Full chapter text
Chapter 147 synthesises a central argument of this volume: language is not only a marker of identity or a
vehicle of literature. It is infrastructure through which households coordinate care, markets exchange
information, schools transmit knowledge, courts and administrations communicate rights, migrants
maintain relationships, and digital systems make people visible or invisible. Roads connect places; language
connects intentions, rules, memories and transactions. When people can understand a notice, fill a form, read
a prescription, negotiate a wage, search an archive or teach a child in a familiar linguistic register,
communication lowers the cost of participation. When language access fails, exclusion can occur even where
a formal service physically exists.
The history of Mithila, Vajji and Anga shows that this infrastructure has always been layered. Oral
traditions coexisted with manuscript cultures; Sanskrit and other learned languages interacted with
vernacular speech; Tirhuta/Mithilakshar, Devanagari, Kaithi and Persianate scribal practices belonged to
different institutional worlds; print created new publics; Hindi and Nepali expanded as state and educational
languages; Maithili, Angika and Bajjika continued through homes, performance, publication, broadcasting
and local politics; and networked digital media have added Unicode, keyboards, search, speech technologies
and machine translation. The relevant historical question is therefore not which single language ‘belongs’ to
the region. It is how multilingual repertoires have been organised, valued and made usable across changing
institutions.
147.1 Language infrastructure begins with coordination, not with official status
A language functions infrastructurally when it helps people coordinate action. Farmers exchange
information about seed, rainfall and labour; traders negotiate quality and credit; families organise marriage,
inheritance and care; patients describe symptoms; teachers explain concepts; migrants transfer instructions
and money across distance. None of these activities requires constitutional recognition in order to matter.
Conversely, formal recognition does not guarantee everyday usability. The useful unit of analysis is therefore
a communication chain: who needs to understand whom, through which medium, with what consequences
if meaning is lost. This approach shifts attention from prestige alone to the practical distribution of linguistic
capability.
147.2 Multilingualism is the normal condition of the region, not an exception
Mithila, Vajji and Anga have never been sealed monolingual territories. People move between household
speech, regional literary forms, market languages, administrative languages and languages associated with
religion or education. The repertoire changes by place and social position. A merchant, teacher, migrant
worker, student or ritual specialist may use several codes for different purposes. Historical sources likewise
represent different linguistic domains rather than a single hierarchy valid everywhere. Treating
multilingualism as normal prevents the historian from converting modern language boundaries into timeless
cultural frontiers and makes it possible to study code choice as a form of social competence.
147.3 Oral language carried institutions long before mass literacy
Before widespread schooling and print, most social knowledge circulated orally: customary obligations,
agricultural timing, songs, genealogical memory, ritual procedure, market reputation and dispute narratives.
Oral transmission was not the absence of infrastructure. It depended on recognised performers, elders,
specialists, repeated occasions and mnemonic forms that stabilised information. Even in literate societies, oral
15131513
GAJENDRA THAKUR
explanation remains crucial because documents are often interpreted collectively. The persistence of song,
story, recitation, announcement and face-to-face mediation reminds us that literacy is one layer of language
infrastructure, not its starting point.
147.4 Script converted speech into durable records but also created gatekeepers
Writing allows language to travel across time and distance without the speaker being present. Land
records, letters, genealogies, contracts, scholastic manuscripts and accounts expand institutional memory. Yet
scripts also create thresholds: one must know how to read them, possess the material, and understand the
conventions of the record. Tirhuta/Mithilakshar, Devanagari, Kaithi, Persian and other scripts therefore
belonged not only to languages but to occupations and institutions. Scribes, teachers, clerks and manuscript
custodians became mediators between oral claims and written authority. The social history of script is
consequently a history of access as well as preservation.
Figure 584 — Language acts as shared infrastructure across household, education, public institutions, markets,
media and digital systems.
