
1140 DAVID GILMARTIN

organization within the British administration and the irrigation system. Central
to many of the petitions received by the government were complaints about the
abolition of rights of canal sarpanches, "leading irrigators" and headmen, who had
served as "representatives" of canal sections within -the administration of the chher
system. In the eyes of most irrigation engineers, the abolition of the special privileges
of these zamindars was an obvious concomitant of the abolition of the chher system
itself. But as H. J. Maynard, the Multan Commissioner, wrote, canal sarpanches
had long done more than simply supervise the collection of chhers; they had assisted
canal officers "on all matters regarding the management of canals, including matters
of clearance and distribution of water," and on questions of controlling canal
department subordinates. Government recognition of privileged irrigation rights for
such men-and of the validity of their knowledge-had, in fact, long been central
to the maintenance of government's political influence in the management of canals
(Maynard 1907).
Nevertheless, for many engineers the necessity of undercutting the position of
these men was at the very heart of "scientific" management. Writing in 1909, E.
S. Bellasis, the Superintending Engineer of the Derajat Circle, attacked the old
system, linking the language of "equity" and "efficiency." The root cause of "popular"
complaint at irrigation reforms, he said, lay in the power to control water that "big
men" had previously exercised all along the inundation canals of southwestern Punjab.
Certainly such men complained of chakbandi arrangements, of the amalgamation of
canals and watercourses, and of the fixing of permanent, departmentally controlled
outlets, he wrote, but this was primarily because the new departmental arrangements
now limited their power. Hitherto, the large owner had often had "control of his
own and his neighbour's water," Bellasis wrote. "Now things are changed." The
same thing applied to upper irrigators on many channels. Those who protested
against such measures, in fact, failed to understand the meaning of efficiency in
water management and distribution. And perhaps most serious in the eyes of Bellasis
and many other engineers, this included many British civil officials. "It may not
occur at once to the unprofessional mind," Bellasis wrote, castigating those who
viewed irrigation within the traditional structure of administrative discourse, "that
any privilege conceded on a canal to any individual is at the expense of some one
else" (Bellasis 1909).
But the political ramifications of such a "professional" view of society were still
to be reckoned with. If "privilege" stood in the way of professionalism and efficiency
in water control, it had long served far different purposes within the context of th
colonial structure of "scientific empire." Grounded in a discourse of classification
and measurement on the land, state recognition of "privilege" (legitimized often in
genealogical terms) lay at the heart of the system of rural administration and authority
in rural Punjab that linked the colonial state to the particularistic language of
indigenous social order. Though many civil officials were sympathetic to demands
to make water management more "efficient," "unprofessional" thinking (in the sense
that Bellasis attacked it) was in a sense at the heart of British power. In stressing
the need for a water management system that removed all "privilege" from its operation
(and as a result all levers of "community" manipulation), engineers were thus seeking
to define implicitly a new, instrumental (and universal) foundation for the state's
authority. The power of scientific efficiency, linking the individual directly to the
state, was to replace the mediation of the local "community," bound to the state
through the recognition of local "privilege." But the political foundations for such
an administrative revolution were unclear.

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
 SCIENTIFIC EMPIRE AND IMPERIAL SCIENCE 1141

The political issues involved in this debate crystallized most

conflict in the early twentieth century over the meaning of irrigation "rights." In
spite of long uneasiness among some officials with the operation of the chher system,
many civil officials came to see the critical issue in Irrigation Department chakbandi
and remodeling policies as the protection of longstanding "rights," which lay at
the heart of British rule's foundations. The Multan Commissioner thus put the issue
succinctly in 1908:

From the point of view of a professional commercial department the cardinal maxim
is to make the best use of the volume of water available. [But) it is quite another
thing to introduce [rules) suddenly, arbitrarily, without compensation and without
regard to established rights and interests that have been possessed and enjoyed for
generations. ... The people have rights to which they are entitled under the law,
if they are entitled to anything .
(Merk 1908)

