Ministry of Tribal Affairs holds the first meeting of the Working Group on Tribal Languages

1. At a Glance

• What it is: The Ministry of Tribal Affairs (MoTA) has set up a multi-institution Working Group on Tribal Languages. It will guide the documentation, preservation, research and promotion of tribal languages across India. Its first meeting was held on 15 October 2026 [1].
• Linked platform: It builds on Adi Vaani, an AI-based translator for tribal languages. MoTA developed it with IIT Delhi. The plan is to grow it into a national tribal language repository of corpora, primers, dictionaries and community stories [1][2].
• Why it matters for UPSC: It touches several syllabus areas: tribal welfare delivery, mother-tongue-based multilingual education, AI for Indian languages, endangered languages, and India's commitments under the UN International Decade of Indigenous Languages (2022–2032) [1][9].

2. Why in the News

• 15 Oct 2026: The Working Group held its first meeting in hybrid workshop mode. It was chaired by Smt. Ranjana Chopra, Secretary, MoTA [1].
• Participants: Padmashree Chamu Krishna Shastry (Chairman, Bhasha Samiti) and Dr. K. Sreenivasa Rao (former Secretary, Sahitya Akademi) took part, along with linguists, academics and community language practitioners [1].
• Agenda: The meeting reviewed progress on Adi Vaani and discussed a framework for a national baseline of tribal languages [1].

3. Background & Evolution

• Rationale: The Secretary said language is central to the outreach and effective delivery of welfare initiatives among tribal communities. She also wanted the work already done by institutions and scholars brought onto one common platform, so that efforts are not duplicated [1].
• Milestones:
• 92nd Constitution Amendment Act, 2003 added Bodo, Dogri, Maithili and Santhali to the Eighth Schedule, taking it to 22 languages. It was published in the Gazette on 8 Jan 2004 [7].
• 2019: UN International Year of Indigenous Languages. This led to the International Decade of Indigenous Languages (IDIL) 2022–2032, with UNESCO as lead agency [9][10].
• 1 Sep 2025: MoTA launched the beta version of Adi Vaani, described as India's first AI-powered translator for tribal languages [2][3].
• Mar 2026: In a Lok Sabha reply, the government said Santali, Bhili, Mundari and Gondi were live, Kui and Garo were in development, and seven more languages (including Katkari and Koya) were proposed for Phase II [4].
• Constitution in Santhali: The Legislative Department published the Constitution in Santhali for the first time, and President Droupadi Murmu released it [8].
• 15 Oct 2026: First meeting of the Working Group. Adi Vaani now offers text-to-text translation in nine tribal languages [1].

4. Core Static Facts

Item	Fact
Nodal ministry	Ministry of Tribal Affairs [1]
Chair of first meeting	Ranjana Chopra, Secretary, MoTA [1]
Member institutions	CIIL (Mysuru); IIT Delhi, Dharwad, Kanpur; Tezpur University; Manipur University; Assam University; Sikkim University; Living Tongues Institute for Endangered Languages; CoDIN (Council for Diversity and Innovation); Bhasha Samiti; others [1]
Proposed framework (6 areas)	1. national baseline of tribal languages; 2. research and scholarship; 3. digital documentation and language technology; 4. community-led documentation of oral traditions; 5. multilingual education and learning materials; 6. partnerships among TRIs, universities and other institutions [1]
Adi Vaani developer	Consortium led by IIT Delhi, with BITS Pilani, IIIT Hyderabad, IIIT Nava Raipur and State TRIs [2]
Adi Vaani features	Text-to-text translation in 9 tribal languages; speech; OCR; video subtitling; searchable access to 300+ TRI primers [1]
Adi Vaani data	About 13.9 lakh text sentences in 10 languages and about 300 hours of speech data, contributed by State TRIs [1]
Adi Vaani reach	48,000+ Android downloads; about 6,500 monthly users [1]
Access	Web portal adivaani.tribal.gov.in plus a Play Store app [6]
IKS support	The IKS Division, Ministry of Education has supported 38 tribal studies projects [1]
Eighth Schedule	22 languages. No objective criteria exist for adding languages [7]

Seven Pillars of language development (Chamu Krishna Shastry) [1]:

1. Sustained proficiency in the mother tongue
2. Use in education, administration, communication and entertainment
3. Contemporary literature and knowledge content
4. Continuous word generation
5. Adoption of technology
6. Teaching-learning materials and systems (online and offline)
7. Support from the Centre, States, corporates and society

5. Multi-Dimensional Analysis

Social / Tribal

• Using tribal languages improves welfare outreach and last-mile delivery [1].
• Mother-tongue-based multilingual education is proposed to address early-years learning gaps [1].
• Speakers at the centre and "no language small or big", in the spirit of Ek Bharat Shreshtha Bharat [1].