147.5 Learned languages formed specialised knowledge networks
Sanskrit, Pali, Prakrit and later Persian and other learned or administrative languages connected the
region to intellectual and political networks larger than any one locality. Their importance should not be
described as simple replacement of vernacular speech. Learned languages often operated in specialised
domains—scholarship, ritual, court procedure, record keeping or elite correspondence—while everyday life
continued in regional speech. Such functional differentiation could widen knowledge networks but also
concentrate authority among trained groups. The history of language infrastructure must therefore
distinguish domain from population: a language can have enormous institutional influence without being
the ordinary mother tongue of most residents.
147.6 Literary Maithili expanded the capacity to address publics beyond the
immediate household
The formation of literary Maithili created a medium in which lyric, drama, devotional expression,
narrative and later prose could circulate beyond a single village. Literary standardisation makes texts more
portable, but it also selects among regional forms. The resulting tradition should not be mistaken for
HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II
evidence that all speakers used one uniform vernacular. Its infrastructural significance lies elsewhere: it
supplied a recognisable written medium, a repertoire of genres, models for teaching and publication, and a
basis for institutions that could later claim a regional literary public.
147.7 The two Vidyāpatis belong to different historical language infrastructures
The governing distinction maintained throughout this volume remains essential here. Mahākavi
Vidyāpati, the earlier Maithili Padāvalī poet, and the later Vidyāpati Ṭhakkura associated with the Oiniwar
court and Sanskrit/Avahaṭṭa works belong to different historical contexts. Conflating them collapses distinct
literary and institutional environments into a single biography and distorts the chronology of language
formation. Language infrastructure is built by texts, audiences, scripts and institutions at particular times;
accurate attribution is therefore not a minor philological issue but part of reconstructing how literary publics
were actually organised.
147.8 Tirhuta/Mithilakshar linked textual memory to a regional scribal ecology
Tirhuta/Mithilakshar served manuscript, genealogical and literary functions in Mithila and became a
visible sign of regional textual culture. Its historical importance cannot be measured only by how many
people use it in daily writing today. A script can persist as archival infrastructure, cultural knowledge and a
technical requirement for reading older documents. Modern digitisation adds a new dimension: encoded
text, fonts and keyboards can reconnect readers to manuscripts only when transcription standards, training
and provenance are maintained. Script revival is strongest when it supports reading, teaching and archival
work rather than functioning only as ornament.
147.9 Persianate and colonial administrations changed the languages of official
legibility
Changes of political authority altered the linguistic interfaces through which land, revenue, petitions and
law became legible to the state. Persianate administrative practices and later colonial regimes introduced new
documentary conventions, vocabularies and clerical skills. The shift was not merely linguistic; it changed
who could mediate between villagers and offices. Colonial standardisation further linked language to
censuses, schools, courts and print. Every such change redistributed transaction costs: some groups acquired
new opportunities through literacy in the administrative language, while others depended more heavily on
intermediaries.
147.10 Census and survey categories transformed fluid speech into countable
populations
Modern censuses require each response to enter a table. This makes language statistically visible but can
simplify multilingual practice. Mother-tongue names may be grouped, separated or reported differently
across censuses, and political recognition can affect what respondents choose to name. Census of India
language tables and Nepal’s population censuses are therefore indispensable but not transparent mirrors of
speech. Their infrastructural effect is nevertheless real: counted languages can become objects of policy,
schooling, representation and advocacy. The historian must read the table as both evidence about population
and evidence about classification.
15151515
GAJENDRA THAKUR
147.11 Print reduced the cost of repetition and created durable language publics
Printing multiplied the reach of grammars, newspapers, journals, textbooks, pamphlets and literary
works. A manuscript may circulate through a small chain of readers; a printed edition can produce
simultaneous publics and standardise spelling and terminology. Print also enables correction, debate and
serial publication. In Maithili, Angika and Bajjika cultural worlds, print institutions helped turn dispersed
speakers and writers into audiences who could imagine a shared field of discussion. The same process,
however, privileges varieties that obtain presses, editors, typefaces, distribution and school demand. Print
infrastructure therefore amplifies both communication and inequality.
Figure 585 — Language infrastructure accumulates across oral, manuscript, print, broadcast and networked
digital media.