Recognition of established "rights" was in this view central to the popular legitimacy
of British political power. But the colonial usage of the term "rights" (or "vested
rights"), whether referring to irrigation or other matters, was rooted in a state-
dominated social science discourse. Legal "rights" in this case were neither universal
natural rights, inhering in the individual, nor "property rights" in the strict sense
of the term-but rather local (if often personal) "rights" defined by state recognition
of particularistic usage and customary "community." Though no one doubted by
the end of the nineteenth century that the government in some sense "owned" the
water that canals delivered (and that the state bureaucracy had played a vital role
in recording and creating the local "facts" of indigenous "custom"), the legal
recognition of these "rights" nevertheless illustrated the state's acceptance of an
indigenous voice (and of "local knowledge") within the discourse of state power.
And for that reason, the question as to how "rights" in water were defined, and
reconciled with engineering "efficiency," was one intimately connected to the principles
linking the people to the state and organizing the political system.
In fact, the developing debate over the meaning of "rights" can be seen more
clearly if we shift the setting from the inundation canals of southwestern Punjab
to the Upper Bari Doab Canal (UBDC) in central Punjab. Much of the conflict can
be traced in the peculiar odyssey of the word "haq," or "right," within the technical
debate over expanding irrigation. The word "haq" was one used early on by canal
engineers on the UBDC to signify the percentage of the "commanded" culturable
land on an outlet that the Irrigation Department agreed to irrigate. But the calculation
and meaning of "haq" became a subject of increasing controversy in the late nineteenth
century as engineers attempted to introduce new scientific principles of irrigation
management. Kennedy's work on silt control required that engineers periodically
remodel most irrigation channels in order to maintain their efficiency in the face
of silting problems. But remodeling normally required the periodic reduction in
size of some outlets (particularly near canal heads) and the consequent modification
of the "haq" (or irrigation "rights") of some villages to secure more efficient distribution
along the canal. That this produced protest was hardly surprising. Few irrigators,
as engineers saw it, really understood "scientific" management. "The point that is
so difficult for the man who has not made a speciality of irrigation engineering to
understand," wrote one engineer, "is the constantly changing conditions with which
we have to contend," and the concomitant need for constant technical adaptation
to keep the irrigating "machine" in order (Gee 1914).

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
1142 DAVID GILMARTIN

Far more serious, however, were the protests that came from British civil officials.
The problem was not just that reduction of outlet size sometimes prompted serious
political protests from irrigators, but that the alteration of the "haq," simply for
reasons of equitable distribution or efficiency, violated certain assumptions that for
many officials defined the basic discourse of colonial authority. Indeed, for most
civil administrators, the word "haq," however technically employed by irrigation
engineers, carried meanings of far wider political importance. In its broader use,
the term suggested "rights" sanctioned by long historical usage or by customary
practice; the word's meaning, in fact, shaded imperceptibly for the British
administration from "rights" to "customs," and it was used clearly in this sense in
the registers of irrigation customs (or "rights"), the haquq-i abpashi (haquq=pl. of
haq) maintained by the British in some western Punjabi districts. For many officials,
this overlap between "rights" and "customs" was not coincidental, but suggested
the manner in which the colonial regime relied on the idiom of customary authority,
tied often to local "community" (and to the recognition of "privilege" within this
idiom), as a fundamental political legitimizer of its rule. The recognition of customary
"rights" within the legal system was, in a sense, the legal counterpart (and legitimizer)
of the recognition of political "privilege" and local "community" within the discourse
of rural administration.
Indeed, many irrigators themselves turned to this idiom to justify the maintenance
of their "rights" to water during remodeling, asserting their power within the discourse
of which this was a part: "[From: more than seventy years ago, we are using this
water and it has become our right now," wrote one Lahore district landowner faced
with canal remodeling, "and it would be a great injustice if we are deprived of this
right" (Singh 1927). The claims of "right" based on an idiom of genealogy were
asserted perhaps most blatantly by Nawab Nisar Ali Khan Qazilbash, a prominent
Shi'a landowner and political magnate of Lahore, in a petition dealing with Irrigation
Department plans to reduce the size of an outlet irrigating his lands on the Niaz
Beg distributary outside the city. When informed by the canal engineer that his
outlet would be reduced in accord with "mathematical calculations" during canal
remodeling, he appealed to the memory of his ancestors, who had received earlier
British land grants on the basis of their descent and service to the state. They, he
emphasized, "did not render service to the British Government after mathematical
calculations" (Qazilbash 193 1). The importance of a discourse in defining the power
of these local leaders was clear, and many British civil officials responded readily
to their appeals, seeing "scientific" canal remodeling in such cases as an attack on
the political foundations of the state's connections with the "people."
Conflict over the Irrigation Department's interference with "vested rights," in
fact, proved to be pervasive. It led to the formulation in 1901 of what were known
as "Haq Rules" for remodeling canals, which were intended as a compromise formula
to allow remodeling to go forward. But the working of these rules-and subsequent
attempts to modify them-simply provided fuel for an ongoing conflict. "Until
the vested right and political expediency arguments have been degraded from their
false pedestal to their proper level," a high-ranking engineer wrote in 1916, equity,
progress and scientific management of irrigation would be rendered impossible (Holmes
1916). But as H. D. Craik, later Governor of Punjab, responded, political "objections
to interference with vested rights" outweighed even the advantages of "a more equitable
distribution of water" (Craik 1916). The underlying ideological assumptions that
shaped the "colonial" view on water "rights" were expressed perhaps most cogently
by the Lieutenant-Governor at the time, Sir Michael O'Dwyer. Principles of abstract

