Full chapter text
Social Categories
Colonial surveys created an extraordinarily dense archive for the social history of Mithila, Vajji and Anga,
but the archive must be read as a history of classification as well as a description of society. From Buchanan-
Hamilton’s early district inquiries to the censuses, Risley’s ethnography, Grierson’s linguistic survey and the
gazetteers, officials progressively standardized names, boundaries and comparative tables. The central method
of this chapter is chronological: a colonial category is evidence first for the date and administrative purpose in
which it appears. It cannot be projected backward without independent period evidence.
59.1 Colonial knowledge became an instrument of administration
Colonial rule generated an expanding demand to describe the people it taxed, policed, recruited, litigated
with and governed. The resulting surveys were never neutral inventories. Their headings reflected
administrative problems: landholding, occupation, religion, language, “tribe,” “caste,” custom and rank. Yet
it is equally misleading to treat every category as a pure colonial fabrication. Officials worked with locally
supplied names, inherited textual vocabularies, previous revenue records and the testimony of intermediaries.
The historical problem is therefore to reconstruct a process of translation. Social identities existed before the
census, but the survey and census increasingly required them to be written into standardized boxes,
compared across districts and made legible to provincial government. The documentary archive is invaluable
precisely because it records this encounter between lived social description and bureaucratic classification.
59.2 The Company survey preceded the decennial census by decades
Long before all-India decennial enumeration, the East India Company commissioned district surveys to
learn the economic and social character of its territories. The sequence associated with Francis Buchanan-
Hamilton in eastern India between 1807 and 1814 covered Purnea, Bhagalpur, Bihar and Patna, Shahabad
and neighbouring districts. His questionnaires ranged far beyond revenue. They recorded cultivation,
manufactures, markets, prices, diseases, religious establishments, learned specialists, languages and divisions
among the population. The form was encyclopaedic because the Company still lacked a settled statistical
apparatus. For Mithila, Vajji and Anga, these early reports are therefore a baseline for the early nineteenth
century, not a transparent survival of medieval society. Their strength is descriptive breadth; their weakness is
the mixture of direct observation, local testimony, estimation and authorial judgement.
59.3 Buchanan-Hamilton’s method joined observation to local testimony
Buchanan-Hamilton travelled, questioned officials and residents, inspected institutions and compiled
numerical returns, but he could not personally verify every statement in large districts. His accounts make
visible the people who supplied information even when they are not always named: revenue servants,
landholders, merchants, religious specialists, cultivators, artisans, physicians and interpreters. A social
category in his report should therefore be read as a negotiated entry produced through a chain of
communication. The same label might carry different meanings in different localities, while a local term
might be translated into an English administrative vocabulary that made it appear more uniform than it was.
This is why the reports are best used comparatively, alongside settlement papers, genealogical records,
vernacular materials and later gazetteers, rather than mined for isolated timeless statements about
communities.
605605
GAJENDRA THAKUR
59.4 Early surveys mixed social, occupational and religious classifications
The early district accounts did not yet possess the rigid appearance of later census tables. Descriptions
could move between occupation, lineage, sect, locality, status, religion and customary practice without
forcing each person into a single analytical axis. That fluidity is important evidence. It reminds us that later
census columns were not merely recording a pre-existing universal taxonomy. A cultivator might also be
described through a caste name, a tenancy relation, a sectarian affiliation and a regional label; a learned
specialist could be located simultaneously in a scholastic institution and a kin network. Historians should
preserve those overlaps. Reducing an early survey entry to one modern identity category can erase exactly the
ambiguity that makes the source useful for tracing how social classifications changed during the nineteenth
century.
59.5 Revenue knowledge and social knowledge were intertwined
Survey categories gained administrative weight because they were linked to fiscal and legal questions.