147.12 Schooling makes language an economic resource as well as a cultural one
The language of schooling affects comprehension, progression, examination performance and the
relationship between home knowledge and formal curriculum. Mother-tongue support can reduce the
cognitive burden of learning unfamiliar content through an unfamiliar code, especially in early grades. At
later stages, competence in wider languages can expand access to higher education and labour markets. The
policy problem is not solved by choosing one side. A strong system needs additive multilingualism: children
should gain literacy and conceptual confidence in familiar languages while acquiring the additional languages
required for wider mobility.
147.13 Textbooks and terminology are the hidden engineering of language policy
A language cannot function in education or administration through symbolic recognition alone. It needs
stable spelling practices, subject vocabulary, dictionaries, edited texts, translation conventions and materials
that teachers and officials can actually use. Terminology work is especially important in science, law, health
and technology because literal borrowing without explanation can reproduce exclusion in a new script. The
quality of language infrastructure is therefore visible in mundane artefacts: a clear form, a usable glossary, a
graded reader, a well-edited dictionary and a consistent place-name list.
147.14 Examinations convert linguistic competence into mobility
Public examinations, university admissions and recruitment tests give language choices direct economic
consequences. A candidate who understands the subject but struggles with the examination language faces an
additional barrier unrelated to substantive knowledge. At the same time, common languages can make large-
HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II
scale assessment administratively feasible. Fair systems require transparent language options, reliable
translation, equivalent question difficulty and clear rules about scripts and terminology. Examination
language is therefore an example of how linguistic design can redistribute opportunity without changing the
formal syllabus.
147.15 Administration is accessible only when citizens can understand
procedures
Government offices generate notices, certificates, land records, welfare forms, health instructions and
grievance procedures. Physical proximity to an office does not guarantee access if the citizen cannot
understand the document or explain the problem. Front-line multilingualism, plain-language drafting,
interpretation and translated interfaces reduce this friction. The objective is not to reproduce every
document in every variety regardless of demand, but to design service chains around actual linguistic needs.
Language access is most valuable where misunderstanding carries high costs: identity records, land, benefits,
policing, health and disaster warnings.
147.16 Courts reveal the difference between legal language and understood
justice
Legal systems require precision, precedent and authoritative records, which encourages specialised
vocabulary and standard languages. Yet a litigant must also understand the accusation, testimony, order and
remedy. Interpreters and translated documents therefore form part of due process, not an optional cultural
service. Historical court language has often differed from household speech, making legal intermediaries
especially powerful. A language-infrastructure perspective asks how reliably meaning moves between litigant,
lawyer, witness, clerk and judge, and where errors or asymmetries can affect rights.
147.17 Health communication turns words into treatment outcomes
Diagnosis begins with description. A patient must communicate symptoms, duration, medication use
and consent; a health worker must explain dosage, warning signs and follow-up. In multilingual regions,
misunderstanding can reduce adherence even where medicines and clinics are available. Visual aids and
multilingual digital tools can help, but they cannot substitute for trust and the ability to ask questions.
Public-health campaigns similarly depend on language choices that people recognise and believe. Health
therefore demonstrates the infrastructural principle clearly: communication is part of the service itself.
147.18 Disaster warnings must be linguistically local as well as technically
accurate
Forecasting systems may generate precise information about floods, storms or heat, but warnings save
lives only if people receive and understand them in time. Local vocabulary for river stages, evacuation routes,
livestock movement and shelter can be more actionable than abstract technical terminology. Multiple
channels—sirens, radio, mobile messages, village announcements and social networks—should reinforce one
another. Language thus forms the last mile of disaster infrastructure. A technically sophisticated forecast that
fails at the communication interface has not completed its public function.
15171517
GAJENDRA THAKUR
Figure 586 — Formal recognition becomes usable public access only when content, interfaces, comprehension
and service outcomes align.
147.19 Markets depend on linguistic trust and repeated reputational information
Prices are numbers, but markets operate through much more than numbers. Buyers and sellers discuss
quality, credit, delivery, authenticity, warranties and risk. Shared linguistic repertoires lower bargaining costs
and support reputational networks, while multilingual brokers connect otherwise separate markets. Regional
languages can therefore carry substantial economic value even when formal contracts use another language.