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
 SCIENTIFIC EMPIRE AND IMPERIAL SCIENCE 1143

equity in water distribution meant nothing to O'Dwyer. In his view, only irrigation
"rights" originating in past customary practice warranted recognition by the
government, whose legal authority derived from its protection and arbitration of
these "rights." His solution to the problem of remodeling was thus to encourage
"representatives of villages" with competing claims to put forward their positions
so that "Government officers (Revenue and Irrigation) will then be in the position
not of asserting any new Government rights but of arbitrating and adjusting between
the rival claims of up and down stream villages" (O'Dwyer 1898). Needless to say,
this view, which suggested a need for the state to balance and interpret local
"community," local "custom," and local "knowledge" in the operation of canal
channels, was not one that appealed to the Irrigation Department. But it reflected
a powerful British principle of administration.
Indeed, beyond an appeal to efficiency and productivity, engineers had few
effective political answers to such arguments. For some, the easiest solution was
simply to eliminate the official use of the word "haq" altogether. Since the 1873
Canal Act had, in fact, given the Irrigation Department a legal mandate to run the
canals in accord with principles of efficiency, some engineers blamed the word "haq"
itself for many of their problems. As the Chief Engineer in 1910 (the appropriately
initialed W. E. T. Bennett) noted, the word was the cause of "serious misunderstanding
in the minds of the people as to their rights in Canal Water." Bennett thus directed
that the word should be avoided in all subsequent correspondence and replaced with
the phrase rakba nahri mujawiza ("permissible canal area"), which carried the appropriate
sense of technical contingency and had none of the politically charged legal resonances
that the word haq carried (Bennett 1910). That this was unlikely to replace haq in
everyday conversation with irrigators (for whom the word offered potential
empowerment) was pointed out by at least one engineer, but it was apparently
hoped that this would at least clarify the Irrigation Department's position in debates
within the government. But so long as "rights" held an important place in another
discourse of power linked to the state itself, such a solution was unlikely.

Conclusion

Indeed, no compromise or modification of language could eliminate the underlying

tension that had shaped the "haq" debate. But this tension did not, it should be
stressed, undermine expanding canal irrigation. Civil administrators and engineers
shared a wide range of values that, in spite of such disputes, fueled an extraordinarily
rapid expansion of British canal irrigation in the Indus Basin in the early twentieth
century.11 Whatever their differences, nearly all British officials shared a strong
commitment to the expansion of the Indus Basin irrigation system as the key to
expanding cultivation, increased government revenue, and enhanced government
prestige and control. The transformation of the Indus Basin by the middle of the
twentieth century into a densely populated region of largely commercialized agricultural
production was the result.