Officials wanted to know who held land, who cultivated it, what services were customary, which groups
migrated, which occupations supported particular markets and how local authority operated. Land and
revenue classification therefore shaped social description even when the report’s chapter title concerned “the
people.” In Permanent Settlement districts, the distinction among proprietors, tenure-holders, tenants and
labourers could be as consequential as caste terminology. Conversely, officials sometimes interpreted agrarian
relations through assumptions about community or hereditary occupation. The archive must be read in both
directions. Social labels could influence administrative reasoning, while administrative forms could make
some dimensions of social life more visible than others.
Figure 232 — From local description to administrative category: the colonial classification chain
59.6 Hunter’s Statistical Account made the district a comparative statistical unit
By the 1870s the government possessed a denser archive of revenue returns, censuses, police reports, road
and trade information and district correspondence. W. W. Hunter’s Statistical Account of Bengal assembled
this material into volumes organized by district. Tirhut and Champaran appeared in Volume XIII; Patna and
neighbouring districts were treated in other volumes. These accounts combined geography, population,
agriculture, landholding, trade, institutions and administrative statistics. Their importance lies less in any
single number than in the new comparative form: districts could now be tabulated against one another. The
method encouraged standard headings and recurring variables, making society increasingly visible as a set of
measurable attributes. At the same time, local complexity was compressed to fit the provincial statistical
template.
HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II
59.7 The 1872 enumeration was a major experiment rather than a fully
standardized modern census
The census taken across the Bengal Presidency in 1872 marked a new scale of population enumeration.
Its reports discussed the accuracy of returns and the problems of supervision, showing that the operation
itself was understood as experimental. For Bihar districts, surviving census publications and divisional
statements provide population counts and classifications that can be compared with Hunter’s statistical
compilations. But “1872” should not be treated as if it had the uniform procedures of a later national census.
Dates, forms and administrative arrangements varied across British India, and enumerators depended heavily
on district machinery. The value of the exercise is therefore double: it produced new quantitative evidence,
and it reveals the practical difficulties involved when government attempted to translate millions of
households into a centralized statistical language.
59.8 Enumeration was mediated by local clerks, household informants and
officials
A census schedule did not fill itself. Enumerators had to ask questions, interpret answers, choose
spellings, resolve synonyms and assign entries to printed headings. Household informants could answer
strategically or according to local usage; clerks might regularize unfamiliar names; district officers could
merge small categories or separate large ones when compiling tables. Social knowledge was thus co-produced,
although the state controlled the final classificatory architecture and publication. Recognizing local
participation does not remove the asymmetry of colonial power. It explains why the archive can contain both
indigenous vocabularies and bureaucratic distortions at the same time. The historian’s task is to reconstruct
the chain from answer to schedule to table instead of assuming that the published category was the only
identity available to the person enumerated.
59.9 The 1881 census regularized decennial enumeration
The 1881 census brought a more systematic and synchronized framework to British Indian population
statistics. The Bengal report included major sections on caste, birthplace and language, occupations,
instruction and illiteracy. Repeated census-taking created comparability across decades, but it also increased
pressure for stable definitions. A category that changed name or boundaries created a statistical discontinuity,
so officials increasingly documented classification rules. This administrative need could encourage the
appearance of sharper social boundaries than everyday life displayed. The gain for historical research is
substantial: recurring tables make it possible to study demographic structure, literacy, occupation and
migration. The cost is that the repeated form can tempt readers to mistake the state’s need for
commensurability for evidence that every social category was fixed and universally agreed.
59.10 A printed table made fluid answers appear mutually exclusive
Tabulation is an intellectual act. When thousands of written answers are aggregated, variant spellings and
local subdivisions must be assigned to headings, while ambiguous cases are resolved so that column totals add
correctly. The printed page then hides much of that editorial labour. This is especially important for caste,
tribe, language and occupation. A table can make a category look bounded because every row has a line and
every number occupies one cell. Yet the original social world may have contained overlapping affiliations,
status disputes and multilingual households. Statistical exactness at the final stage should therefore not be
confused with ontological exactness. The numbers may precisely reproduce the rules of compilation while
those rules themselves simplify social reality.