The rise of digital commerce extends the same principle to product descriptions, customer service and search.
A seller who cannot describe a product in the buyer’s language may be invisible in a market that is technically
open.
147.20 Migration makes language portable infrastructure
Migrants carry speech practices to destinations and acquire new languages for work, housing, transport
and bureaucracy. They also maintain home-language networks through calls, messages, songs, video and
remittances. This produces multi-local linguistic households in which children, elders and migrants may have
different dominant repertoires. Language competence can affect occupational mobility and vulnerability at
destination, but home-language continuity can sustain care, identity and information flows. Migration
therefore does not simply erode regional languages; it reorganises the places and media through which they
are used.
147.21 Gender shapes who acquires, transmits and can publicly use linguistic
capital
Women are major transmitters of songs, ritual vocabulary, child language and household knowledge, yet
historically their access to formal schooling, print authorship and public institutions has often been more
restricted. Educational expansion and digital media can widen participation, but device ownership, mobility,
time and safety still affect whose voice becomes public. A language may be vigorous in domestic transmission
while women remain underrepresented in publishing, administration or digital production. Measuring
vitality therefore requires attention to gendered control over both speech domains and communicative
technologies.
HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II
147.22 Caste and class affect language through unequal access to institutions,
not through fixed speech essences
Speech varieties can correlate with locality, occupation and social networks, but caste or class should not
be treated as possessing timeless linguistic essences. What matters historically is access to schooling, scripts,
clerical employment, literary institutions, migration and prestige registers. Groups excluded from formal
literacy may sustain rich oral repertoires; upward mobility may encourage acquisition of state or
metropolitan languages without erasing home speech. The key mechanism is unequal access to institutions
that reward particular forms of linguistic capital. This approach avoids turning social hierarchy into linguistic
determinism.
147.23 Angika and Bajjika demonstrate why public naming and linguistic
relationship must be kept analytically distinct
Angika and Bajjika occupy important regional speech, literary and cultural fields within the wider history
of Mithila–Vajji–Anga. Their relationship to neighbouring Maithili varieties has been described differently
by linguists, activists, institutions and census systems. A historical synthesis should not resolve that debate by
fiat. It should record how speakers and institutions name these forms, where they are used, how they appear
in censuses or academies, and how literary production develops. Public life depends on usable
communication regardless of whether a variety is classified as a language, dialect or regional style in a
particular scheme.
147.24 Nepal’s Madhesh shows how federalism turns multilingualism into an
institutional design problem
Madhesh contains large Maithili and Bajjika-speaking populations alongside Bhojpuri, Nepali, Urdu and
other linguistic communities. Federalism creates additional arenas—province and local government—in
which language can be used for education, administration and cultural policy. Constitutional principles and
language-commission processes provide frameworks, but implementation depends on legislation, staffing,
materials and political agreement. The lesson is broader than Nepal: decentralisation can bring language
policy closer to speakers, yet it also requires careful coordination so that local inclusion does not create new
barriers to mobility or intergovernmental communication.
147.25 Broadcasting created a shared listening public before the internet
Radio, recorded music, cassette culture, television and video expanded regional-language publics across
distance and literacy levels. Broadcasting gives speech a new scale: pronunciation, performance styles and
vocabulary that were once local can become widely recognisable. It also creates gatekeepers through
programming schedules, licensing and audience measurement. For Maithili and related regional forms,
broadcast history matters because sound media preserve voices that print cannot capture and can reach
households with limited literacy. Digital streaming now extends this archive while making discoverability
dependent on platform metadata.
147.26 Unicode changed scripts from images into searchable text
Digital preservation requires more than photographing a page. When a script is encoded in Unicode and
supported by fonts and input methods, text can be searched, copied, indexed, compared and processed by
software. Tirhuta’s inclusion in the Unicode Standard therefore created a technical foundation for digital
use, but encoding alone does not guarantee adoption. Fonts, keyboards, OCR, teaching materials and
15191519
GAJENDRA THAKUR
correctly encoded corpora are also needed. The distinction mirrors the chapter’s wider argument: formal
recognition creates a possibility; infrastructure turns that possibility into routine capability.