"The total area irrigated by government canals in Punjab increased from just over 1.32
million acres in 1878 to just over 9.06 million in 1918. In Sind, the increase was from
roughly 1.5 to 3.16 million acres from 1880 to 1918. By the mid-1940s, the canal-irrigated
acreage in Punjab totaled about 14 million acres, and in Sind about 4.5 million (Public
Works Department, Government of India 1919:47; Sorley 1968:464-67; Sharma 1948:45).

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
1144 DAVID GILMARTIN

But the political differences embedded in the contrasting d

empire" and "imperial science" did have critical consequences for the "development"
of the region. The British "science of empire," with its reliance on the systematic
description and classification of Indian society, underlay the power of an increasingly
bureaucratic state that mobilized indigenous communities in support of the colonial
order. By its very reliance on the scientific processing and ordering of local knowledge,
it defined a critical place within the structure of power for local people (i.e., those
whose power was rooted in local relationships and particularistic idioms). State concerns
with "discipline" and "control" were certainly not lacking, but to the degree that
the object of social scientific knowledge in India was both the individual and the
"cultures" and "communities" of India, the discourse of "scientific" administration
was one in which powerful Indians were joined. Recognition in the courts and in
the administration of the language of "rights" and "custom" (usually closely joined-
and distinct from universal natural rights) reflected this and was thus critical to
the entire colonial political structure.
This discourse of "scientific empire" was in some ways not separate from that
of "imperial science," with its emphasis on the classification and transformation of
the natural world. In some respects, the one reinforced the other. As Michael Adas
writes, "The small numbers of Europeans who actually governed the colonized peoples
relied on their superior technology not only for the communications and military
clout that made the ongoing administration of vast areas possible but also for the
assurance that they had the 'right,' even the 'duty,' to police, arbitrate disputes,
demand tribute, and insist upon deference" (Adas 1989:204). But in the colonial
political context, the application of transformative technology carried implications
that were in many ways radically different from those inherent in the discourse of
"'scientific empire." These implications were, in fact, evident in the administrative
structure of the state itself, where generalist administrators were increasingly joined
in the late nineteenth and early twentieth centuries by cadres of technical "experts,"
concerned with forestry, medicine, irrigation, and other technical concerns. As colonial
administrators, these men shared many attitudes, but the differences between them
were in large part defined by the central value placed by generalists on the
understanding of particularistic "local knowledge" within a scientific discourse. For
the technical experts, concerned primarily with controlling the natural environment,
"local knowledge" was subordinated to a universal, technical discourse of "science."
Indeed, only subordination of "local knowledge" to the universal principles of science
would allow a productive transformation of the environment such as they sought,
a concern central to the field of irrigation.
The exclusion of "local knowledge" from the discourse of power, however,
presented serious political problems-not because Indians themselves clung to
"traditional" ideas-but because the scientific processing of "local knowledge" was
central both to the discourse and to the political system that defined the power of
the colonial state itself. As debates on irrigation "rights" illustrated, the imperatives
of technological transformation raised political problems that could not be easily
resolved within the colonial political framework. In this sense, the problem for
colonial irrigation was not the conflict between technical, engineering values and
the cultural imperatives of indigenous social organization, but rather the conflicts
internal to the ideology and scientific authority of the colonial state itself. And as
the colonial era drew to a close, these conflicts raised serious questions also concerning
the political foundations for a state-directed program of economic "development."

