Precolonial Intellectuals and the

Production of Colonial Knowledge

PHILLIP B. WAGONER
Wesleyan University

Recent years have seen the emergence of a lively debate over the nature of
“colonial knowledge”—those forms and bodies of knowledge that enabled Eu-
ropean colonizers to achieve domination over their colonized subjects around
the globe. Lying at the heart of the debate are two opposing evaluations of the
role played by colonized subjects in the production of colonial knowledge. One
position holds that the role of the colonized was negligible—at most, permit-
ting some of them to serve as passive informants, providing raw information to
the active European colonizers who produced the new knowledge by imposing
imported modes of knowing upon the raw data of local society. In contrast, the
other holds that indigenous intellectuals in reality contributed actively to the
process, and that colonial knowledge was thus produced through a complex
form of collaboration between colonizers and colonized, and an attendant
process of epistemic confrontation and adjustment between European and in-
digenous knowledge systems. Although this debate has focused primarily on
one colonial context—that of British India—it has important ramifications for
the broader history of colonialism, and is complemented by contributions re-
lating to other areas of European colonialism (Cooper and Stoler 1997:11–18).
The first position—which, at the risk of oversimplification, I will refer to as
postcolonialist—is well known through the influential works of Edward Said
(1978), Ronald Inden (1986, 1990) Bernard Cohn (1987, 1996), Nicholas Dirks
(1989, 1993, 2001), Gauri Viswanathan (1990), and Thomas Metcalf (1994),
among others. To these scholars we owe the rich development of the insight,
ultimately deriving from Gramsci and Foucault, that European colonial con-
quest was dependent not just upon superior military, political, and economic
power, but also upon the power of knowledge—or “cultural technologies of
rule” in Dirks’ formulation (2001:9). These scholars have shown how a whole

This essay has benefited greatly from discussions with Richard Eaton, Peter Gottschalk, Sumit
Guha, Brian Hatcher, Lisa Mitchell, Vijay Pinch, Sheldon Pollock, Peter Schmitthenner, and Cyn-
thia Talbot, as well as from the comments and suggestions of four anonymous readers and CSSH
editor Thomas Trautmann, to all of whom I am much indebted. Two periods of research in the
Mackenzie collections at the British Library were made possible by a USDE Fulbright-Hays Fac-
ulty Research Fellowship (1999 –2000) and by a Project Grant from Wesleyan University (2001).

0010-4175/03/783–814 $9.50 © 2003 Society for Comparative Study of Society and History

783

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 784 phillip b. wagoner

host of colonial forms of knowledge—ranging from more clearly administra-

tive forms like the census to seemingly more “pure” forms of inquiry like his-
tory and ethnology—have not simply described an objective reality, but rather,
have called that reality into being in ways that served the interests of the colo-
nial state. They have also highlighted the self-perpetuating, reciprocal rela-
tionship between colonial knowledge and conquest: if knowledge enables con-
quest, this in turn enables the production of further knowledge, leading to more
effective consolidation of political power. Although this colonial knowledge is
largely produced in the colony, it is not of the colony: rather, it comes about
through the conjunction of pre-defined, imported European forms of knowl-
edge—the “investigative modalities” of Cohn (1996:5 ff.)—with the raw data
provided by the indigenous social and cultural forms of the colonized society.
In other words, colonial knowledge is produced by the active agents of the col-
onizing society, operating upon the passive patients of the colonized. Because
of the attendant loss of agency on the part of the colonized, and because in-
digenous categories and forms of thought are not just ignored, but forcibly dis-
placed by the imported epistemes of the colonizers, the production of colonial
knowledge thus represents a form of “epistemological violence” waged by the
colonial state against its colonized subjects (Dirks in Cohn 1996:xii). An im-
portant corollary following from this position is that colonialism introduces a
profound epistemic disjuncture, a rupture in the historical fabric of the society
subjected to colonialism. There can be no significant continuities across the
great rift generated by colonial knowledge, for all indigenous forms of knowl-
edge and bodies of cultural practice are effectively superceded and displaced
through the imposition of new, imported epistemes.
A second position, largely conceived as a revisionist critique of this post-
colonialist view, has emerged more recently through the writings of Eugene
Irschick (1994), C. A. Bayly (1996), Thomas R. Trautmann (1999a, 1999b),
William R. Pinch (1999), Richard M. Eaton (2000), Norbert Peabody (2001),
Mohamad Tavakoli-Targhi (2001), and others. Most of these contributions
would appear to share the point of departure of the established postcolonialist
position—namely, that knowledge is power and that colonial knowledge did
play a fundamental role in the consolidation of colonial rule. They depart from
the postcolonialists, however, in questioning the assumption that the colonized
were mere passive bystanders to the process, and in pointing to significant el-
ements of continuity, epistemic and otherwise, running across the presumed
colonial divide. In particular, some of these works have called attention to the
active role of the colonized in producing colonial knowledge. Whereas the post-
colonialists have tended to see native scholars merely as “informants,” provid-
ing raw data with which active Europeans produced colonial knowledge, the
“collaborationists” have instead viewed these indigenous intellectuals as active
partners in the process, bringing their own forms of knowledge and epistemic
regimes to the dialogue. Thus, in a recent issue of this journal, Norbert Peabody

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 785

has suggested that the position argued by Arjun Appadurai—namely, that caste-
wise censuses represent a colonial innovation introduced into India by the
British, following upon the prior development of an “enumerative habit” in
the domain of British land surveys (see Appadurai 1993)—should be “turned
on its head” (Peabody 2001:823). Peabody demonstrates that to the contrary,
enumeration by caste in early colonial censuses was conceptually dependent
on a familiarity with indigenous precolonial registers of households, such as
Munhata Nainsi’s “Account of the Districts of Marwar” (produced between
1658 and 1664), which in fact had already adopted a caste-based form of clas-
sification. Indeed, the earliest British censuses reveal that colonial officials had
not originally been interested in classifying the populations according to fine-
grained caste distinctions. Rather, Peabody suggests that “this style of classi-
fying the population appears to have crept into the colonial census in a some-
what more backhanded manner owing to the reliance of British administrators
on native informants and petty officials who were familiar with precolonial
‘household lists’ that had long been caste-sensitive . . .” (2001:841). Peabody’s
analysis also suggests that what these indigenous collaborators chose to con-
tribute—or to withhold—was determined by a desire to “further their own lo-
cally determined agendas” (2001:841), thus pointing to the inadequacy of the
starkly dichotomized view of the postcolonialists, which reduces all colonizers
to the role of agents and all colonized to the status of patients.
If indeed colonial knowledge was formed through such a process of intel-
lectual “dialogue” or “conversation” (the terms are Irschick’s and Trautmann’s,
respectively), one would expect to find the impress of indigenous conceptual
categories and even forms of thought on the final form and content of the re-
sultant knowledge. This is precisely what we find in the case of the concept of
the Dravidian language family, as Trautmann has demonstrated in persuasive
detail (1999a and 1999b). Although the Dravidian concept is largely taken for
granted today—it is the generally accepted means of accounting for linguistic
similarities between the four major languages of the south Indian peninsula, as
well as numerous tribal languages dispersed throughout other parts of the sub-
continent—at the time of its formulation in early nineteenth-century Madras,
the concept was “revolutionary and unprecedented, both for Europeans and In-
dians.” It was “the product of a new way of looking at things that came about
through the interaction of European and Indian mental frames under colonial
conditions, creating something that went beyond the limits of each” (1999b:51).
According to Trautmann’s analysis, the European mental frame contributed the
concept of ramifying families of languages, based on the notion of a “segmen-
tary, branching Tree of Nations . . . whose origin was the narrative of Noah’s
descendents in the Book of Genesis” (1999b:44); while the Indian frame con-
tributed the Sanskritic discipline of vyākaran! a (“linguistic analysis”), accord-
ing to which any language could be reduced to two elements: a comprehensive
list of roots on the one hand, and a series of concise transformational rules, on

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 786 phillip b. wagoner

the other, that “like a computer program applied to an input stream of data,
would generate the entire lexicon . . . by application to the list of roots” (45).
These two mental frames were brought together through the colonial relation,
as British scholar-administrators employed learned brahmins and other pandits
to teach them the vernacular languages employed in the peninsula. Although
the British undeniably held the upper hand in this relationship and set the agen-
da for the conversation, the colonial knowledge thus produced would not have
taken the form it did, had it not been for the fact that the Indian intellectuals
provided not merely raw data, but a key analytical framework that led to the
formulation of the new form of knowledge. Moreover, neither the fact that
British colonizers had initiated and directed the dialogue, nor their subsequent
employment of the Dravidian hypothesis to further their colonial agenda, in any
way prevented later members of the colonized society from employing the same
knowledge to further a different, nationalist agenda of their own (Trautmann
1999b:52–54).

historical epigraphy as a form of colonial knowledge

The present essay attempts to further develop this “collaborationist” model,
through a case-study of one category of colonial knowledge—historical epi-
graphy—that has been almost totally overlooked to date.1 Although in many
other areas of the world the study of inscriptions serves merely as the hand-
maiden of history, in the study of India’s past, epigraphy holds a position of cen-
tral importance. The nature of ancient and medieval cultural practices leading
to the production of epigraphic texts, the documentary functions served by such
inscriptions, their sheer numbers, and the comparative meagerness of other in-
digenous historiographic traditions prior to the beginning of the second mil-
lennium, are all factors which have conspired to put epigraphy at the center of
the study of Indian history. Indeed, according to the estimate of D. C. Sircar,
one of the leading epigraphists of the twentieth century, something on the or-
der of 80 percent of what we know about the basic framework of Indian histo-
ry before 1000 a.d. is derived entirely from inscriptional sources (Sircar, cited
in Salomon 1998:3).2 Without them, we would have only the shakiest knowl-
edge of chronology, historical geography, the actual boundaries of regions and
territories, the changing nature of language use, and most forms of political, so-
cial, and economic life as they were actually lived. To borrow a metaphor from
the ubiquitous benedictory verse with which so many of these records begin,
epigraphy is truly the “central pillar” upon which the entire edifice of precolo-
nial Indian history rests.3

1 What would seem to be the sole exception is Daud Ali’s brief appendix, on “Archaeology and

the History of India,” to his essay “Royal Eulogy as World History: Rethinking Copper-plate In-
scriptions in Cola India” (2000:225–29). Although this appendix focuses more broadly on archae-
ology as a form of colonial knowledge, it briefly discusses epigraphy within this context.
2 The estimate is D. C. Sircar’s, as cited by Richard Salomon (1998:3).
3 The verse is the famous benediction borrowed from the opening of Bān !a’s Hars! acarita, prais-

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 787

While the inscriptions themselves were present and available long before the
advent of European colonialism, they had not been systematically used as an
avenue for understanding the past by precolonial Indians.4 This historical use
of inscriptions was only a product of the colonial encounter, as British admin-
istrators, scholars, and intellectuals strove to understand the pasts of the terri-
tories they were acquiring. This development has been described in detail by
Richard Salomon, in terms of a model which has the British bringing the dis-
cipline of epigraphy to India, essentially as a pre-defined set of theories and
practices derived from the study of the classical European past, and applying it
to the new body of data represented by the inscribed texts found in the new en-
vironment. The only problems they faced in this endeavor were, first, that the
alien languages used in these texts had to be learned, and second, that the un-
recognizable forms of their archaic scripts had to be deciphered, so that the texts
themselves could be read. The first difficulty was overcome readily enough by
employing native Indian assistants to teach the relevant classical and vernacu-
lar languages to the European scholars; the second was over time surmounted
by the collective efforts of the Europeans themselves, by applying paleograph-
ic principles (putatively another European import). This method allowed them
to work backward step by step from the scripts currently employed, to the
slightly earlier forms used in older Sanskrit manuscripts, and then through pro-
gressively older epigraphic scripts, all the way back to the ancient Brahmi and
Kharoshthi alphabets that had been used in the third century b.c. (Salomon
1998:199–217).
The hub of all this epigraphic activity described by Salomon was late eigh-
teenth- and early nineteenth-century Calcutta, the heart of the Bengal Presi-
dency and center of the British colonial government in India. Orientalist schol-
arship in Calcutta was institutionalized in a triad of overlapping institutions,
which often shared the same personnel, both British and Indian. These were the
College of Fort William (founded in 1800 as a center for instruction in Indian
languages for young British civil servants), the Asiatic Society of Bengal, with
its influential journal Asiatic Researches, and, of course, the government itself
(Trautmann 1999b:37). Epigraphic work was carried out by a number of major
Calcutta orientalists, including the likes of Charles Wilkins (1749–1836), Hen-
ry Colebrooke (1765–1837), and William Jones (1746–1794).5 Although it is
acknowledged that their Indian assistants played an important role in this de-

! bha) upon which rests the cosmic edifice of

ing Siva in the form of the “central pillar” (mūla-stam
the three-fold universe: “Adoration to Śam! bhu [!Siva], who is the main pillar in the building of
the city of the three worlds—beautiful with the moon hung on him like a white royal chowrie, as
it kisses his lofty brow” (Cowell and Thomas 1968:1).
4 There is some indication that more recent inscriptions were appealed to as documentary evi-

dence in support of property claims, but more ancient inscriptions were not even readable because
the forms of the various scripts used had changed so significantly in the intervening period (Sa-
lomon 1998:199).
5 For a detailed review of their epigraphic activities, see the survey by Salomon (1998:199–

203).