Scientific / Technological

• Adi Vaani uses AI for translation, speech, OCR and subtitling [1][3].
• IIT Delhi asked the Working Group for help with linguistic validation of datasets, standardisation of spelling and orthography, and domain-specific vocabulary for pilots [1].
• The problem is data scarcity for low-resource languages. Only about 300 hours of speech data exist so far [1].

Administrative / Governance

• The roadmap is cooperative federalism: State TRIs supply data, the Centre provides the platform [1].
• CIIL proposed phase-wise prioritisation of languages, region-wise teams with coordinators and a common repository [1].
• Sreenivasa Rao proposed a two-year action plan with quarterly reviews, and district-level community resource groups with trained tribal youth as researchers [1].

Cultural / Historical

• Proposals include community-led documentation of oral traditions, community digital archives on Adi Vaani [1], and an atlas of tribal languages and mother tongues across all States (Prof. Shailendra Mohan, former Director, CIIL) [1].

Legal / Constitutional

• Only a few tribal languages, such as Santhali and Bodo, are in the Eighth Schedule (added 2003). Many demands are pending and there are no objective criteria for inclusion [7].

International

• The initiative fits with IDIL 2022–2032. UNESCO's Global Action Plan has ten thematic areas, including education, digital inclusion and public services [9][10].

6. Recent Developments (last 12–18 months)

• 1 Sep 2025: Adi Vaani beta launched. Launch materials listed Santali, Bhili, Mundari and Gondi, with Kui and Garo in development [2].
• Nov 2025: A PIB explainer for Janjatiya Gaurav Divas listed Garo and Kui as covered, which conflicts with other releases [5].
• Mar 2026: Lok Sabha reply: 4 languages live, Kui and Garo in development, 7 proposed for Phase II [4].
• 15 Oct 2026: First Working Group meeting. Adi Vaani is reported at 9 languages for text-to-text translation [1].
• 2026–27: Four new Co-Chairs appointed to the IDIL Global Task Force [9].

7. Prelims Hooks

• The Working Group on Tribal Languages was set up by the Ministry of Tribal Affairs, not the Ministry of Education or Ministry of Culture [1].
• First meeting: 15 Oct 2026, hybrid workshop mode, chaired by Secretary, MoTA [1].
• Adi Vaani is MoTA's AI-powered tribal language translator, developed by a consortium led by IIT Delhi [2].
• Adi Vaani beta launched 1 Sep 2025 [2].
• Adi Vaani offers text-to-text translation in nine tribal languages, as of Oct 2026 [1].
• Adi Vaani hosts 300+ primers prepared by Tribal Research Institutes [1].
• State TRIs contributed about 13.9 lakh sentences in 10 languages [1].
• Phase-II candidate languages for Adi Vaani include Katkari and Koya [4].
• The "Seven Pillars" of Indian language development were set out by Chamu Krishna Shastry, Chairman, Bhasha Samiti [1].
• The IKS Division, Ministry of Education has supported 38 tribal studies projects [1].
• CIIL is in Mysuru [1].
• Santhali and Bodo entered the Eighth Schedule through the 92nd Amendment, 2003 (Gazette: 8 Jan 2004) [7].
• UNESCO is lead agency for the International Decade of Indigenous Languages, 2022–2032 [9].
• Living Tongues Institute for Endangered Languages is a member of the Working Group [1].

8. Two Ministries, Two Language Archives: Will They Join Up?

• A tribal language archive already exists, under a different ministry
• The Ministry of Education started SPPEL (Scheme for Protection and Preservation of Endangered Languages) in 2013. CIIL, Mysuru runs it [11][12].
• In its first phase it picked 117 endangered languages and mother tongues to study. Its longer-term aim is to document about 500 lesser-known languages [11][12].

• Its fieldwork has produced grammars, dictionaries in two or three languages, picture glossaries and ethno-linguistic profiles (a short record of who speaks the language, where, and how). Since 17 July 2025 this work has been on the Sanchika portal [11].

• So the same kind of material is now being built in two places

• Adi Vaani (MoTA) also collects primers, dictionaries and community stories, and it plans to become a national repository [1].

• The Secretary's own reason for setting up the Working Group was to stop work being duplicated [1]. The first real test is whether SPPEL's material on Sanchika is brought into the new repository, or whether the two keep running side by side.

• CIIL is the link between the two

• CIIL runs SPPEL, and it also sits on the MoTA Working Group [1][11].
• At the meeting CIIL asked for a common repository [1]. If that common repository does not include CIIL's own SPPEL archive, the main duplication stays in place.