147.27 Artificial intelligence exposes the consequences of low-resource language
inequality
Speech recognition, text-to-speech, translation, OCR and language models improve when they have high-
quality corpora, lexicons and representative recordings. Languages with little digitised material may be poorly
recognised, while regional accents and code-mixed speech can produce systematic errors. Public programmes
such as BHASHINI expand the technical ecosystem for Indian languages, but smaller varieties still require
deliberate data creation and evaluation. Community consent, provenance and privacy matter because
linguistic datasets contain voices and cultural knowledge. AI policy should therefore treat data stewardship as
part of language infrastructure, not as a purely technical procurement problem.
Figure 587 — Digital language capability depends on encoding, corpora, speech tools, platforms, archives and
governance working together.
147.28 Archives, dictionaries and corpora are productive infrastructure for future
language use
An archive is not only a repository of the past. Digitised manuscripts, newspapers, recordings,
dictionaries, thesauri and aligned translations can support education, publishing, search, OCR and machine
translation. Their value increases when metadata records date, place, author, script, edition and provenance.
Poorly documented scans preserve appearance but limit reuse; normalised transcriptions without page links
can erase uncertainty. Sustainable language infrastructure therefore combines preservation-quality images
with searchable text, version history and clear licensing. This creates a bridge from cultural memory to future
research and technology.
147.29 A language-infrastructure dashboard should measure usability rather
than prestige alone
Useful indicators would include early-grade materials available, teachers trained, administrative forms
translated, interpreter access, broadcast hours, books published, digital corpus size, keyboard and font
availability, search quality, speech-recognition accuracy, archived hours of audio, and the share of users who
HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II
can complete a service in a preferred language. These measures reveal where the chain breaks. A language may
have literary prestige but weak digital tools; another may have vigorous oral use but few textbooks.
Measurement should therefore be domain-specific and should never reduce vitality to a single speaker count.
147.30 Conclusion: language is infrastructure because society repeatedly has to
move meaning
Across the long history of Mithila, Vajji and Anga, the central linguistic problem has never been simply
which language exists. It has been how meaning moves between generations, households, markets, schools,
courts, states and technologies. Oral transmission, scripts, print, broadcasting and digital systems have
accumulated rather than replacing one another. Maithili, Angika, Bajjika and the wider multilingual
repertoire of the region remain socially consequential when they enable people to learn, transact, remember,
claim rights and create. The strongest future policy is therefore neither linguistic homogenisation nor
symbolic recognition without capacity. It is durable multilingual infrastructure that makes communication
usable, interoperable and open to new generations.
Table 147.1 — Evidence architecture for reconstructing language as social infrastructure
Evidence source What it can establish Main caution
Census language tables mother-tongue reporting; categories and self-reporting
broad spatial distribution; can change; multilingual
changes in public naming practice is simplified
Manuscripts, inscriptions & scripts, documentary survival is socially selective;
records domains, scribal institutions written language does not
and historical terminology equal everyday speech
Print & publishing archives standardisation, public publication counts do not
debate, readership, textbook measure everyday
and literary production comprehension
Education records medium of instruction, formal provision may differ
enrolment, curriculum, from classroom practice
examinations and language
options
Administrative & legal official interfaces, forms, recognition does not prove
materials court language, translation service-level usability
and language rights
Broadcast & performance spoken registers, reach, music, audience size and passive
archives drama and regional public exposure may be difficult to
culture reconstruct
Digital corpora & platform encoding, searchability, platform metrics are
data corpus scale, speech tools and proprietary and can change
online creation quickly
Interviews & ethnography code choice, comprehension, local samples cannot
language attitudes, household automatically be generalised
transmission and service to the whole region
barriers
Dictionaries, thesauri & lexical resources, normative resources may
terminology standardisation, translation underrepresent variation and
capacity and machine- living usage
readable data
15211521