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 788 phillip b. wagoner

velopment of epigraphy, it is nonetheless understood to have been largely a sec-

ondary one—whether they were serving as language teachers and linguistic in-
formants, or as understudies to the active Europeans, learning to apply the new,
imported epigraphic methods in the service of their colonial masters.6

col. colin mackenzie and colonial epigraphy in madras

Such a scenario may very well have been true for Calcutta, but things were quite
different in the Madras Presidency to the south. The undeniable importance of
the early Calcutta epigraphers, especially for their leading role in the decipher-
ment of the most ancient Indian scripts, has hindered recognition of a simulta-
neous and equally vital tradition of epigraphic study located in the Madras Pres-
idency. In certain respects, as I shall argue below, this Madras school of colonial
epigraphy would ultimately exercise an even greater impact on the subsequent
development of the discipline, particularly with respect to how it was eventu-
ally institutionalized within the Archaeological Survey of India (founded 1861).
While the Calcutta school is embodied in the work of many individual schol-
ars, the primary locus of the early Madras school was restricted largely to a sin-
gle collaborative project undertaken by the well-known British surveyor, Col.
Colin Mackenzie (1753–1821) together with a team of some twenty Indian “as-
sistants.”
The fact that modern scholarship has thus far failed to recognize the contri-
butions of Mackenzie’s project is in keeping with the larger tendency to assume
the primacy of Bengal as a center for Orientalist scholarship in nineteenth-
century India.7 Recently, however, it has been suggested that the particular con-
cerns and institutions of the Bengal school were not reproduced in other parts
of the subcontinent, and that Madras, in particular, was the home of a distinct
school of Orientalist scholarship that differed in decisive ways from its better
known Calcutta counterpart. Credit for recognizing a “Madras school of Ori-
entalism” goes to Trautmann, whose studies of the role of the Madras school in
the formation of the Dravidian concept have been mentioned above (1999a;
1999b). Although Trautmann has called attention to the Madras school’s de-
pendence on the Calcutta school in its institutional embodiment (thus, the
College of Fort William was replicated in Madras’s College of Fort St. George,
and the Asiatic Society of Bengal was reproduced in the Literary Society of
Madras), he stresses that the Madras school was in competition with its Cal-
cutta counterpart, as it “sought to assert its authority over the Orientalist study
of the Indian South, against the Orientalist establishment of Calcutta and the
Asiatic Society” (1999a:59). Thus, the “discovery” of a Dravidian family by

6 As, for example, in the case of some of Jones’s learned pandits, including Rādhākānta Tarka-

vāgı̄śa and Rāmalocana Pandit, both of whom contributed their own epigraphic articles to Asiatic
Researches. See Salomon 1998:202, and, on Rādhākānta, Rocher 1989 and 1993.
7 On various aspects of the “Bengal School” and its history see the works of Kopf (1969), Ke-

jariwal (1988), Rocher (1989 and 1993), and Hatcher (1996).

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 789

Francis Whyte Ellis and his Indian collaborators went against the official doc-
trine of the Calcutta school, which held that “all the languages of India, north
and south . . . were descendants of Sanskrit” (Trautmann 1999a:58).8
Trautmann’s notion of a “Madras school” provides a productive analytical
framework for approaching Mackenzie’s project and understanding its unique
contributions to the emergent field of historical epigraphy. Mackenzie was first
and foremost a cartographer, whose influential Survey of Mysore (1800–1810),
carried out at the behest of the colonial government after the defeat of the
Mysore ruler Tipu Sultan and the annexation of his territories, has been called
“the first completely organized survey expedition to take the field in India”
(Phillimore 1950:93).9 Mackenzie’s widespread recognition as a surveyor led
to his being appointed as the first Surveyor General of India, a post which he
held from 1815 until his death in 1821. To historians, however, he is better
known for the vast archive of inscriptions, “local histories,” and other manu-
script materials that he collected during the course of his cartographic surveys
in southern India. It is important to recognize, however, that these two activi-
ties of Mackenzie’s—mapmaking and collecting—were integral components
of a single, governmentally sponsored colonial project. Thus, in the official or-
ders commissioning Mackenzie to undertake his Survey of Mysore, Richard
Wellesley, the Governor General of Madras, explicitly stipulated that the sur-
vey should not be confined merely to military and geographical mapping, but
should incorporate also a “statistical account of the whole country” which
would contribute directly to the successful administration of the new provinces:
[S]uch a survey is in the first place absolutely necessary to the accurate settlement of
our frontier; it will also tend to augment our knowledge of Indian Geography, and to pro-
duce immediate and important benefits in establishing and conducting our government
in the conquered provinces, for I propose that the attention of the Surveyor should not
be confined to mere military or Geographical information, but that his enquiries should
be extended to a statistical account of the whole country, and that he should be supplied
with the best means in our power to assist him (quoted in Phillimore 1950:91; my em-
phasis).

8 In a similar vein, Peter Schmitthenner has identified differences relating to sponsorship, pro-

fessional identity, and linguistic orientation that set Madras Orientalism apart from similar schol-
arship in Bengal. Whereas Orientalist scholarship in Bengal received official government patron-
age, was carried out primarily by Company officials, and was devoted largely to the goal of
recovering the classical Sanskritic past, Indological research in Madras generally lacked direct gov-
ernmental sponsorship, was dominated by individuals who came from missionary backgrounds,
and tended to be directed more toward the study of vernacular languages and local regional cul-
tures (Schmitthenner 2001:28– 32). While this profile fits well for figures like Charles Philip Brown
and Robert Caldwell, it does not fit Mackenzie at all on the first two of its three counts, as Schmit-
thenner himself concedes (2001:29).
9 For the details of Mackenzie’s biography, see in addition the work of Phillimore (1945:349–

52 and 1950:419–28), and the introduction by David M. Blake to Cotton et al. 1992. For discus-
sion of Mackenzie’s cartographic work, see Phillimore 1950:91–121, 152– 56 passim; and Edney
1990:152ff. For discussions of Mackenzie’s work as a collector of texts and antiquities, see Cohn
1992 (repr 1996:76–105); and Dirks 1989 (and repr. in Seneviratne 1997:120–35) and 1993.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 790 phillip b. wagoner

The “statistical account” requested by Wellesley effectively mandated that

Mackenzie should investigate any and all aspects of the economic, social, and
political conditions of the country—both past and present—that might prove
relevant to its successful administration by the colonial state.10 Mackenzie
chose to do this by collecting ethnographic information, oral accounts, antiq-
uities, and written historical accounts, both literary and epigraphic.
Mackenzie’s conception of the potential value of inscriptions for the success
of his statistical account is revealed in a document he compiled in 1800, pro-
viding a list of thirty-two “Hints or Heads of Enquiry for facilitating our knowl-
edge of the more southerly parts of the Deckan” as guidelines for his statistical
enquiries.11 Under item number thirteen, “inscriptions on stone or metal” are
singled out as one of several types of sources that could be used—in the ab-
sence of “regular histories” among the Hindus—to reconstruct “the history of
the succession of the Rajahs, Naigs, Poligars, or Chief families” who issued
them, and eventually by extension, of the broader history of the Country itself
(OIOC, MSS Eur F 128/213 [pp. 3–6]).12 Accordingly, the “able native assis-
tants” who accompanied his surveying parties as translators were instructed to
seek out and collect any inscriptions they could find in the course of their pere-
grinations across the Mysore country. Although in some cases the actual stone
stelae or copper plates were physically transported back to Madras, in most
cases the epigraphic texts were collected as facsimiles drawn by hand, repro-
ducing as closely as possible the precise forms of the individual engraved let-
ters, their disposition across the surface of the inscribed stone or copper plate,
and any accompanying sculptural ornaments or borders. By 1807, the Indian
10 Clearly, as Cohn has suggested (1996:80), the model for Wellesley’s “statistical account” was

provided by Sir John Sinclair’s Statistical Account of Scotland. According to his own account in
the History of the Origin and Progress of the Statistical Account of Scotland (1798), it was Sinclair
who first introduced the word “statistics” into English, borrowing it from German, where it was
used in the sense of “an enquiry for ascertaining the political strength of a country, or questions re-
specting matters of state.” Sinclair indicated that he intended it in the slightly different (and appar-
ently more benign) sense of “an inquiry into the state of a country, for the purpose of ascertaining
the quantum of happiness enjoyed by its inhabitants, and the means of its future improvement (Sin-
clair 1798:v). Given the explicit nature of Wellesley’s charge to Mackenzie, it is a mistake to as-
sume, as Dirks seems to, that Mackenzie’s obsessive “interest in collecting narratives and facts”
somehow represented a private project in which he was engaged “on the side” while carrying out
his official, public duties as a surveyor (1989:45, 49).
11 See the copy presented to the colonial magistrate and friend of Elphinstone, Edward Strachey

(1774 –1832), preserved in Strachey’s papers in the British Library (OIOC, MSS Eur F 128/213,
pp. 3 – 6).
12 In thus recognizing the historical value of inscriptions, Mackenzie was in all probability fol-

lowing the lead of the Calcutta orientalists, who had turned to inscriptions after largely failing in
their attempts to access India’s ancient history from Sanskrit texts such as the Puranas. By the end
of the eighteenth century, they were turning increasingly to other sources, of which inscriptions
were hailed as the most revealing (see Trautmann and Sinopoli 2002:3– 6). As a member of the Asi-
atic Society of Bengal since 1797 (the year in which Captain Colin Mackenzie is first included in
the list of “Members of the Asiatic Society”; see Asiatic Researches, vol. 5), and as a reader of its
journal Asiatic Researches, Mackenzie would certainly have been familiar with the Calcutta Ori-
entalists’ growing disillusionment with the Puranas, and their attendant enthusiasm for inscriptions.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 791

members of his survey team had collected the texts of “upwards of 1100” in-
scriptions, and had translated and analyzed enough of them for Mackenzie to
be able to write an “Introductory Memoir on the use and advantage of inscrip-
tions and sculptured Monuments in illustrating Hindoo History” (OIOC, Mac-
kenzie General Collection, vol. 18, recensions A and B). We will return to this
important manuscript below, but first, let us consider more closely the nature
of the collaboration between Mackenzie and his Indian assistants.

mackenzie’s indian collaborators

Mackenzie himself was quite frank about his own ignorance of Indian lan-
guages, and openly acknowledged his intellectual debt to his Indian collabora-
tors. Writing to Charles Wilkins in 1808, for example, Mackenzie confessed
that “My own want of knowledge of the languages has rather impeded my
progress, but I have the advantage of able native assistants, and I have been for-
tunate enough to obtain much interesting materials . . .”13 While the importance
of Mackenzie’s collaborators has thus been long recognized, the few available
accounts of them have been largely restricted to the activities of the three Cavel-
ly (! Kāvali) brothers—Venkata Borriah, Venkata Lakshmiah, and Venkata
Ramaswamy—who were arguably the most important and influential members
of Mackenzie’s establishment.14 Before his untimely death in 1803 at the age
of twenty-seven, Borriah served briefly as Mackenzie’s “head translator,” and
had penned a number of scholarly articles in English on ancient Indian history,
as well as a number of Telugu literary works. Lakshmiah inherited the post of
“head translator” after Borriah’s death, and was assisted by the younger Ra-
maswamy who appears to have served as Lakshmaiah’s deputy and also worked
as a translator in Mackenzie’s Madras office. During the time of his service with
Mackenzie, Lakshmaiah was accorded the distinction of being the first Indian
to be inducted into the Madras Literary Society (until then an exclusively Eu-
ropean establishment) and later, in the early 1830s, he founded and became the
first president of the Hindu Literary Society in Madras, modeled on its Euro-
pean counterpart. Ramaswamy was likewise admitted to the Madras Literary
Society, and after what appears to have been a somewhat turbulent and check-
ered career with Mackenzie, went on to pursue a productive and distinguished
literary and scholarly career, and is remembered as the author of Biographical
Sketches of the Dekkan Poets (1829 and numerous later editions), Descriptive
and Historical Sketches of Cities and Places in the Dekkan (1828), and various
other works, including an English rendering of an early modern Telugu cook-

13 Mackenzie to Charles Wilkins, Esq., 25 Oct. 1808 (OIOC Mackenzie General collection, vol.

18 (recension A, bound in before page 279).