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
 SCIENTIFIC EMPIRE AND IMPERIAL SCIENCE 1145

This article has only hinted at some of the conflicts that emerged in this context
in the earlier decades of irrigation expansion in the Indus Basin. With the large-
scale settlement of cultivators in the newly irrigated wastes of Punjab and Sind,
these problems became all the more evident in the twentieth century. Environmental
problems, particularly those of salinity and waterlogging, multiplied as the integration
of irrigation works into a vast network proceeded apace. 12 Political problems persisted
as well. These were evident in the continuing tensions between political "privilege"
and irrigating "efficiency" in the canal colonies of the Punjab, and in the "corruption"
created by the ongoing interaction between a lower-level irrigation bureaucracy
(intended to carry out the Irrigation Department's mandate) and local irrigators
whose power was embedded in the discourse of "custom" and "privilege" supported
by the state itself. Some scholars have seen these problems as significantly undermining
the Indus Basin's potential for development, in spite of the massive colonial expansion
of irrigated acreage (e.g., Ali 1987; Merrey 1983; for a more positive view, see
Michel 1967; Lieftinck, et al. 1968; Sir Alexander Gibb & Partners, et. al. 1966).
The gradual emergence of an indigenous, educated "middle class" in the Indus
Basin and the movement toward self-government, of course, changed the political
dynamics of state policy. Educated Punjabis and Sindis showed an increasing interest
in technical science. But the politics of colonial Punjab and Sind were noteworthy
for the particularly sharp cleavage in the twentieth century between urban and rural
interests, a product, in part, of the discourse of "scientific empire." Though the
urban middle class gained in political influence, its power was limited by the
contradictions inherent in the colonial state system itself.
Partition in 1947 split the Indus Basin in two. Although transformations in
the discourse of state power since independence have been substantial (and divergent
in India and Pakistan), the role of scientific discourse in defining state power remains
important for understanding post-colonial programs of irrigation development.
Controversies surrounding the role of local organization, for example, continue to
influence irrigation policy in both India and Pakistan. With low irrigated yields a
common problem in Indus Basin agriculture (particularly in Pakistan), considerable
scholarly attention has focused on the roles of local communities as key institutions
in affecting irrigation water's effective use (e.g., Chambers 1988; Uphoff 1986;
Radosevich 1975; Painter, et al. 1982). This concern, in fact, spurred government
efforts in both countries in the 1970s and 1980s to establish new forms of local
water-users' associations to manage local water use within the context of larger
irrigating networks. Encouraged by Western experts, this was part of a more general
recent effort to encourage "appropriate" technology aimed toward the more effective
adaptation of technology to the local social and natural environment. But it also
suggests the continuing relevance for development of an understanding of the place
of "local community" and "local knowledge" within the discourse of power and
knowledge supporting the state. New users' associations have hardly been an
unqualified success. If the history of colonialism and irrigation provides a clue, the
problem of "appropriate" technology may have as much to do with the place of
technology and local organization,within the scientific discourses of power underlying

12The history of the environmental problems caused by waterlogging and salinity is complex.
By 1970 the percentage of arable land exhibiting waterlogging and salinity was close to 20
percent in many districts and higher in others (Siddiqi 1981:528; see also Michel 1967:455-
514). A general history of the environmental impact of canals on the Indus Basin, however,
remains to be written.

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
1146 DAVID GILMARTIN

the political system as with its seemingly objective "appropriateness" to the social
or natural environment of the Indus Basin.