607607
GAJENDRA THAKUR
Figure 233 — Major stages in the nineteenth-century classificatory archive
59.11 Caste names in colonial records were situated claims, not medieval fossils
When a census or gazetteer records a caste name, the safest first conclusion is that the name was used or
recognized in that administrative setting at that date. More ambitious historical claims require earlier
evidence. Some designations had long histories; others changed scope, acquired new prestige, split into
subgroups or were standardized only in the nineteenth century. The colonial archive is exceptionally rich for
studying these transformations because officials repeatedly asked about rank, marriage, occupation and
custom. It becomes misleading only when its categories are projected backward without chronological
control. The present volume therefore treats a colonial caste label as evidence for colonial social history unless
inscriptions, dated genealogies, manuscripts, Panji material or other earlier records independently establish an
older chronology.
59.12 Occupation and caste were related but never interchangeable
Colonial officials often searched for correspondences between caste and hereditary occupation, partly
because such correspondences promised an intelligible social system. District evidence repeatedly
complicated that assumption. Cultivators could belong to many named communities; service groups could
combine agriculture with specialized work; merchants and artisans moved between occupations; landholding
altered status; migration created new labour combinations. Census occupational tables therefore answer
different questions from caste tables. They are strongest when used to reconstruct the distribution of work at
a particular time, not when used to infer an unchanging occupational essence for a community. For Mithila,
Vajji and Anga, this distinction is especially important in a nineteenth century transformed by indigo,
opium, railways, urban services, migration and expanding clerical employment.
59.13 Religious classification also required administrative boundaries
Religious enumeration appears simpler than caste classification only when later categories are taken for
granted. Census officers had to decide how to count sects, reform movements, local devotional identities and
people whose practice crossed neat textual boundaries. Gazetteers often described popular worship,
pilgrimage and local cults in ethnographic detail, while census tables reduced religion to a smaller set of
aggregate headings. Both source types are useful, but for different purposes. The table measures the
administrative population assigned to a category; the descriptive chapter records a selective account of
practice. Neither should be substituted for the other. Ritual participation, sectarian affiliation and census
religion were related dimensions of social life, not necessarily identical ones.
HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II
59.14 Language classification converted multilingual practice into named
languages
The same methodological problem appears in language statistics. Households could use one speech form
at home, another in markets, a different literary language and yet another language in school or
administration. Census forms nevertheless required a countable answer. Officials had to decide whether a
local name denoted a separate language, a dialect or a synonym. The resulting tables are essential for studying
the history of language naming, but they do not map speech mechanically. In Mithila, Vajji and Anga, later
distinctions among Maithili, Angika, Bajjika, Hindi/Hindustani and other administrative language labels
must be interpreted through the wording and classification rules of each source. Modern political or
constitutional language categories cannot be read backward into every colonial table.
59.15 Hunter’s accounts linked population categories to economy and
administration
One advantage of the Statistical Account over isolated census tables is that it places population categories
beside agriculture, markets, education, transport, land revenue and public institutions. This permits a
genuinely socio-economic reading. A group’s numerical presence can be examined alongside the district
economy in which people lived, without assuming that the group itself explains the economic pattern.
Hunter’s compilations also preserve the administrative state of knowledge in the 1870s: what officials
thought worth measuring, which earlier reports they trusted and how they summarized local conditions.
Used critically, the volumes are therefore sources for both regional history and the history of colonial
knowledge. Their authority must be disaggregated claim by claim rather than either accepted wholesale or
dismissed wholesale.
59.16 The district gazetteer became a durable administrative archive
Early twentieth-century district gazetteers consolidated decades of surveys, settlement reports, censuses
and departmental statistics into compact reference volumes. Darbhanga, Muzaffarpur, Monghyr and
Bhagalpur gazetteers contain chapters on physical geography, history, population, agriculture, industry,
communications, education, health and administration. They are especially valuable for tracing institutional
change because compilers frequently summarized earlier conditions before describing the present. Yet the
gazetteer is a retrospective synthesis, not a contemporary source for every period it mentions. Where it
recounts eighteenth- or early nineteenth-century society, the underlying source must be identified whenever
possible. For its own early twentieth-century administrative world, by contrast, it is primary evidence of what
the government officially knew and how it organized that knowledge.