14 The Cavelly brothers, and especially the near-mythic Borriah, inevitably loom large in most

accounts of Mackenzie’s project. Of works devoted primarily to them, the most important is Ra-
machandra Rao 2003, which moves well beyond the sources in the Mackenzie archive to focus pri-
marily on the Kāvalis’ own intellectual activities in Telugu as well as in English.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 792 phillip b. wagoner

book (Pākaśāstra, otherwise called Sūpaśāstra, or the Modern Culinary

Recipes of the Hindoos, Madras 1836).
Compared to the Cavelly brothers, Mackenzie’s other collaborators have
proven far more elusive. From the published materials available, it is not even
possible to know with certainty how many different collaborators Mackenzie
employed, much less form a coherent idea of their individual personalities and
backgrounds. From various sources—including, most helpfully, the entries in
H. H. Wilson’s catalogue of the Mackenzie collection (1828; see under Trans-
lations, Class XII “Letters and Reports”)—one notes the names of some sev-
enteen or more assistants in addition to the three Cavelly brothers: these are
Cristna Row, Ananda Row, Sooba Row, Venkat Row, Narrain Row, Moba Row,
Seva Row, Sreenivassiah, Mallayya, Ramadoss, Nitala Naina, Gundapah Sin-
garachari, Baskeiah Bramin, Dharmaiah (and son), Abdul Aziz, and Swaramai-
ah Sastree. The thoughts and actions of the individuals behind these names can
be partially recovered, however, thanks to the many volumes of their unpub-
lished writings preserved in manuscript form in Madras and London. While
they were in the field working for Mackenzie, these assistants kept Lakshmai-
ah and Mackenzie informed of their progress by regularly dispatching letters
and detailed, monthly reports on their activities. The originals of these docu-
ments, written mostly in Telugu, Marathi, and Tamil, are preserved in the
Mackenzie collections of the Government Oriental Manuscripts Library in
Madras (Chennai). Because of Mackenzie’s own ignorance of Indian lan-
guages, he had Lakshmaiah and his staff translate them into English for his
benefit. After Mackenzie’s death, these translations were eventually sent to
London, where they are presently housed in the Oriental and India Office
Collections of the British Library.15 Thanks to the fortuitous preservation of
these richly detailed documents, it should eventually prove possible to recon-
struct with some precision the careers of most of these other Mackenzie col-
laborators.
It must be acknowledged that it was Cohn and Dirks, in their respective stud-
ies of the Mackenzie archive, who first called attention to the existence of these
translated letters and reports, but it should also be pointed out that they have by
no means exhausted the potential of these documents as historical evidence.
Thus, Cohn (1996:84) presents only a general, one-paragraph summary of Laksh-
maiah’s monthly reports for 1804 (Mackenzie Translations, Class XII, #9:39–
99), documents which shed little light on the particular styles and methods of

15 Apart from the logistical difficulties posed by the originals and their translations being housed

in separate libraries on two different continents, an additional measure of complexity is introduced

by the fact that the Madras Mackenzie collections have been re-catalogued, and are thus no longer
accessible under their Wilson catalogue designations. I am grateful to Peter Schmitthenner for his
generous help in confirming the continued existence and accessibility of these original letters and
reports. Unfortunately, however, I have not yet had the opportunity to consult the originals in
Madras.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 793

the majority of the less-known collaborators.16 Dirks has cast his net more
widely, discussing the activities of at least five collaborators other than the
Cavellys, but his approach remains largely anecdotal, and ultimately fails to
penetrate into the deeper system of relations that binds these men to the project.
For example, to support his proposition that Mackenzie’s assistants were
often foiled in their attempts to collect texts—on account of the local people’s
distrust of the British—Dirks quotes from the account of Narrain Row’s un-
successful interview at the court of the Maharaja of Gadwal (Mackenzie Trans-
lations, Class XII, no. 14[5]), and concludes by observing that Narrain Row was
honored with the customary gift of betel nut and then escorted out of the king-
dom empty handed (2001:103; 1993:293).17 This episode indeed provides a
vivid example of the difficulties Mackenzie’s collaborators often faced in their
work, but at the same time—if one considers it from a broader perspective—it
just as clearly demonstrates their dedication and ingenuity in surmounting these
obstacles. In fact, another letter, written a month later (but filed out of sequence
in the London archive), reveals that Narrain Row persisted in his attempts de-
spite this initial failure, and ultimately succeeded in acquiring “a good deal of
old papers of that Samastanum [i.e. the Gadwal court].” Narrain Row accom-
plished his task through the agency of a local acquaintance, whose father-in-
law was an employee of the court and agreed to help Narrain Row in return for
some “pecuniary assistance”; he was duly paid ten rupees for his efforts, plus
another ten for road expenses (see Translations Class XII, no. 14[3]).18

16 Cohn also presents a three-paragraph summary of the “Report of Baboo Rao,” describing his

inquiries carried out along the Coromandel coast, but this is based on the printed transcription of
H. H. Wilson, published in his 1828 catalogue of the Mackenzie collection, not on the original man-
uscript translation in the archive (Cohn 1996:84– 85). Wilson selected this specimen and included
it in his catalogue as an example of the process through which the literary mss. were collected by
Mackenzie’s team.
17 Dirks himself cites “‘Letters and Reports,’ [i.e. Mackenzie Translations, Class XII], nos. 26,

27” (2001:330, note 62), but the account in question is actually contained in Class XII, no. 14 [item
5]. Class XII, no. 26 contains three separate reports of Narrain Row’s, for the years 1811, 1812, and
1813, during which time he was working in the Ceded Districts (the present Rayalaseema region
of Andhra) further south from Gadwal, which was in the Nizam’s dominions (the present Kurnool
District of Telangana). According to the index listing in the Wilson catalogue, Class XII, no. 27 ap-
pears to have been a duplicate of the previous number, but it is marked “missing” in the British Li-
brary’s 1986 shelf-list.
18
Another example of the often significant inaccuracies in Dirks’ representation of these doc-
uments may be found in his discussion of the report describing the activities of “C. V. Ram” in
Aunomacondah and Vorungole (the present Hanamkonda and Warangal, in Andhra Pradesh; see
Class XII, no. 47). The report was actually written by Narrain Row, as is clearly indicated in the
heading at the top of its first page, “Report of Narrain Row from April to August 1816”; Dirks’
“C. V. Ram” is a misreading for “C. V. Ram.y,” the abbreviated signature of Cavelly Venkata Ra-
maswamy who translated Narrain Row’s report into English from its original Telugu, and certified
his work thus at the bottom of the last page: “True Tran.tion by C. V. Ram.y.” Dirks explains that
the author of the report befriended an aged, learned Brahman at Aunomacondah, and that “the sub-
sequent information in the report was then based almost exclusively on the words of the old Brah-
man” (2001:101). Careful reading of the report, however, reveals that this is not at all the case; to
the contrary, folios 41 verso through 52 recto are the synopsis of an old palm-leaf manuscript pre-

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 794 phillip b. wagoner

The point of these criticisms is in no way to belittle the important contribu-

tions of Cohn and Dirks, but simply to suggest that further research into the
Mackenzie archive is needed. Clearly, if we are to retrieve the personalities of
Mackenzie’s unknown assistants without falling prey to easy generalizations
and a priori assumptions, what is required is a far more systematic approach to
the documents than can be achieved simply by sampling their contents at ran-
dom. One possible method that suggests itself is to adopt a biographical frame
of analysis, focusing on the activities of individual figures, as revealed through
the totality of their preserved letters and reports. I have made a first attempt in
this direction by examining the career of just one of Mackenzie’s unknown col-
laborators: Narrain Row Brahmin, who worked for Mackenzie for a period of
fifteen years between 1803 and 1818. Much of the interpretation that follows
is predicated on my systematic analysis of the more than 100 translated letters
and reports of Narrain Row that are preserved in London, as well as upon my
careful reading of the one historiographic work written by Narrain Row that is
available in published form in its original Telugu—the Śrı̄śaila Dēvālaya
Kaiphiyatu, composed in 1810.19

sented to Narrain Row by the brahmin; 52v.– 59r. are taken up by Narrain Row’s own antiquarian
account of the ruins and inscriptions of the city of Orungole (Warangal), based on his own direct
observation and study; 59v.– 63v. are blank in the ms.; 64r.– 69v. are the synopsis of another histo-
riographic text procured by Narrain Row in Warangal; and finally, the last pages (70r.–75r.) recount
Narrain Row’s antiquarian observations carried out at sites in the vicinity of Warangal, and then his
return journey to Hyderabad.
19
The text was first published in an English translation by P. Sitapati in 1981, and was followed
by an edition in the original Telugu prepared by Sitapati the following year (see Nārāyan!a Rāvu
1810). Sitapati’s translation in fact abridges and summarizes the text and does not provide a reli-
able indication of the complex texture and historiographic sophistication of the original. Narrain
Row’s authorship of the text is explicitly indicated through the intriguing, one-sentence “colophon”
written in Persian (but in Telugu script) at the end of the text. This statement reads “dariyāpti
nārāyan! a rāvu mutassaddi yilāke mējar kālan mekanjı̄ sāhebu; tārı̄ku 13 māhe me sannu 1810
yisavi.” It can be recognized, however, as a Persian sentence that would be transcribed in ortho-
graphically correct form as: dar-yāfti narrain rāvu mutas! addı̄ !alāqa-i mejar kālan mekanjı̄-sāhab;
tārı̄kh 13 māh-i me sana 1810 !ı̄sawı̄. This can be translated as “An investigation of the accoun-
tant Narrain Rao, in the service of Major Colin Mackenzie-sahib, the thirteenth day of the month
of May, year 1810 of the Christian era.” Sitapati misconstrues the word dariyāpti as Narrain Rao’s
family name (in!!ti-pēru), and takes him to be “the scribe” who “recorded” the account (Sitapati
1981:xii). Sitapati was correct in identifying Narrain Rao as one of Mackenzie’s mutas! addı̄s, al-
though he was confused about his name and his precise role in producing the text. None of the let-
ters and reports I have examined identifies Narrain Rao by means of any in!!i-pēru.
t Nor does Sita-
pati seem to have registered the crucial point that this last sentence in the text is written in Persian.
Almost all of the words used have in fact been borrowed into Telugu from Persian or Hindustani,
but the syntax and grammar of this statement are unmistakably Persian, not Telugu. Although there
is a lacuna in the surviving letters and reports of Narrain Row’s activities, covering almost the en-
tirety of 1810, it is significant that his last recorded location before this hiatus, on 6 January 1810,
is at Kampili, in the present Bellary district of Karnataka, not far from Srisailam, located in the ad-
jacent Kurnool district of Andhra Pradesh. (See British Library, OIOC Eur. mss., Mackenzie Trans-
lations, Class XII/14, 4 [sheets 6v and 7], letter to Cavelly Venkata Lakshmaiah dated 6 Jan. 1810).
When Narrain Row reappears after this hiatus, in January of 1811, he is in the “Ceded Districts,”
that is, the present Cuddapah district in southern Andhra.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 795