List of References

ADAS, MICHAEL. 1989. Machines as the Measure of Men: Science, Technology and
Ideologies of Western Dominance. Ithaca: Cornell University Press.
ALI, IMRAN. 1987. "Malign Growth? Agricultural Colonization and the Roots of
Backwardness in the Punjab." Past and Present 114.
. 1988. The Punjab under Imperialism, 1885-1947. Princeton: Princeton
University Press.
AMIN, SHAHID. 1989. "Editor's Introduction." In William Crooke, A Glossary
of North Indian Peasant Life, ed. by Shahid Amin. Delhi: Oxford University
Press.
BARRIER, N. G. 1967. "The Punjab Disturbances of 1907: The Response of the
British Government in India to Agrarian Unrest." Modern Asian Studies 1.4.
BELLASIS, E. S. 1907. Note by Superintending Engineer, Derajat Circle, on questions
raised by the Deputy Commissioner, Dera Ghazi Khan, August 15, 1907. Punjab
Public Works Department (hereafter, PWD), Irrigation, file no. 412 of 1909.
PWD Secretariat, Lahore.
. 1909. Report dated 5 May 1909, and undated note included in W. R.
H. Merk to Senior Secretary to Financial Commissioners, 16 Sept. 1908. Punjab
PWD, Irrigation, file no. 412 of 1909. PWD Secretariat, Lahore.
BENNETT, W. E. T. 1910. Chief Engineer's summation of arguments on the
discontinuation of the word "haq" in official correspondence. Punjab PWD,
Irrigation, file no. 74 of 1910. PWD Secretariat, Lahore.
CHAMBERS, ROBERT. 1988. Managing Canal Irrigation: Practical Analysis from South
Asia. Cambridge: Cambridge University Press.
CLARKE, R. T. 1908. Deputy Commissioner, Multan to Commissioner, Multan,
August 11, 1908. Punjab PWD, Irrigation, file no. 412 of 1909. PWD
Secretariat, Lahore.
COHN, BERNARD. 1987. An Anthropologist among the Historians and Other Essays.
Delhi: Oxford University Press.
CRAIK, H. D. 1916. Note by the Revenue Secretary, June 2, 1916. Punjab PWD,
Irrigation, file no. 78 of 1898. PWD Secretariat, Lahore.
CROFTON, COL. J. 1875. Letter to Secretary to Government, Punjab, March 9,
1875. Punjab Revenue, Agriculture and Commerce, A Procs., September 1875,
no. 8. India Office Records, London.
DEWEY, CLIVE J. 1991. The Settlement Literature of the Greater Punjab. Delhi: Manohar.
DUTHY, G. W. 1919. "Remodelling Inundation Canals in the Muzaffargarh
District." In Minutes of the Proceedings of the Punjab Engineering Congress VII.
ELPHINSTONE, CAPT . N . W. 1860. Report of the Revised Settlement of the Googaira
District. Lahore: Punjab Government.
FIFE, J. G. 1871. Note by Under-Secretary, PWD, January 24, 1871. Bombay
PWD, Irrigation, Vol. 753 of 1868-1890, no. 125 of 1871. Maharashtra State
Archives, Bombay.
FOUCAULT, MICHEL. 1977. Discipline and Punish, trans. by A. Sheridan. New
York: Pantheon.

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
 SCIENTIFIC EMPIRE AND IMPERIAL SCIENCE 1147