59.17 Gazetteers intertwined ethnography with practical governance
The gazetteer’s social descriptions sat beside roads, ferries, schools, police stations and land systems
because colonial administration treated knowledge as practical infrastructure. A district officer needed a
portable account of terrain, population, customary institutions and economic resources. Ethnographic
paragraphs therefore acquired an official afterlife: later administrators could repeat classifications first
assembled by earlier surveyors. Repetition could harden a tentative observation into an apparent fact,
especially when citations were omitted. Historians should track this textual genealogy. If the same statement
appears in a 1907 gazetteer and an 1870s statistical account, it may represent one inherited observation rather
than two independent confirmations. Administrative continuity is itself historically important, but it must
not be mistaken for independent evidence.
609609
GAJENDRA THAKUR
59.18 Risley’s ethnographic programme sought a more systematic taxonomy
Herbert Hope Risley’s Tribes and Castes of Bengal, published in the early 1890s, represented a more
ambitious attempt to systematize social classification. The work combined an ethnographic glossary with
anthropometric data and drew on district information, questionnaires and local reports. Risley treated caste
not merely as a list of names but as a structure to be classified and compared. That programme deeply
influenced later discussions because it connected administrative enumeration to a theory of social hierarchy
and physical type. For modern historical use, its rich descriptive material must be separated from its
theoretical apparatus. Kin terms, customs and local names can preserve evidence unavailable elsewhere, while
racial and anthropometric interpretations belong to a now-discredited framework and cannot be treated as
biological explanations of social history.
59.19 Anthropometry attempted to turn social hierarchy into bodily
measurement
The anthropometric component of Risley’s work measured physical traits and sought correlations
between bodily indices, race and caste rank. Such reasoning reflects late nineteenth-century racial science
rather than a valid biological account of Indian social structure. Its historical significance lies in what the state
and learned institutions attempted to measure, the procedures they devised and the authority numerical
measurement claimed. The presence of a decimal or index gave a speculative racial theory the appearance of
precision. For this volume, anthropometric tables are therefore evidence about colonial knowledge-making,
not evidence that caste groups possessed fixed racial essences. This distinction also protects regional social
history from a false move in which administrative categories, social rank and biological ancestry are collapsed
into one supposedly objective hierarchy.
59.20 Social precedence made ranking itself a field of inquiry
Risley and census officials also experimented with arranging caste groups by “social precedence.” Such
ranking required criteria and local testimony: who accepted food or water from whom, which groups
claimed particular ritual status, how Brahmanical opinion was reported, and how competing claims were
reconciled. The exercise did not simply discover a single ladder waiting to be written down. Rank could vary
by locality, context and informant, and communities could contest official placement. The methodological
lesson is not that hierarchy was unreal; hierarchy powerfully shaped social relations. It is that the published
order was an administrative synthesis of multiple, sometimes conflicting status practices. The hierarchy in a
census report is therefore itself a historical event in the politics of classification.
HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II
Figure 234 — Classification as a feedback loop between society and the state
59.21 Local intermediaries were indispensable to ethnographic classification
District officers lacked the linguistic range, local genealogical knowledge and everyday access needed to
classify populations unaided. They relied on subordinate officials, schoolmen, pandits, maulvis, zamindari
staff, community spokesmen and other informants. These intermediaries brought their own intellectual
frameworks and social interests to the process. A classification could therefore contain genuine local
knowledge while also privileging the perspectives of literate and institutionally connected groups. The
archive rarely represents every social stratum equally. The most marginalized could be described by others
rather than speaking through their own texts. Recognizing this asymmetry helps explain why administrative
ethnography must be compared with petitions, court records, oral traditions and vernacular publications
when reconstructing self-description.