niyogis at the arcot court

Before turning to some of the details of Narrain Row’s personal background
and contributions, it is necessary first to clarify his social identity as a brahmin.
Although the majority of Mackenzie’s assistants can in fact be identified as
brahmins, it would be a serious mistake to think of them in terms of the tradi-
tional stereotype of “the brahmin” as described in normative Dharma-śāstra
texts. Through their use of the distinctive caste-suffix “Rao” (Telugu, rāvu), a
majority of Mackenzie’s brahmins can be identified in far more specific terms
as Niyogis, a class of secular (laukika) brahmins who typically supported them-
selves by accepting appointments (niyōgamu) to political and administrative
service. Some Niyogis served at the village level as accountants (karan! am),
while others served in a more cosmopolitan courtly setting as ministers, resi-
dents, or other functionaries in the administrative bureaucracy.
The emergence of the Niyogi community can tentatively be traced back to
the twelfth and thirteenth centuries, in the period of the Kakatiya kingdom’s as-
cendancy over the Andhra country (Lakshmana Rao 1965[1923]). Although the
term Niyogi itself does not appear to be documented until much later, Kakatiya
inscriptions point to the existence of a loosely defined sub-class of brahmins
who took up secular occupations in the political society of the time. As Cynthia
Talbot has shown, these brahmins often identified themselves in inscriptions
with the title rāju, from the Sanskrit rājan and literally meaning “king” or
“prince,” but here used in the sense of one who is employed by a lord or prince
(Talbot 2001:57; other titles used include āmātya, mantri, and pregad! a, all used
in the sense of “minister”). These secular brahmins thus distinguished them-
selves from other brahmins who continued to follow their traditional calling as
religious specialists or scholars and used the titles bhat!!ta or pan! d! ita. Signifi-
cantly, this early brahmin usage of the title rāju appears to prefigure the later
title “Rao” (rāvu), which is likewise derived from rājan, but through the
Marathi form “Rao.” By the Vijayanagara period (fourteenth to sixteenth cen-
turies), the tradition of secular brahmin service in political and even military
contexts had become so widespread and well-established, that the early six-
teenth-century emperor Krishnadevaraya especially recommends that brah-
mins be appointed as commanders of forts.20 Although it appears to have first
originated in the Telugu country, the Niyogi brahmin community had by the late
eighteenth century become widely dispersed across most of the Deccan, reach-
ing more or less the area of its present distribution.21
20
See Krishnadevaraya’s Āmuktamālyada IV, 205 – 85, and especially verse 207. For transla-
tion and discussion, see Rangasvami Sarasvati 1926.
21 Today, Niyogis are found in significant numbers in four linguistic states of South India:

Andhra Pradesh, Maharashtra, Karnataka, and Tamil Nadu. Those living outside of Andhra are rou-
tinely bilingual, speaking Telugu in the home and using the local language—Marathi, Kannada, or
Tamil—when interacting with outside communities. See K. S. Singh 1998:2644–48 for a contem-
porary ethnographic account of the various Niyogi brahmin communities.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 796 phillip b. wagoner

From a traditional perspective, the salaried, administrative employment em-

braced by Niyogis was held to be demeaning, and resulted in the widespread
perception that Niyogi brahmins were inferior to their more conservative coun-
terparts in terms of purity and ritual hierarchy. But at the same time, this polit-
ical service could also be both economically and socially advantageous, as it
provided Niyogi families with regular sources of cash income and access to an
expanded range of social networks. Moreover, because of their willingness to
serve as social and economic intermediaries between the local world of the vil-
lage and the cosmopolitan world of the court, Niyogis as a class came to em-
body an unusual constellation of linguistic skills and attitudes toward language.
On the one hand, they tended to identify more closely with the local south In-
dian vernaculars than they did with Sanskrit, which was seen as the intellectu-
al preserve of their more traditional counterparts; on the other hand, they also
cultivated skills of literacy in a succession of cosmopolitan administrative lan-
guages, from Marathi and Hindustani to Persian and eventually English.
By the turn of the nineteenth century, when Mackenzie embarked upon his
Survey of Mysore, there was a major repository of Niyogi talent located right
in the city of Madras, concentrated primarily at the court of the Nawab of Ar-
cot.22 Although the primary seat of the Nawab had originally been in Arcot
town, some sixty-five miles west of Madras, much of the court bureaucracy had
shifted to Madras after 1768, when the Nawab Muhammad Ali built a new
palace just adjacent to Fort St. George.23 Contemporary records indicate that
by this period Arcot’s revenue administration had become thoroughly domi-
nated by Niyogi brahmins—an understandable development given the advan-
tages of their broad multilingualism and literacy related skills.
The most detailed information relating to the dominance of Niyogis in the
Arcot revenue administration comes from the records of Avadanum Paupia
Braminy et al. versus Reddy Row and Anunda Row, a celebrated legal case tried
in the Supreme Court at Madras on 10 October, 1808, and documented in co-
pious detail thanks to the publication of its proceedings in 1811 (see PREIA III
1811 and PREIA IV 1809). The two defendants in the case had been members
of the Arcot darbār during the reign of Umdat al-Umra (1795–1801) and had
allegedly been involved in a conspiracy to forge government bonds. Of the
eighty-four witnesses called during the course of the case, fifty-nine had been
employees of the Arcot darbār, and out of these fifty-nine, as many as thirty-

22 The Cavelly brothers were themselves Niyogi brahmins, although they hailed not from Ar-

cot, but from Eluru in the present West Godavari District of Andhra Pradesh (Ramachandra Rao
2003). Nonetheless, they did have familial connections with Arcot—Lakshmaiah at one point men-
tions visiting his relative Seeteramiah who lived there—which suggests the likelihood of their hav-
ing been instrumental in recruiting additional Arcot Niyogis, whether from Madras city or Arcot
town itself, to serve their employer. (See British Library, OIOC Eur. mss., Mackenzie Translations,
Class XII, no. 9, “Report of C. V. Letchmiah to Major C. Mackenzie Esq. . . . for the month of June
1804,” entries for 13 and 14 June.)
23 See Ramaswami 1984:317–22.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 797

eight (or 64 percent) are identifiable as Niyogis on the basis of their names. The
two next largest groups—Muslims and Kayasthas (or analogous immigrant
writing castes from northern India)—fall far behind, at only 10 percent and 5
percent respectively. The records of this case are also invaluable for the cir-
cumstantial details they record relating to matters of language use at the Arcot
court in general, and in particular, in the revenue administration of the state.
The general picture that emerges is that Persian and English were used as the
two main mediums for written correspondence, and that Persian and Marathi
were used as the principal languages for keeping accounts. Telugu and Tamil
were evidently the two primary languages used for in-group communication
among the Niyogi bureaucrats, as well as for out-group communication at the
local level. The accuracy of this picture is vouchsafed by the fact that it is large-
ly corroborated by the testimony of contemporary Persian histories of the Ar-
cot court, such as Muh!ammad Karı̄m’s Sawānih! āt-i Mumtāz (translated in
Nainar 1940 and 1944).24

narrain row and his collaboration with mackenzie

Narrain Row Brahmin produced more than one hundred letters, reports, and in-
terpretive documents during his fifteen years with Mackenzie (1803–1818),
and to judge from the vivid picture that emerges from them, he must have been
one of the more impressive of the Niyogi recruits. Narrain Row explicitly iden-
tifies himself as a native of Arcot, and circumstantial evidence overwhelming-
ly suggests the likelihood of his having previously been employed in the
Nawab’s administration.25 In any event, by the time he joined Mackenzie, he
had mastered all of the languages used in the Arcot administration—and then
some—with documented proficiency in Telugu, Marathi, Kannada, Tamil, San-
skrit, Hindustani, Persian, and English. Thanks to the survival of his letters and
reports, it is possible to reconstruct his peregrinations across the length and
breadth of the Deccan, month by month, and often day by day, from Tirupati
and Mysore in the south to Pune and Ahmadnagar in the Maratha dominions to
the north. In the course of these travels, Narrain Row collected scores of me-
dieval historical texts in Telugu, Kannada, Marathi, Sanskrit, and Persian, wrote
many original historiographic works, including an innovative “statistical”
analysis of Ahmadnagar district and the first coherent antiquarian descriptions

24 I have provided a more detailed discussion of these distinct bodies of evidence and their bear-

ing on the social and linguistic composition of the Arcot court as a chapter in a larger study, still in
progress, of Narrain Row’s collaboration with Mackenzie. The book’s working title is Mapping the
Past: The Career of Narrain Row Bramin and Colin Mackenzie’s Survey of South India, 1803–
1818.
25 One of the most important sources of information on Narrain Row’s personal background is

his account of his audience with Sitaram Bhupala Raja at Gadwal, a petty king and mansabdar of
the Nizam of Hyderabad. This account is contained in Narrain Row’s letter to Cavelly Venkata Laksh-
maiah, written from Corucondah and dated 5 October 1809 (British Library, OIOC Eur. mss.,
Mackenzie Translations, Class XII, no. 14[5])

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 798 phillip b. wagoner

of such sites as Ellora and Warangal, and collected hundreds upon hundreds of
inscriptions, mostly in Telugu, Kannada, Sanskrit, and Tamil.26
Just what would an Arcot Niyogi like Narrain Row have brought to Macken-
zie’s project? There were, I believe, many factors, not the least of which was a
professional orientation and work ethic which were strikingly congruent with
Mackenzie’s. Because of their years of experience in the revenue administra-
tion at Arcot, Narrain Row and his Niyogi colleagues would have been perfectly
comfortable working within a tightly organized bureaucratic hierarchy. So, too,
was Mackenzie, for whom years of military service and official government ap-
pointments made this the most natural model for organizing his staff. Indeed,
there was an almost uncanny convergence between precolonial and colonial ad-
ministrative structures, as Mackenzie’s organization closely replicated the of-
fices and hierarchies established at the Arcot court. Thus, Mackenzie’s Niyogi
staff members are referred to in the contemporary records not as “native infor-
mants”—as they have been conceived in our own time through what is essen-
tially a backward projection of a modern paradigm for anthropological field re-
search27 —but rather, as “writers” (munshı̄s), “accountants” (mutas! addı̄s), and
“administrative assistants” (gumāshtas), using the very same Persian adminis-
trative titles that these men would have borne while employed under the
Nawab.28 There are other important areas in which this convergence of outlook

26 The statistical account of Ahmadnagar is titled “Particular Account of the Agriculture, Pro-

duction, Manufactures, Commerce, Imports and Exports, with the Coins, Weights and Measures
and also an Account of the Soil, Seasons, Rains, Animals, etc. etc. of the Tallook of Ahmednuggur
in the Deckan, from Enquiries on the Spot from Intelligent Natives, in 1806” (British Library, OIOC
Eur. mss., Mackenzie General, XIV, no. 7). This work appears to have been jointly produced by
Narrain Row and his collaborator Ananda Row, with whom he traveled and worked in the Maratha
country during this period. The antiquarian description of Ellora was likewise co-authored with
Ananda Row, and is titled “Memoir Descriptive of the Ancient Place of Ellola [sic] near
Dowlatabad—compiled from the Mahatyams of that Ancient Stullum and an inspection of the Sev-
eral Ancient Monuments existing there by Narrain Row and Anunda Row in 1806” (British Library,
OIOC Eur. mss., Mackenzie General, XIV, no. 8). The antiquarian account of Warangal is contained
in Narrain Row’s letter to Cavelly Venkata Lakshmaiah, written from “Urungole” [! Warangal]
and dated 1 August 1816 (British Library, OIOC Eur. mss., Mackenzie Translations, Class XII, no.
49), and again in a fuller version in his “Report of Narrain Row from April to August 1816” (British
Library, OIOC Eur. mss., Mackenzie General, XIV, no. 47; see especially folios 53–59, and 69 –
74). A detailed account of Narrain Row’s travels and scholarly activities will be included in my
forthcoming Mapping the Past (see note 24 above).
27 See, for example, the title of Dirks’ 1993 essay, “Colonial Histories and Native Informants:

Biography of an Archive.”
28 A munshı̄ was by definition a professional writer—someone who might produce correspon-

dence, write, copy, or translate documents, or produce reports. In contrast, a mutas! addı̄ was tech-
nically an accountant—someone who had not only the necessary literate skills, but also and more
importantly, the advanced skills of numeracy needed to keep accounts and other records relating to
revenue, money, and finance. Although the two offices were thus technically differentiated, it ap-
pears that there was often a considerable degree of overlap between the munshı̄ and the mutas! addı̄.
As it was explained to the court by one of the witnesses in the case of Avadanum Paupia Braminy
et al. versus Reddy Row and Anunda Row, “The business of Moonshee is to write, that of Moota-
suddie to keep accounts; the two offices are often discharged by the same person” (PREIA IV
1809:55; the witness in question is named “Narrain Row,” but this is clearly not the same Narrain

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 799

can be detected. Both Mackenzie and his Niyogis accepted the proposition that
the production of knowledge meant the generation of “statistical” information
for the use of the state, whether colonial or pre-colonial, and both accepted that
such knowledge was most effectively produced through the modality of field-
survey, whether that be devoted to the production of maps (in the case of
Mackenzie) or to the collection of revenue information (in the case of his Niyo-
gis). Most importantly, both accepted that an investigation of the past was an
important component of this survey activity, and that present-day economic
arrangements could not be properly understood without simultaneously con-
sidering the historical conditions of their production. For this reason, both
Mackenzie and his Niyogis ascribed a high evidentiary value to donative in-
scriptions, in that the vast majority of such inscriptions documented the terms
of religious gifts involving the right to collect revenue from agricultural and
other taxes.
So much for the existence of shared bodies of experience and orientation that
made the collaboration such a natural and productive one. But what about those
unique qualities and skills that men like Narrain Row brought to the collabora-
tion, skills and abilities that were not otherwise available to Mackenzie, and
which would thus have constituted their most important contribution to the de-
veloping discipline of epigraphy?
Perhaps the single most important factor relates to the heightened critical
awareness of language and writing that had emerged among these Niyogis in
the course of performing their duties in the Arcot bureaucracy. In this connec-
tion, it is important to stress that the Niyogi approach to language differed de-
cisively from that of the more tradition-minded brahmins whose interests in
language centered on Sanskrit and its discipline of vyākaran! a, or “linguistic
analysis.” Although the Sanskrit vyākaran! a tradition did produce systematic
and sophisticated theories of language, we must nonetheless recognize that
there were limitations to this kind of work, in that it concentrated primarily on
the synchronic, structural aspects of the Sanskrit language to the exclusion of
its diachronic or pragmatic aspects. The linguistic interests that developed
among Niyogis could hardly have been more different from those of the
vyākaran! a tradition. On the one hand, they had developed a more sophisticat-
ed ability to think of language in comparative terms, apparently as a result of
their simultaneous professional involvement with so many linguistically dis-
tinct languages. In their professional administrative duties, they would have
been regularly called upon to translate documents from one language to anoth-
er, to make judgments about the commensurability or incommensurability of
certain terms and concepts, and ultimately, to transcend these specific lan-

Row who worked with Mackenzie). The term gumāshta designated a personal “assistant,” “agent,”
or “secretary” to a higher ranking munshı̄ or mutas! addı̄. It is interesting to note that as a nominal
derivative from the Persian verb gumāshtan, “to assign, to appoint,” the word is thus equivalent to
the Sanskrit niyogi, “one who is appointed.”

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 800 phillip b. wagoner

guages and the cultures they embodied, to arrive at an overarching theoretical

understanding of language as a universal category of human experience, of
which Sanskrit—contra the claims of the vyākaran! a tradition29 —was but one
of many equally valid manifestations. On the other hand, they were also far
more concerned with the particulars and variations of individual speech acts
(and, by extension, what we might think of as “writing acts”) because of their
daily involvement with administrative records and documents that could all too
readily be forged and falsified. It was thus important for Niyogis to develop a
heightened critical sense for the historically contingent and time-bound aspects
of language and writing, from the distinctive graphological features that reveal
an individual’s unique hand, to the specific lexical and syntactic choices that
might serve to identify a document as the product of a particular chronological
moment. In effect, what they developed was an implicit theory of secondary
signification, according to which these types of unconscious formal variation,
which were not directly linked to the expression of a statement’s intended
meaning, could be variably construed as signifiers of the authorship of a writ-
ten document, or of the chronological moment of its production.
Niyogis did not, on the whole, give formal, theoretical expression to these
views on language, but if we consider carefully the implications of their docu-
mented professional practice, we must nonetheless infer their existence as an
implicit set of assumptions. As for the graphological ability to identify the au-
thor of a given document simply on the basis of its handwriting, the records of
the 1808 court case against Reddy Row and Anunda Row clearly imply that this
was a widespread and highly developed skill among the Arcot Niyogis. Read-
ing these records, we learn, for example, that the defendant Anunda Row was
repeatedly asked over the course of his deposition to judge the identity of the
persons who had written certain documents, solely on the basis of their hand-
writing:
[Q:] In whose hand-writing are the words “inspected,” “examined,” on the back of the
sowal?
[Anunda Row:] One of them I know to be in the handwriting of Reddy Row, and the
other I believe to be the hand-writing of Roy Sreenevas Row.
...
[Q:] Do you know in whose hand-writing the swad on the account current and sowal is?
[Anunda Row:] I know it to be the swad of the Nabob Omdut ul Omrah.
...
[Q:] Did you ever see the Nabob write?
[Anunda Row:] Never.

29 The traditional position had argued that only Sanskrit bore the powers of primary and sec-

ondary signification, thereby qualifying it as the sole communicative medium. If words in vernac-
ular languages also appeared to function in this way, it was held to be either through the “memo-
ry” of the original Sanskrit words from which they were believed to be derived, or through a mere
“illusion” of primary signification. See Pollock 2001:26–30.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 801

[Q:] Then how do you know these characters to be the hand-writing of the Nabob?
[Anunda Row:] I have seen his handwriting frequently, and from the resemblance these
characters bear to it, I know them to be his hand-writing (PREIA III 1811:294–95).

Given this heightened graphological awareness on the part of the munshı̄s and
mutas! addı̄s of Arcot, it was but a short step to the development of a practical
method for the collection, decipherment, and classification of the various earli-
er forms of scripts used in inscriptions. Anunda Row’s testimony points to his
reliance on an abstract notion of individual graphological norms—“the hand-
writing of Reddy Row,” “the handwriting of Roy Sreenevasa Row,” “the hand-
writing of the Nabob”—and his use of a comparative formal method to identi-
fy any given sample as a particular manifestation of that norm—“I have seen
his hand-writing frequently, and from the resemblance these characters bear to
it, I know them to be his handwriting.” Taking the same kind of sensitivity to
subtle formal variations that is implied in such a critical operation, and apply-
ing it to the act of accurately transcribing a textual specimen so as to reproduce
these significant formal details, stands as one important step toward the formu-
lation of a paleographic methodology. A second crucial step consists in the use
of the formal comparative method, as a means to conclude that the initially un-
recognizable letters in an ancient text in fact represent archaic forms of the same
letters as they would be written in a familiar modern script, simply by virtue of
“the resemblance” they bear to their modern forms. It is for this reason that
Mackenzie’s Niyogis tended to work backwards in their paleographic deci-
pherments, starting with more closely related forms of script from the later me-
dieval period, and then moving progressively farther back to the more distant
early scripts. Thus, Narrain Row and his colleagues had no problems reading
inscriptions in the Hala-Kannada script dating back to as early as the eleventh
century, but the still earlier seventh-century forms of this script—which they
called “Poorwad-Halla-Canara” ( pūrva-hāla-kannad! a), or “earlier old Kanna-
da,” implicitly revealing that they recognized it to be an earlier genealogical
form of Hala-Kannada30 —seems to have remained undeciphered by any of
Mackenzie’s collaborators. But the most important point to recognize here is
that they nonetheless proceeded to make faithful copies of these undeciphered
inscriptions, and that if the hand-copies they thus produced are readable today
by someone informed by the many further advances in paleographic under-
standing made after the time of these pioneering Niyogis, it is precisely because
of the heightened sensitivity to distinctive formal variations that they had de-
veloped in connection with their earlier administrative work at the Arcot court.

30
See “specimen number 3” of the facsimiles in the “recension A” copy of Mackenzie’s Regis-
ter of a Collection of Ancient Sassanums or Grants and Inscriptions on Public Monuments . . .
(British Library, OIOC Eur. mss., Mackenzie General, vol. XVIII). The facsimile of this copper
plate from Nagamangalam District is labeled as being in the character called “The Poorwad-Halla-
Canara yet undecyphered October 1806.”

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 802 phillip b. wagoner

The Niyogi’s critical awareness of language was not just restricted to this
graphological level, however. Of still greater significance is the fact that these
men possessed an awareness of the chronological implications of recurrent styl-
istic formulae and other distinctive lexical usages. Let us consider for a moment
the deposition of another witness in the case of Avadanum Papiah vs. Reddy
Row and Anunda Row. Here, the witness Narsinga Row is being questioned
about his abilities to pass judgment on the authenticity of bonds issued by the
Nawab’s treasury department. Although he initially appeals to his knowledge
of the various hands in which they are written, he goes on to describe his use
of a similar method of formally based criticism applied to the syntactic and lex-
ical forms of the text itself:
[Q:] Have you any information which can guide you, in forming a judgment as to the
authenticity or spuriousness of any particular Bonds out-standing against the Nabobs
Wallajah and Omdut ul Omrah?
[Narsinga Rao:] [My opinion on the genuineness of bonds] has been founded principal-
ly upon a knowledge of the hand-writing of the persons in the Nabob’s service, who were
generally employed to write Bonds; and I sometimes judged of the spuriousness of the
Bonds from the tenor of them, which was different from that in which true Bonds were
usually written.

And further:
The Bond in the name of Gopaul Row being shown to the Witness, he is asked, Whether
he has any reason to know whether it is a genuine Bond of the Nabob Omdut ul Omrah
or not?
[Narsinga Rao:] This Bond is not in any hand-writing which I know, and the Bonds
which I have recognized as genuine have never contained the words, “benaberi zuroorut
karikhood,” or, in English, “on account of the necessity of my affairs,” and the expres-
sion at the bottom, “een tumusook wapis girifte khahud shood,” or, in English, “this bond
will be taken back,” are not usual in the Bonds which I have considered genuine (PREIA
III 1911:308; my emphasis).31

The implications of Narsinga Row’s testimony are simple and clear enough,
but at the same time they hold a tremendous potential for the development of
an epigraphic methodology. The testimony implies not only that he recognizes
the time-bound nature of certain lexical and syntactic choices, and that he has
built up a mental catalogue of typical phrases and expressions occurring in the
formal language of a given historical context, but most importantly, that he can
also employ this knowledge to pass judgment on the chronological age—and
by extension, the authenticity—of a given document.
The potential of such a method for the dating of inscriptions is tremendous
indeed. One of the recurrent hurdles of epigraphic research in India is that so
many inscriptions are either undated or have lost their dates through abrasion

31 Modern transcription of the two Persian phrases quoted by Narsinga Rao would be “binā-

bar-i zarūrat-i kār-i khūd ” and “ı̄n tamassuk wāpas gerefteh khvāhad shūd,” respectively.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 803

and breakage. But this problem may readily be surmounted by extending this
critical method of authenticating recent administrative documents and apply-
ing it to the similar task of dating undated inscriptions from a more distant era.
Significantly, just two years after Narsinga Rao explained his critical proce-
dures in his court deposition, we have evidence of Mackenzie’s collaborator
Narrain Row making use of precisely the same methodology in his critical
analysis of some of the inscriptions he had collected. In his Śrı̄śaila Devālaya
Kaiphiyatu, or “Particular Account of the Srisailam Temple” written in 1810,
Narrain Row twice appeals to the concept of “vakkan! a” as grounds for dating
an otherwise undated inscription. Although the Telugu word “vakkan! a” is more
often used in the sense of the formulaic greetings at the opening of a letter, he
uses it here in an extended sense to refer to the string of titles or birudas that
are applied to a ruler, and which occupy a comparable position at the opening
of an inscription. The very fact that he uses the word “vakkan! a,” however, is
strongly suggestive of a process of adaptation, as concepts and practices that
had originally been developed in the context of court administration are being
transferred to the new historiographic task. In the first passage, Narrain Row
describes a damaged inscription in which the name of the issuing king is ef-
faced, but which he ascribes to the fourteenth-century ruler Vemaya Reddi sim-
ply on the basis of the “vakkan! a” used, which he already recognizes from his
familiarity with other inscriptions issued by this ruler:
Next, there is a slightly damaged stone inscription with the vakkan! a of Vemaya Reddi,
lord of the lion throne of Addanki. This inscription is on a pillar in the mandapa hall in
front of the eastern doorway to the sanctum of the Srisailam temple. [There follows a
verbatim transcription of the string of titles constituting the “vakkan! a” of the grant.]
Because of the damage to this inscription, the donor’s name and the Śaka date are not
preserved. However, the Śaka date 1283 and cyclic year Plavanga are mentioned in oth-
er inscriptions of Vemaya Reddi, which have precisely the same vakkan! a (Śrı̄śaila
Devālaya Kaiphiyatu, 12; my translation).