FREITAG, SANDRIA B. 1991. "Crime in the Social Order of Colonial North India,"
Modern Asian Studies 25.2.
FROST, H. F. B. 1904. Note by Superintending Engineer, Bari Doab Circle, on
the History of the Bari Doab Canal. Punjab PWD, Irrigation, file no. 355 of
1904. PWD Secretariat, Lahore.
GEE, C. D. 1914. Note by Chief Engineer, May 19, 1914. Punjab PWD, Irrigation,
file no. 78 of 1898. PWD Secretariat, Lahore.
GIBB, SIR ALEXANDER, & PARTNERS, INTERNATIONAL LAND DEVELOPMENT
CONSULTANTS, AND HUNTING TECHNICAL SERVICES LIMITED. 1966. Programme
for the Development of Irrigation and Agriculture in West Pakistan. London and
Arnhem: International Bank for Reconstruction and Development.
GLADSTONE, C. E. 1879. Deputy Commissioner, Dera Ghazi Khan to
Commissioner, Derajat, December 8, 1879. Punjab Foreign, A Procs., January
1880, no. 12. India Office Records, London.
GORDON, W. B. 1909. Note by Chief Engineer on complaints against the canal
administration in the districts of Multan, Muzaffargarh and Dera Ghazi Khan
(with abstract of petitions submitted to the Chief Engineer), May 31, 1909.
Punjab PWD, Irrigation, file no. 412 of 1909. PWD Secretariat, Lahore.
HABIB, IRFAN. 1970. "Presidential Address." In Proceedings of the Indian History
Congress, 31st Session, Varanasi, 1969. Patna: Indian History Congress.
HEADRICK, DANIEL R. 1988. The Tentacles of Progress: Technology Transfer in the
Age of Imperialism. New York: Oxford University Press.
HOLMES, W. F. 1916. Note dated August 11, 1916. Punjab PWD, Irrigation,
file no. 78 of 1898. PWD Secretariat, Lahore.
HOUK, IVAN E. 1956. Irrigation Engineering, II. New York: John Wiley and Sons.
JACOB, LT. COL. JOHN. 1855. "Official Correspondence on the Abolition of
Statute or Forced Labour in Sind," February 23, 1855. Selections from the
Records of the Bombay Government, no. XXXIV-New Series. India Office
Records, London.
JAMES, LT. HUGH. 1849. "Canals of the Mooltan District" (November 29, 1849).
Selections from the Records of the Punjab Administration, Old Series, No. 1.
India Office Records, London.
LAMBRICK, H. T. 1975. John Jacob ofJacobabad. Karachi: Oxford University Press.
LIEFTINCK, PIETER, A. ROBERT SADOVE, and THOMAS C. CREYKE. 1968.
Water and Power Resources of West Pakistan: A Study in Sector Planni
Baltimore: Johns Hopkins University Press, for the World Bank.
LYALL, JAMES B. 1882. Memorandum by Financial Commissioner, Punjab, August
2, 1882. Punjab Revenue and Agriculture, Irrigation, A Procs., November
1882, no. 1. India Office Records, London.
MAYNARD, H. J. 1907. Commissioner, Multan to Superintending Engineer, Derajat
Circle, November 19, 1907. Punjab PWD, Irrigation, file no. 412 of 1909.
PWD Secretariat, Lahore.
MEREWETHER, W. L. 1870. Commissioner in Sind, to Sir W. R. Seymour V.
Fitzgerald, Governor of Bombay, October 1, 1870. Bombay PWD, Irrigation,
Vol. 753 of 1868-1890, no. 466 of 1870. Maharashtra State Archives, Bombay.
MERK, W. R. H. 1908. Commissioner, Multan to Junior Secretary to Financial
Commissioners, September 16, 1908. Punjab PWD, Irrigation, file no. 412
of 1909. PWD Secretariat, Lahore.
MERREY, DOUGLAS. 1983. "Irrigation, Poverty and Social Change in a Village
of Pakistani Punjab: An Historical and Cultural Ecological Analysis." Ph.D.
diss., University of Pennsylvania.