59.22 Communities could use the classificatory state strategically
Once census and ethnographic categories acquired public authority, communities had incentives to
influence them. Petitions, memorials, genealogies and claims to revised rank formed part of a wider
nineteenth- and early twentieth-century politics of representation. A community could emphasize a
preferred name, origin narrative, occupation or ritual practice in order to seek recognition. Such activity
should not be dismissed as mere fabrication. It shows that classification had consequences and that people
understood the state’s categories as a resource to be negotiated. At the same time, a successful administrative
label could feed back into later self-understanding. The direction of influence therefore ran both ways:
society shaped the census, and repeated census categories could reshape social politics.
59.23 Classification fed back through schools, courts, associations and print
The census was not isolated from other colonial institutions. School textbooks, legal decisions,
recruitment practices, municipal representation, newspapers and caste or regional associations could repeat
classifications first made authoritative in official publications. A label thereby moved from a schedule into
public debate. This feedback is one reason colonial categories mattered even when they simplified lived
reality. The effect was not uniform. Some labels remained administrative conveniences; others became
organizing terms for associations and political claims. The historian should therefore ask what happened after
classification: Was the term used in petitions? Did it appear in print? Did officials employ it in recruitment or
policy? Did local writers accept, reject or redefine it? Only this second stage reveals whether a category
became socially consequential.
59.24 Grierson’s Linguistic Survey created a different classificatory map
George A. Grierson’s Linguistic Survey of India, published in multiple volumes from 1903 onward after
long preparatory work, classified speech through linguistic comparison rather than caste rank. Volume V,
Part II treated the eastern Indo-Aryan “Bihari” and Oriya languages and supplied specimens. The survey is
fundamental for the history of Maithili and neighbouring speech forms because it documented grammatical
and lexical evidence on a scale not previously assembled by the state. Yet its map too was analytical. Decisions
about language versus dialect, boundaries and nomenclature reflected comparative philology, administrative
data and the available specimens. The survey should be read as a major scholarly achievement and as a
historically situated act of classification, not as a timeless map of immutable linguistic borders.
611611
GAJENDRA THAKUR
59.25 Linguistic boundaries rarely coincided with district boundaries
Administrative districts were designed for government, while speech continua followed settlement,
migration, rivers, markets and kin networks. Gazetteers themselves sometimes noted linguistic transitions
within a district. North and south of the Ganga could show different dominant speech patterns; border
tracts could mix forms that later maps assign to separate named languages. This mismatch is
methodologically productive. It prevents the history of Mithila, Vajji and Anga from being reduced to
modern district lines or to one-to-one language territories. Language data should be located at the smallest
scale the source permits and compared across time. A later district majority cannot prove the speech of every
village, and a linguistic label recorded in 1901 cannot by itself define a medieval cultural frontier.
59.26 Maithili, Angika and Bajjika require source-specific naming discipline
The present volume treats Maithili, Angika and Bajjika through historically controlled evidence rather
than forcing every earlier source into current terminology. Colonial records may use “Bihari,” “Maithili,”
“Tirhuti,” local dialect names, Hindi/Hindustani or other labels according to the period and classifier. The
correct procedure is to report the source’s terminology, identify the speech area or linguistic features where
possible, and then explain how that usage relates to later nomenclature. This protects regional history from
two opposite errors: erasing distinct linguistic traditions under an overly broad administrative label, or
projecting today’s standardized boundaries into a period when naming conventions and written standards
were different. The history of a language name is part of social history, not an obstacle to it.
59.27 Self-description and administrative category must remain analytically
separate
A census entry is strongest evidence for how the state recorded a person or household; it is not
automatically a complete statement of how that person understood identity. Self-description can be
recovered more directly from petitions, letters, associations, inscriptions, genealogies, literary texts,
community histories and oral testimony, though each has its own limits. The most reliable reconstruction
compares the two domains. Agreement between self-naming and administration can show institutional
consolidation; disagreement can reveal translation, resistance or competing vocabularies. This distinction is
especially important when later political identities are at stake. Historical analysis gains precision when it asks
“who used this name, in what document, for what purpose, and at what date?” rather than assuming that one
label has always meant the same thing.