Narrain Row here is clearly concluding that this inscription must also belong
to Vemaya Reddi, and that it would thus most likely have been issued at some
point around Śaka 1283.
In a second instance of his appeal to the concept of vakkan! a, Narrain Row
follows a similar chain of logic in ascribing an inscription to the reign of the
Vijayanagara ruler Krishnadevaraya. In this case, the epigraph records a series
of donations made to the Srisailam temple by one Koti Chinnisetti, but the name
of the reigning king has been effaced: “The place where Krishnaraya’s name
would have appeared at the top of the inscription has been broken off. Howev-
er, the vakkan! a is entirely that of Krishnaraya, and moreover, the Śaka date that
is given falls within the reign of Krishnaraya, so there can be no doubt [that this
inscription can be considered as pertaining to the reign of Krishnaraya]”
(Śrı̄śaila Devālaya Kaiphiyatu, 14; my translation).
From our present perspective, in which the history of Vijayanagara’s ex-

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 804 phillip b. wagoner

panding territorial holdings is so well known, it might appear that the mere Śaka
date alone should have been sufficient for Narrain Row to conclude that the in-
scription pertained to Krishnaraya. But we must remember that in 1810, the his-
torical geography of India was still in its infancy, and Mackenzie’s collabora-
tors were in fact pioneers in the process of plotting premodern polities along
the dual axes of time and space, recognizing that the expansion and contraction
of territories could be measured through epigraphic means. In approaching this
inscription, Narrain Row already knew of both the Vijayanagara state and the
contemporaneous Gajapati kingdom in coastal Andhra, but would not yet have
known where the boundary between these two states was drawn in Śaka 1434
(1512 a.d.). By reasoning in terms of the vakkan! a, which he could easily rec-
ognize as Krishnaraya’s, Narrain Row could infer not only that the inscription
would have been issued in that ruler’s name, but also, and more significantly,
that in 1512 at least, it was Vijayanagara that dominated the Srisailam country.

madras versus calcutta: the impact of the

mackenzie project on later epigraphy
To those familiar with the history of epigraphic studies in nineteenth-century
India, my emphasis on the historical significance of Mackenzie’s project may
appear to run counter to received characterizations of the project as essentially
a “beached whale.” Although Mackenzie’s epigraphic project is often assumed
to have died without begetting intellectual progeny, careful consideration of just
what it produced and how it was received suggests otherwise. In particular, the
project generated two primary products which appear to have exercised a sig-
nificant impact on the subsequent development of epigraphy in India. These
were the actual corpus of collected epigraphic texts, and the analytical “Regis-
ter” that served as an index to the texts and summarized the historical data they
contained.
Although the completely unprecedented nature of the Mackenzie epigraph-
ic corpus has thus far escaped notice, it constituted the first attempt to produce
a comprehensive collection of Indian epigraphic texts through a systematical-
ly organized survey. None of the early Calcutta epigraphers had ever attempt-
ed to carry out systematic collections; rather, they remained content to analyze
individual inscriptions which were, in Colebrooke’s own words, “occasionally
discovered through various accidents” (1807:398). While each such acciden-
tally discovered epigraph of course contributed in some small measure to the
understanding of India’s past, ultimately, real progress would come only through
the systematic assembling of comprehensive collections of inscriptions. All too
often, the potential of a given inscription as a historical document remains un-
realized, until one is able to relate its contents to the testimony of other records
pertaining to the same time and place—just as we have seen in the case of Nar-
rain Row’s success in dating the undated inscriptions he found at Srisailam,
thanks to his ability to relate them to other inscriptions already in the project’s
growing corpus.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 805

If Mackenzie was able to move beyond the Calcutta orientalists in conceiv-

ing of the idea of such a systematic undertaking, it was at least in part due to
his professional outlook as a cartographic surveyor, and to the fact that he had
been explicitly charged to carry out a “statistical survey” in tandem with his
map-making activities. But at the same time, that he was able to implement this
scheme so successfully depended just as much on his collaboration with his
Niyogi assistants, whose own professional outlook as revenue administrators
provided such a perfect complement to Mackenzie’s. By the time the project
concluded in about 1818, Mackenzie’s team had produced an epigraphic cor-
pus that was the largest ever produced in India, one that would not be surpassed
for many decades. Estimates of the collection’s size vary from “more than
7,500” to the precise figure of 8,076,32 but whichever figure one prefers to ac-
cept, it is clear that Mackenzie’s collection of inscriptions far exceeded any
made by his contemporaries.33
Given that Mackenzie never published this corpus, one may reasonably won-
der how much of an impact it could have exercised on subsequent scholarship.
Although the collection remained unpublished, Mackenzie himself corre-
sponded regularly with other epigraphers, often sharing the fruits of his proj-
ect’s labors. In this regard, it is noteworthy that he was in close contact with the
leading epigraphers of the Calcutta school, and that it was often Mackenzie
from whom the information flowed in these intellectual exchanges. Thus, he
appears to have been in correspondence with Henry Colebrooke from at least
as early as 1800, and indeed, a significant number of the inscriptions published
by Colebrooke in Asiatic Researches had been provided by Mackenzie from
among the materials he had collected. For example, among the inscriptions
translated and discussed by Colebrooke in his article “On Ancient Monuments,
containing Sanscrit Inscriptions,” published in Asiatic Researches in 1807,
there are five from the Mysore country which were provided to him by Macken-
zie.34 Mackenzie also corresponded with Charles Wilkins in London, promis-
ing in October 1808 to send Wilkins “the alphabet of the HallaCanara charac-

32 The first estimate is H. H. Wilson’s, from his catalogue of the Mackenzie collection (Wilson

1828). The second is Alexander Cunningham’s, writing in 1871 as the first Director General of the
newly instituted Archaeological Survey of India (Cunningham 1871:xxix).
33 C. P. Brown notes that of the other two collections he consulted in the compilation of his

Cyclic Tables (see discussion below), one included 595 inscriptions (the collection assembled by
Walter Elliot in the southern Maratha country), and the other “about three hundred” (the collection
made by “Captain Newbold, another zealous antiquarian” [Brown 1863:iv]). This makes Macken-
zie’s collection at least twelve times larger than either of the others that was available to Brown.
34 These included an original copper plate from Chitradurga, a facsimile of a second copper plate

from the same site, an original stone inscription from Kurugodu, and copies of two copper plates
from Nidigal and Goujda (Colebrooke 1807:398 –453; see nos. III, IV, and VIII). Colebrooke’s high
estimation of his colleague in Madras is revealed by his gracious acknowledgement of Mackenzie,
whom he characterized as “a gentleman, whose zeal for literary research, and indefatigable indus-
try in the prosecution of inquiries, cannot be too much praised” (1807:430).

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 806 phillip b. wagoner

ter” that his team had just deciphered.35 He also made his inscriptions available
to contemporaries involved in the study of India’s history and languages, in-
cluding his friend Mark Wilks, who used the “extensive and valuable collec-
tion” in Madras as source material for his Historical Sketches of the South of
India (1810),36 and A. D. Campbell, who appealed to the authority of inscrip-
tions in the collection to confirm the chronology of the Vijayanagara kings,
whose succession he summarized in the introduction to his A Grammar of the
Teloogoo Language (1820).37
Even after Mackenzie’s death, it is clear that the collection of inscriptions re-
mained accessible in Madras. One prominent scholar who used Mackenzie’s
epigraphic materials later in the nineteenth century was Charles Philip Brown,
who is remembered today primarily for his work on Telugu language and liter-
ature. In addition to his literary researches, however, Brown also produced an
important work on Indian chronology, published originally in 1850 as Cyclic
Tables of Hindu and Mahomedan Chronology, and revised in 1863 under the
new title of Carnatic Chronology, the Hindu and Mahomedan Methods of Reck-
oning Time Explained. In both versions of the work, Brown openly acknowl-
edges his indebtedness to the epigraphic materials collected by Mackenzie. In
the preface to Cyclic Tables, Brown explains that he first encountered Macken-
zie’s historical materials in the collection of the College of Madras, where they
had been placed when the Government had acquired them after Mackenzie’s
death. Although Brown’s first interests in Mackenzie’s collection were in the
more properly “literary” works and local records, he became encouraged to
study the inscriptions as well after reading H. H. Wilson’s 1828 catalogue of
the Mackenzie collection, and learning that the inscriptions were “more than
seven thousand five hundred in number” (Brown 1850[1994]:iii–iv). Brown
used these inscriptions as the primary sources for determining the regnal dates
of the various kings and dynasties that featured in his “Cyclic Tables” that

35 Mackenzie to Charles Wilkins, Esq., 25 Oct. 1808 (OIOC Mackenzie General collection, vol.

18 (recension A, bound in before p. 279).

36 In describing the various categories of sources he had employed, Wilks writes of “the exten-

sive and valuable collection of grants, generally of a religious nature, inscribed on stone or copper,
which are in the possession of my friend Lieutenant-Colonel Colin Mackenzie of the corps of en-
gineers on the establishment of Fort St. George. These ancient documents are of a singularly curi-
ous texture; they almost always fix the chronology, and frequently unfold the genealogy and mili-
tary history of the donor and his ancestors, with all that is remarkable in their civil institutions, or
religious reforms; and the facts derived from these inscriptions are illustrated by a voluminous col-
lection of manuscripts, which can only be trusted with confidence, so far as they are confirmed by
these authentic documents. . . . at the period of my departure from Madras, [this collection] amount-
ed to near one thousand seven hundred grants, and six hundred MSS” (Wilks 1810[1930], I:xxv).
37 Campbell writes, “I am indebted to the friendship of that able and distinguished officer

Colonel McKenzie C. B. of the Madras Engineers, now Surveyor General of India, for the follow-
ing translation of an extract from the Gutpurtee Manuscript in his valuable and extensive collec-
tion, containing, in the form of a prophecy, a chronological account of these kings. Numerous in-
scriptions, and grants of land, in the possession of Colonel McKenzie confirm the correctness of
this account” (1820:xi, my emphasis).