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
1148 DAVID GILMARTIN

MICHEL, ALOYS A. 1967. The Indus Rivers: A Study of the Effects of Partition. New
Haven and London: Yale University Press.
NASIR, SARDAR A. D. 1988. The Canal and Drainage Laws of Pakistan. Lahore:
Mansoor Book House.
O'DWYER, M. F. 1898. Note by the Lieutenant-Governor, June 5, 1916. Punjab
PWD, Irrigation, file no. 78 of 1898. PWD Secretariat, Lahore.
PAINTER, JAMES E., EMILY BALDWIN, SANDRA MALONE, ERNEST T.
SMERDON, AKBAR S. AHMED, MASUD A. SIDDIQUI, AND MAHMUD
KHAN, 1982. The On-Farm Water Management Project in Pakistan. Project Impact
Evaluation No. 35. United States Agency for International Development (June
1982).
PERKINS, H. E. 1882. Commissioner, Multan to Secretary to Financial
Commissioners, May 8, 1882. Punjab, Revenue and Agriculture, A Procs.,
August 1882, no. 4. India Office Records, London.
PRINSEP, E. A. 1865. Report of the Settlement of the Sealkote District, 1855-1860.
Lahore, Punjab Government.
Public Works Department, Government of India. 1919. Irrigation in India: Review
for 1917-1918. Simla: Government of India.
Public Works Department, Irrigation Branch, Punjab. 1940. Departmental noting,
May 22, 1940. Punjab PWD, Irrigation, file no. 124 of 1908. PWD Secretariat,
Lahore.
. 1943. A Manual of Irrigation Practice, 1st ed. Lahore: Punjab Government.
QAZILBASH, NAWAB NISAR ALI KHAN. 1931. Petition to Deputy Commissioner,
Lahore, October 5, 1931. Punjab PWD, Irrigation, file no. 45 of 1907. PWD
Secretariat, Lahore.
RADOSEVICH, GEORGE E. 1975. "Water User Organizations for Improving Irrigated
Agriculture: Applicability to Pakistan." Water Management Technical Report
No. 44. Fort Collins: Engineering Research Center, Colorado State University
(December 1975).
SANDEMAN, CAPT. ROBERT. 1870. Deputy Commissioner, Dera Ghazi Khan,
to Commissioner, Derajat, August 4, 1870. Punjab Land Revenue, A Procs.,
October 1870, no. 5. Punjab Archives, Lahore.
SAUNDERS, LESLIE S. 1873. Report on the Revised Land Revenue Settlement of the
Lahore District. Lahore: Punjab Government.
SHARMA, K. R. 1948. Irrigation Engineering, I. Jullundur: India Printers.
SHOUBRIDGE, H. 0. B. 1898. "Desert Canal, Remodelling Project." Bombay
PWD, Irrigation, Vol. 79 of 1899-1903, file no. 50 of 1899. Maharashtra
State Archives, Bombay.
SIDDIQI, AKHTAR HUSAIN. 1981. "Salinization and Waterlogging in the Upper
Indus Plain (Punjab) During the Early Colonial Period: 1850-1900 A.D.: A
Problem of the Past Neglects." Asian Profile 9.6 (December 1981).
SINGH, CHETAN. 1985. "Well-irrigation Methods in Medieval Punjab: the Persian
Wheel Reconsidered." Indian Economic and Social History Review 22. 1.
SINGH, HIRA. 1927. Member of the Legislative Council, of Narli, Lahore District,
to Sir Fazli Husain, December 3, 1927. Punjab PWD, Irrigation, file no. 45
of 1907. PWD Secretariat, Lahore.
SORLEY, H. T. 1968. The Gazetteer of West Pakistan: The Former Province of Sind.
Karachi: Government of West Pakistan.
STEEDMAN, E. B. 1882. Report of the Revised Settlement of the Jhang District,
1874-1880. Lahore: Punjab Government.

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms
 SCIENTIFIC EMPIRE AND IMPERIAL SCIENCE 1149

STONE, IAN. 1984. Canal Irrigation in British India: Perspectives on Technological

Change in a Peasant Economy. Cambridge: Cambridge University Press.
TEMPLE, RICHARD. 1852. Memorandum, September 9, 1852. Punjab Revenue,
April 17, 1852, nos. 92-102. Punjab Archives, Lahore.
UPHOFF, NORMAN. 1986. Improving International Irrigation Management with Farmer
Participation: Getting the Process Right. Boulder and London: Westview Press.
WARD, T. R. J. 1917. Comments of the Punjab Chief Engineer, Irrigation, in
minutes of committee to consider the remodelling of outlets of the UBDC,
November 29-30, 1916. Punjab PWD, Irrigation, Works, April 1917, nos.
1-4 ("Reduction of Outlets on the Upper Bari Doab Canal"). Punjab PWD
Secretariat Library, Lahore.
WHITCOMBE, ELIZABETH. 1982. "Irrigation." In Dharma Kumar, ed., The
Cambridge Economic History of India, II. Cambridge: Cambridge University Pr
WORSTER, DONALD. 1985. Rivers of Empire: Water, Aridity and the Growth of t
American West. New York: Pantheon Books.
YANG, ANAND. 1989. The Limited Raj: Agrarian Relations in Colonial India, Saran
District, 1793-1920. Berkeley: University of California Press.

This content downloaded from 70.162.91.107 on Mon, 25 May 2020 10:42:24 UTC
All use subject to https://about.jstor.org/terms