Figure 235 — Evidence firewall against retrospective social classification
HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II
59.28 Colonial classifications are evidence with defined limits
Used carefully, surveys and censuses can establish a great deal: the categories administrators employed;
approximate population structures; the distribution of reported occupations; literacy and schooling;
language naming; migration and birthplace; institutional locations; and contemporary arguments about
status. They are much weaker for proving deep antiquity, biological ancestry, fixed cultural essences or a
single consensual hierarchy. Gazetteer statements about origins often repeat tradition or earlier scholarship
and require independent corroboration. Census numbers can be compared across decades only after
checking changes in definitions and boundaries. These source limits are not reasons to abandon colonial
statistics. They are the conditions that make the statistics historically meaningful rather than deceptively
exact.
59.29 Śrotriya chronology is protected by a specific Panji-based firewall
This project applies a governing chronological correction to the category Śrotriya. It does not project
Śrotriya as a fixed Maithil Brahmin sub-caste into the medieval period. The adopted Panji and Śākhā-pustak
chronology places the emergence of Śrotriya as a distinct sub-caste around c. 1800. Nineteenth- and early
twentieth-century surveys, censuses, ethnographies or gazetteers can therefore document colonial-era usage,
status claims and administrative classification, but they cannot be carried backward as proof of a medieval
sub-caste. If a colonial source uses the term, the statement is dated to that source. Earlier social organization
must be reconstructed from earlier evidence. This firewall exemplifies the wider method of the volume: later
administrative clarity must never substitute for period evidence.
59.30 Conclusion: categories were historical processes as well as historical data
Colonial surveys did not encounter a society without names, ranks, languages or communities, but they
changed the scale, form and consequences of naming. Buchanan-Hamilton’s broad district inquiries,
Hunter’s statistical compilations, the 1872 and 1881 censuses, Risley’s ethnographic and anthropometric
programme, Grierson’s linguistic survey and the district gazetteers progressively turned local descriptions
into comparable administrative knowledge. The process involved officials and local informants,
standardization and contestation, observation and theory. Its archive is indispensable for the nineteenth and
early twentieth centuries if each category is kept attached to its date, compiler and purpose. The safest
historical question is therefore not “what was this community really?” in an abstract sense, but “how was it
described, by whom, under which institutional rules, and how did that description change social action?”
Table 59.1 — Colonial classification sources: evidentiary value and limits
Source / institution What it can establish Main interpretive caution
Buchanan-Hamilton Early nineteenth-century local Breadth is exceptional, but
district surveys, descriptions of land, work, observation, local testimony and
1807–14 institutions, language and social estimation are mixed; do not
names. project entries backward.
Hunter statistical Administrative synthesis of Repeated statements may derive
accounts and district population, economy, from one earlier source; later
gazetteers institutions and earlier reports at synthesis is not independent proof
district scale. for earlier periods.
1872 and 1881 Enumerated population, Definitions, boundaries and
censuses reported compilation changed; final tables
caste/religion/language, hide translation and aggregation
occupation, birthplace and decisions.
literacy under stated rules.
Risley ethnography Contemporary social names, Anthropometric racial explanations
613613
GAJENDRA THAKUR
Source / institution What it can establish Main interpretive caution
and anthropometry customs, status claims and the are obsolete; ranked administrative
history of colonial syntheses are not timeless social
racial/ethnographic theory. ladders.
Grierson linguistic Comparative linguistic evidence, Language/dialect boundaries are
survey specimens and historical naming analytical and need not coincide
of eastern Indo-Aryan speech with district or modern political
forms. boundaries.
Panji, petitions and Period-specific self-description, Each source has its own
self-authored genealogy, claims and internal institutional bias; use dated
materials naming practices. evidence to test rather than merely
repeat colonial categories.