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 807

formed the core of the book. In the later version of this work, Brown further un-
derscores his dependence upon Mackenzie—and seemingly expresses admira-
tion for his predecessor—by appending a “Memoir of Lieutenant-Colonel
Colin Mackenzie,” adapted from the sketch given in the preface to Wilson’s cat-
alogue (Brown 1863:88).38
In terms of the institutional history of Indian epigraphy, it is especially im-
portant to recognize that Alexander Cunningham was familiar with Macken-
zie’s work, for it was under Cunningham’s direction that the Archaeological
Survey of India was established, dedicated to the systematic exploration of In-
dia to locate, document, and preserve the material vestiges of the country’s an-
cient past—including, of course, inscriptions. Although Cunningham mentions
Mackenzie only in passing—referring to “Mackenzie’s great collection of
8,076 inscriptions” in the context of his 1871 summary of earlier studies of In-
dian antiquities—the very organization and conception of the archaeological
department he founded suggests an inspiration in Mackenzie’s earlier Survey
of Mysore. Thus, in his 1861 memorandum proposing the establishment of the
Survey, Cunningham lamented that most earlier investigations of India’s past
had been “due to the unaided efforts of private individuals,” and that as a result
they had always been “desultory and unconnected and frequently incomplete”
(1871:iv). What was now needed, opined Cunningham, was an officially spon-
sored government undertaking, as scientific and systematic as the Trigonomet-
rical Survey of India (the great imperial cartographic enterprise launched at
mid-century under the auspices of the Survey of India), which would not only
concern itself with the surveying of sites and documentation of buildings and
monuments, but would also extend to the systematic collection of inscriptions
(through the preparation of “careful facsimiles”), coins, and even local tradi-
tions preserved orally at each site (1871:iv, viii). The parallel with Mackenzie’s
earlier project is striking, both in Cunningham’s emphasis on a systematic mode
of collection organized through what is essentially a cartographic paradigm
(thus, the Archaeological Survey of India), and in his emphasis on inscriptions
as one class of material evidence (together with coins, buildings, and other
monuments), which “in the almost total absence of any written history, form
the only reliable sources of information as to the early condition of the coun-
try” (1871:iii). Although Mackenzie had been dead for fifty years by the time
Cunningham was writing, the value of his systematic approach to the collec-
tion of inscriptions and other antiquities was now being affirmed in the estab-
lishment of the Archaeological Survey of India (ASI). Within a few years, by
1883, the ASI would establish a special epigraphic branch dedicated to the sys-
tematic survey and collection of inscriptions, the immediate forerunner of to-
day’s Chief Epigraphist’s Office in Mysore. That the office of this direct de-

38 I am indebted to Peter Schmitthenner for calling my attention to Brown’s reliance upon

Mackenzie’s materials. See Schmitthenner 2001.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 808 phillip b. wagoner

scendent of Mackenzie’s survey should be located in the capital of the state that
had been the focus of his statistical survey, is a fitting—if unintentional—trib-
ute to the importance of Mackenzie’s collaborative project to the development
of epigraphy as a systematic discipline.
The second significant product of the Mackenzie project was equally un-
precedented, and took the form of a manuscript volume titled “Register of a
Collection of Ancient Sassanums or Grants and Inscriptions on Public Monu-
ments . . .” and dated 31 December 1807. Although it was never published, the
“Register” is preserved in two slightly different manuscript copies, both housed
in the Oriental and India Office Collections of the British Library.39 In this
work, Mackenzie and his collaborators compiled a “register” for the some 1100
inscriptions that they had collected up to that point. In tabular form, the regis-
ter summarizes the attributes and contents of each inscription under the head-
ings “Date when Received,” “Granters Names and Reigning Sovereigns,”
“Place Where They are Preserved,” “Language,” “Character,” “Date of the In-
scription (Cyclic Year; Sakum era date),” “To whom Granted,” “The subjects
and contents of the grants and inscriptions” and “Remarks and Observations.”
The register is preceded by an index to the inscriptions arranged alphabetical-
ly by the name of the reigning king (this is present only in the “recension B”
copy), and is followed by an appendix, containing “copies of Facsimiles, Trans-
lations, and Drawings of some of the Inscriptions registered” (the “recension
A” copy has a complete set of facsimiles; while “recension B” includes only
one). In his introduction to the “Register,” Mackenzie sketches out a typology
of inscriptions, distinguishing between “Sassanums,” “Danapatrum,” and “Nero-
opum,” by virtue of such factors as contents and purpose, and also with refer-
ence to the materials typically used, whether stone or copper plates. Detailed
observations are also made on the general formats and conventions employed
in the various types of inscriptions, as well as on the identities of the languages
and characters used, and the dynasties who issued them.
The Mackenzie “Register” must be recognized as nothing less than the first
synthetic overview of the subject of Indian epigraphy, produced at a time when
the Calcutta epigraphers were still focusing their efforts on the decipherment
of earlier scripts and the translation and interpretation of individual epigraphic
records. Compared to anything his Calcutta contemporaries had yet produced,
the Mackenzie “Register” is not only vastly more detailed, but also far more
“emic” in its awareness of the cultural purposes originally served by inscrip-
tions. Although the specific contents of Mackenzie’s overview would soon
enough be rendered obsolete as epigraphic study advanced still further (com-
pare, for example, the far more detailed account in A. C. Burnell’s Elements of
South-Indian Palaeography, 1878), it is interesting to note that the format it
adopted lives on in countless epigraphic publications even today. With its tab-

39
British Library, OIOC Eur. mss., Mackenzie General, vol. XVIII (recensions A and B).

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 809

ular analysis of the contents and attributes of multiple inscriptions, with its in-
dex by reigning king, and with its graphic reproduction of selected facsimiles,
it must be recognized as the direct precursor of such epigraphic publications as
the Annual Reports on Indian Epigraphy and similar tabular compilations.

conclusion
Clearly, the continued preservation and accessibility of both the epigraphic cor-
pus and the analytical “Register” must be kept in mind as we attempt to under-
stand the Mackenzie project and assess the nature of its impact upon the sub-
sequent development of the epigraphic discipline in the later nineteenth century.
But in the end, we must also remember that the products of this dialogue were
not limited simply to collections, archives, and texts, but extended as well to
the new ways of thinking that had arisen in the minds of all who had partici-
pated in this fruitful intellectual conversation. Among the twenty-some Niyo-
gis who had helped formulate the new epigraphic methods, some, at least, con-
tinued to practice the new forms of epigraphic and historical knowledge even
after the close of Mackenzie’s project in 1818. Some of them continued to seek
opportunities for employment within the colonial government; thus, Cavelly
Venkata Lakshmiah went on to assist H. H. Wilson in producing the catalogue
of the various Mackenzie collections, and when Wilson succeeded in deci-
phering the ancient central Indian “box-headed script” dating back to the fourth
century, it was with the help of Sri Verma Suri, who had been one of Macken-
zie’s Jain assistants.40 Others, however, went on to practice the new forms of
colonial knowledge themselves, without depending on the direct patronage of
the colonial state. Here, the best documented case is provided by the subsequent
career of Cavelly Venkata Ramaswamy, who put many of the project’s materi-
als—including the work produced by Narrain Row—to good use in writing his
Descriptive and Historical Sketches of Cities and Places in the Dekkan, which
he published in Calcutta by subscription in 1828. Interestingly enough, Ra-
maswamy’s entry on Warangal (“Ekasilanagar”) is based almost entirely on
Narrain Row’s unpublished letters and reports from the Mackenzie archive, and
on some of the materials he had collected in Warangal in 1816.41
As for Narrain Row and the other lesser-known assistants, what they did af-
ter the close of the collaboration with Mackenzie remains shrouded in dark-
ness—at least in the present state of our knowledge. Perhaps a more thorough
examination of the colonial archive may eventually shed some light on their lat-
er activities, as is suggested by the following note of Walter Elliott, explaining
the provenance of a historical manuscript that had come into his collection in

40 This decipherment was published in Asiatic Researches 15 (1825): 499–515, according to

Salomon (1998:203).
41 See Ramaswami 1828:13 and compare with Narrain Row’s Report for the period April to Au-

gust 1816 (OIOC Mackenzie Translations Class XII: no. 47) and letter to C. V. Lechmyah Bramin,
dated Urungole 1 Aug. 1816 (OIOC Mackenzie Translations Class XII: no. 49).

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 810 phillip b. wagoner

January of 1844: “This statement was sent to me by Mr. Dighton who rented
the Warangal Districts from the Nizam for several years. It was prepared by a
Bramin in his service who had been formerly employed by Col. Mackenzie.”42
Perhaps even more promisingly, an examination of later nineteenth-century lit-
erature in Telugu may some day yield works produced by former Mackenzie
Niyogis who opted not to work in English, but in the lingua franca of their own
social class. In any case, it is clear that out of the collaboration with Macken-
zie, a new class of Niyogis had emerged, small at first, but clearly capable of
replicating itself 43 and applying the new mode of colonial knowledge to its own
ends, long after the close of Mackenzie’s project.
Hopefully, this essay has demonstrated that Mackenzie’s Niyogi assistants
brought far more to the collaboration than has previously been suspected. Far
from being mere passive informants, they contributed actively to the produc-
tion of new epigraphic knowledge, and more fundamentally, even to the defi-
nition of epigraphy itself as a method for historical enquiry. The intellectual di-

42
OIOC MSS Eur. D327, inside, opposite p. 1, emphasis added; for the dating, see OIOC MSS
Eur. F. 46, which is a slightly revised translation of the same text.
43
One striking pattern that emerges from Narrain Row’s correspondence relates to his training
and “subcontracting” of other Niyogis who were not formally in Mackenzie’s employ. On more
than one occasion, Narrain Row found himself unable to complete his objectives on schedule,
whether due to illness and logistical difficulties, or simply because of the overabundance of mate-
rials encountered. His answer to this dilemma was to befriend local Niyogis who shared his gen-
eral literate skills and sensibilities, and to hire them as his personal assistants after training them in
the new methods of historical and epigraphic enquiry. Thus, on one occasion, between January and
March of 1811, we find Narrain Row referring in his reports to a local Cuddapah Niyogi named
Letchmen Row, at first in the capacity of a general factotum and go-between, and then as an active
deputy engaged in the making of paleographically accurate transcriptions of epigraphs when Nar-
rain Row was incapacitated due to fever. See Narrain Row’s report for the year 1811 (Mackenzie
Translations Class XII, no. 26). In his entry for 23 January, he notes that “as I am unable even to
get up, I told Letchmen Row to copy those inscriptions of the above mentioned place. He copied
according to the original ” [i.e., he made an accurate hand-copy reproducing their paleographic
form; my italics].
Later, in 1818, we learn that Narrain Row had at least two local Niyogis in his employ when he
was working in the Hyderabad territories; one of these, whom he refers to as “my Goomashta”
Venkat Row, had evidently proven to be such a quick study—Narrain Row notes that he is a “very
clever man”—that he was able to make some very significant collections of old Kannada inscrip-
tions and historical texts in the region around Gulbarga and Malkhed. For Narrain Row’s references
to Venkat Row, see Mackenzie Miscellaneous, 174 (68 [par. 4]; 69 [par. 5]; 74 [par. 2]; and 75 [par.
2]). The Madras office appears to have disapproved of Narrain Row’s earlier hiring of Letchmen
Row—C. Venkata Lakshmaih wrote to Narrain Row in March of 1811 “with an order to discharge
the said Letchmen Row” (Mackenzie Translations, class XII: 26 [entry for 8 Mar.]). But it eventu-
ally concurred with Narrain Row’s estimation of Venkat Row, and engaged him to continue mak-
ing collections even after the bulk of the collecting activity was completed and Mackenzie’s es-
tablishment had shifted to Calcutta. See the five letters of “Venkat Rao employed at Hydrabad”
(Translations, class XII, no. 16; note that this designation apparently serves to distinguish him from
a different Venkat Rao who worked further south in the Tamil country). Of particular interest is that
Narrain Row’s assistant Venkat Row in turn employed his own assistant, suggesting that the seg-
mentary mode favored by Niyogis for organizing their labor would likely have contributed still fur-
ther to the spread of the new forms of colonial knowledge among the members of this precolonial
intellectual class.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 811

alogue through which epigraphy emerged in colonial Madras was a particular-

ly productive one, as Mackenzie, with his cartographic and statistical approach
to the past, engaged in conversation with his Niyogi collaborators, with their
critical awareness of the diachronic and pragmatic aspects of language, and
their impressive abilities to digest and synthesize data from great masses of oth-
erwise discrete documents and records. Indeed, it was the vital quality of this
intellectual dialogue that made the Madras school of epigraphy both more pro-
gressive than that of its Calcutta counterpart, and more central to the subsequent
definition of the discipline in the later nineteenth century.

re f ere nc es

u n pu b l i sh e d so u r c e s
British Library. Oriental and India Office Collections. European Manuscripts. Macken-
zie Collections (General Collection; Miscellaneous Collection; and Translations,
Class XII, “Letters and Reports”).
pu b l i sh e d so u r ce s
Ali, Daud. 2000. “Royal Eulogy as World History: Rethinking Copper-plate Inscriptions
in Cola India.” In Ronald Inden et al., eds., Querying the Medieval: Texts and the His-
tory of Practices in South Asia. New York: Oxford University Press, 165–229.
Appadurai, Arjun. 1993. “Number in the Colonial Imagination.” In Carol A. Brecken-
ridge and Peter van der Veer, eds., Orientalism and the Postcolonial Predicament:
Perspectives on South Asia. Philadelphia: University of Pennsylvania Press, 314–39.
Bayly, C. A. 1996. Empire and Information: Intelligence Gathering and Social Com-
munication in India, 1780–1870. Cambridge: Cambridge University Press.
Brown, Charles Philip. 1850[1994]. Cyclic Tables of Hindu and Mahomedan Chronol-
ogy, Regarding the History of the Telugu and Kannadi Countries, to which are Added
the Genealogies of Particular Hindu Families, with Essays on Various Matters of En-
quiry. Madras: Christian Knowledge Society. [repr. New Delhi and Madras: Asian Ed-
ucational Services].
———. 1863. Carnatic Chronology, the Hindu and Mahomedan Methods of Reckon-
ing Time Explained: With Essays on the Systems, Symbols Used for Numerals, a New
Titular Method of Memory, Historical Records; and Other Subjects. London: Bernard
Quaritch.
Burnell, A. C. 1870. A Few Suggestions as to the Best Way of Making and Utilizing
Copies of Indian Inscriptions. Madras: Higginbotham and Co.
———. 1879[1874]. Elements of South-Indian Palaeography from the Fourth to the
Seventeenth Century A.D., Being an Introduction to the Study of South-Indian In-
scriptions and MSS, 2d ed. London: Trubner & Co.
Campbell, A. D. 1820. A Grammar of the Teloogoo Language, Commonly Termed the
Gentoo, Peculiar to the Hindoos Inhabiting the North Eastern Provinces of the Indi-
an Peninsula, 2d ed. Madras: College of Fort St. George.
Cavelly Boria and Colin Mackenzie. 1809[1807]. “Account of the Jains, Collected from
a Priest of this Sect; at Mudgeri.” Asiatic Researches 9:244–78.
Cohn, Bernard S. 1987. “The Census, Social Structure and Objectification in South
Asia.” In An Anthropologist among the Historians and Other Essays. Delhi: Oxford
University Press, 224– 54.
———. 1992. “The Transformation of Objects into Artifacts, Antiquities and Art in

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 812 phillip b. wagoner

Nineteenth-Century India.” In The Powers of Art: Patronage in Indian Culture. Del-

hi: Oxford University Press, 301–29.
———. 1996. Colonialism and Its Forms of Knowledge: The British in India. Prince-
ton: Princeton University Press.
Colebrooke, H. T. 1809[1807]. “On Ancient Monuments, Containing Sanscrit Inscrip-
tions.” Asiatic Researches 9:398– 453.
Cooper, Frederick, and Ann Laura Stoler. 1997. Tensions of Empire: Colonial Cultures
in a Bourgeois World. Berkeley: University of California Press.
Cotton, J. S., J.H.R.T. Charpentier, and E. H. Johnston. 1992. Oriental and India Office
Collections: Catalogue of Manuscripts in European Languages Belonging to the Li-
brary of the India Office, Volume I, Part II: The Mackenzie General and Miscella-
neous Collections. With an Introduction by David M. Blake. London: The British Li-
brary.
Cowell, E. B., and F. W. Thomas, trans. 1968. The Hars! a-Carita of Bān! a, 2d ed., by
arrangement with the Royal Asiatic Society, London. Delhi: Motilal Banarsidass.
Cunningham, Alexander. 1871. Archaeological Survey of India Report, Volume I: Four
Reports Made During the Years 1862–65. Simla: Government Central Branch Press.
———. 1879. The Inscriptions of Asoka. Corpus Inscriptionum Indicarum. Vol. I. Cal-
cutta: Office of the Superintendent of Government Printing.
Dirks, Nicholas. 1989. “The Invention of Caste: Civil Society in Colonial India.” Social
Analysis 25:42– 52.
———. 1993. “Colonial Histories and Native Informants: Biography of an Archive.”
In Carol A. Breckenridge and Peter van der Veer, eds., Orientalism and the Post-
colonial Predicament: Perspectives on South Asia. Philadelphia: University of Penn-
sylvania Press, 279– 313.
———. 2001. Castes of Mind: Colonialism and the Making of Modern India. Prince-
ton: Princeton University Press.
Eaton, Richard. 2000. “(Re)imag(in)ing Other2ness: A Postmortem for the Postmodern
in India.” Journal of World History 11, 1:57–78.
Edney, Matthew H. 1990. Mapping an Empire: The Geographical Construction of
British India, 1765–1843. Chicago: University of Chicago Press.
Fisher, Michael H. 1984. “Indirect Rule in the British Empire: The Foundations of the
Residency System in India (1764–1858).” Modern Asian Studies 18, 3:393–428.
Hatcher, Brian. 1996. Idioms of Improvement: Vidyāsāgar and Cultural Encounter in
Bengal. Calcutta: Oxford University Press.
Inden, Ronald. 1986. “Orientalist Constructions of India.” Modern Asian Studies 20,
3:401–46.
———. 1990. Imagining India. Oxford: Basil Blackwell.
Inden, Ronald, et al. 2000. Querying the Medieval: Texts and the History of Practices
in South Asia. New York: Oxford University Press.
Irschick, Eugene F. 1994. Dialogue and History: Constructing South India, 1795–1895.
Berkeley: University of California Press.
Kejariwal, O. P. 1988. The Asiatic Society of Bengal and the Discovery of India’s Past
1784–1838. Delhi: Oxford University Press.
Kopf, David. 1969. British Orientalism and the Bengal Renaissance: The Dynamics of
Indian Modernization, 1773–1835. Berkeley and Los Angeles: University of Cali-
fornia Press.
Lakshmana Rao, K. V. 1965[1923]. “Āndhra Brahmulalōni Niyōgi-Vaidiki-bhēda kāla
nirn! ayamu.” In Laks! manarāya Vyāsāvali. Vijayavada: Adarsa Granthamandali, 1–17.
Metcalf, Thomas R. 1994. Ideologies of the Raj. The New Cambridge History of India,
II.4. Cambridge: Cambridge University Press.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 intellectuals and the production of knowledge 813

Mitter, Partha. 1977. Much Maligned Monsters: History of European Reactions to In-
dian Art. Oxford: Clarendon Press.
Nainar, S.M.H. 1940 –1958. Sources of the History of the Nawwabs of the Carnatic.
Madras University Islamic Series, 5 vols. Madras: University of Madras.
[Nārāyan! a Rāvu] 1982[1810] Śrı̄śaila Devālaya Kaiphiyatu. P. Sitapati, ed. Andhra
Pradesh Government Epigraphical Series, no. 17. Hyderabad: Government Press.
Peabody, Norbert. 2001. “Cents, Sense, Census: Human Inventories in Late Precolonial
and Early Colonial India.” Comparative Studies in Society and History 43, 4:819 –50.
Phillimore, R. H. 1945. Historical Records of the Survey of India: Volume I: Eighteenth
Century. Dehra Dun: Survey of India.
———. 1950. Historical Records of the Survey of India: Volume II 1800 to 1815. Dehra
Dun: Survey of India.
Pinch, W. R. 1999. “Same Difference in India and Europe.” History and Theory 38,
3:389 –407.
Pollock, Sheldon. 2001. “New Intellectuals in Seventeenth-century India.” The Indian
Economic and Social History Review 38, 1:3–31.
PREIA III. 1811. Papers Relating to East India Affairs III. Nabob of Carnatic Debts—
Forgeries—Prosecutions. London: House of Commons, 3 May.
PREIA IV. 1809. Copies of two letters from Sir Thomas Strange, Knight, Chief Justice
of the Supreme Court of Judicature at Madras to the Right Honourable Robert Dun-
das; dated 27 Feb. and 4 May 1809, with enclosures. In Papers Relating to East In-
dia Affairs IV. Nabob of Carnatic Debts—Forgeries—Prosecutions. London: House
of Commons, 11 June 1811.
Ramachandra Rao, C. V. 2003. The Kavali Brothers, Col. Colin Mackenzie and the Re-
construction of South Indian History and Cultural Resurgence in South India. Nel-
lore: Manasa Publications.
Ramaswami, N. S. 1984. Political History of Carnatic under the Nawabs. Delhi: Abhi-
nav Publications.
Ramaswami, Cavelly Venkata. 1828. Descriptive and Historical Sketches of Cities and
Places in the Dekkan; To which is prefixed an Introduction, Containing a Brief De-
scription of the Southern Peninsula, and a Succinct History of Its Ancient Rulers: The
Whole Being Intended to Serve as a Book of Reference to a Map of Ancient Dekkan.
Calcutta: Messrs. W. Thacker and Co. St. Andrew’s Library.
Rangasvami Sarasvati, A. 1926. “Political Maxims of the Emperor-Poet, Krishnadeva
Raya.” Journal of Indian History 4, 3:61– 88.
Rocher, Rosane. 1989. “The Career of Rādhākānta Tarkavāgı̄śa, an Eighteenth-century
Pandit in British Employ.” Journal of the American Oriental Society 109:27–33.
———. 1993. “British Orientalism in the Eighteenth Century: The Dialectics of Knowl-
edge and Government.” In Carol A. Breckenridge and Peter van der Veer, eds., Ori-
entalism and the Postcolonial Predicament: Perspectives on South Asia. Philadel-
phia: University of Pennsylvania Press, 215–49.
Said, Edward. 1978. Orientalism. New York: Routledge and Kegan Paul.
Salomon, Richard. 1998. Indian Epigraphy: A Guide to the Study of Inscriptions in San-
skrit, Prakrit, and the Other Indo-Aryan Languages. New York: Oxford University
Press.
Schmitthenner, Peter L. 2001. Telugu Resurgence: C. P. Brown and Cultural Consoli-
dation in Nineteenth-century South India. Delhi: Manohar.
Seneviratne, H. L. 1997. Identity, Consciousness and the Past: Forging of Caste and
Community in India and Sri Lanka. Delhi: Oxford University Press.
Sinclair, Sir John. 1798. History of the Origin and Progress of the Statistical Account of
Scotland. N.p.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /

 814 phillip b. wagoner

Singh, K. S. 1998. India’s Communities, N-Z. People of India National Series, vol. VI.
Delhi: Anthropological Survey of India and Oxford University Press.
Sitapati, P., trans. 1981. Srisailam Temple Kaifiyat, Andhra Pradesh. Vol. 1. Hyderabad:
Government of Andhra Pradesh.
Talbot, Cynthia. 2001. Precolonial India in Practice: Society, Region, and Identity in
Medieval Andhra. New York: Oxford University Press.
Tavakoli-Targhi, Mohamad. 2001. “Orientalism’s Genesis Amnesia”. In Refashioning
Iran: Orientalism, Occidentalism, and Historiography. Hampshire: Palgrave, 18–34.
Trautmann, Thomas. 1999a. “Hullabaloo about Telugu,” South Asia Research 19, 1:53–
70.
———. 1999b. “Inventing the History of South India.” In Daud Ali, ed., Invoking the
Past: The Uses of History in South Asia. London: Oxford University Press, 36–54.
Trautmann, Thomas, and Carla Sinopoli. 2002. “In the Beginning was the Word: Exca-
vating the Relations between History and Archaeology in South Asia.” Journal of the
Economic and Social History of the Orient 45, 3:1–32.
Viswanathan, Gauri. 1990. Masks of Conquest: Literary Study and British Rule in In-
dia. London: Faber and Faber.
Wilks, Lieutenant-Colonel Mark. 1810[1930]. Historical Sketches of the South of India,
in an Attempt to Trace the History of Mysore. 2 vols. Ed. with notes by Murray Ham-
mick. Mysore: Government Branch Press.
Wilson, H. H. 1828. Mackenzie Collection: A Descriptive Catalogue of the Oriental
Manuscripts and Other Articles . . . Collected by the Late Lieut.-Col. Colin Macken-
zie, Surveyor General of India. Vol. 1. Calcutta: Asiatic Press.

1::7 . 0 3 1/. 53 5/ 4 .0/ 5 / : /