9. Why a 'National Baseline' Cannot Simply Be Copied from the Census

• The Census does not list small languages on their own
• The Census places every mother tongue with more than 10,000 speakers under one of its main language headings [11].
• SPPEL counts a language as endangered when it has fewer than 10,000 speakers [11][12].

• Put the two rules together and the problem is clear. The languages most at risk are the ones the Census is least likely to show clearly.

• Most mother tongues are still unclassified

• Census 2011 recorded 2,843 mother tongues. Experts could identify only 1,369 of them as languages. The other 1,474 remain unclassified [11].

• Many of these unclassified names are probably tribal speech forms. A baseline built only from Census headings would leave them out.

• What this means for the Working Group

• Its first proposed area of work is a national baseline of tribal languages [1]. That baseline needs fresh fieldwork, not just Census tables.
• This is why the proposals for an atlas of tribal languages and mother tongues and for district-level community resource groups [1] matter so much. They are the only ways to reach languages the Census cannot see.

10. Why AI Helps the Bigger Tribal Languages First

• AI learns from data, and endangered languages have the least data
• A translator like Adi Vaani has to be trained on large amounts of text and recorded speech.
• The data so far is about 13.9 lakh sentences and 300 hours of speech, spread across 10 languages [1].

• A language spoken by a few thousand people, with no books and no fixed script, cannot supply data on this scale.

• So the tool started with the biggest tribal languages

• The first live languages were Santali, Bhili, Mundari and Gondi [2][4]. These are some of India's largest tribal languages.

• The 117 languages SPPEL treats as endangered [11][12] are exactly the ones an AI model finds hardest to learn. The technology can widen the gap between tribal languages that already have data and those that do not.

• Fixing one spelling can mean choosing one dialect over the rest

• IIT Delhi asked the Working Group to standardise spelling and orthography (the writing system and spelling rules of a language) [1].
• A model needs one standard form to learn from. But many tribal languages are spoken differently from village to village.

• Whichever form becomes the standard is the form the machine will spread. Speakers of the other dialects may stop seeing their own speech as 'correct'.

• Few downloads turn into regular use

• Adi Vaani has 48,000+ downloads but only about 6,500 monthly users [1]. That is roughly one regular user for every seven downloads.
• Downloads show curiosity. Monthly users show real need. The Working Group should follow the second number.

11. Recording a Language Is Not the Same as Keeping It Alive

• The strongest objection: this is an archive, not a rescue
• Critics can argue that dictionaries, corpora and AI tools only record a language. A language stays alive only if children grow up speaking it.
• Chamu Krishna Shastry's own Seven Pillars start with sustained proficiency in the mother tongue and its use in education and administration [1]. Technology is only the fifth pillar.

• In the Working Group's six-area framework, multilingual education comes fifth. Documentation and technology come before it [1].

• What the objection gets right

• SPPEL has been recording endangered languages since 2013 [11][12]. A recorded language can still lose its speakers if schools and offices never use it.

• The note gives no target for how many classrooms or offices will start using tribal languages. All the numbers are about data, downloads and primers [1].

• Where the objection goes too far

• Teaching children in their mother tongue needs textbooks, word lists and trained teachers. Most small tribal languages have none of these.
• The 300+ TRI primers and the corpora are exactly what classroom teaching would need first [1].
• Documentation is therefore a necessary first step. It just cannot be the last one. A fair test of the Working Group is whether its two-year plan [1] ends in classrooms or only in an archive.

12. Who Should Do What Next

• MoTA and CIIL: put SPPEL into the common repository

• SPPEL's material on Sanchika [11] should go into the national repository CIIL proposed [1], under one search system. Two separate public archives defeat the reason the Working Group was set up.

• The Working Group: rank languages by risk, not by data size

• CIIL proposed phase-wise prioritisation of languages [1]. SPPEL's test of fewer than 10,000 speakers [11] is a ready-made rule for setting that priority.

• Languages with almost no data should get fieldwork and recording first. AI translation can follow later.

• MoTA: publish one dated progress table

• Official releases have given 4, 6 and 9 as Adi Vaani's language count [1][2][4][5].

• Under Sreenivasa Rao's proposed quarterly reviews [1], one public table should show languages live, languages in development, monthly users and data collected, each with a date.

• TRIs and district resource groups: get community consent first

• Community-led documentation of oral traditions and community digital archives are already proposed [1].
• Songs, stories and ritual knowledge belong to the community. Before any of it is used to train AI or published online, the community should agree, and it should keep a say over how the material is used.
• The trained tribal youth in the district-level community resource groups [1] are the right people to ask for and record that consent.

13. Anchors for Answers

• Data: Census 2011 recorded 2,843 mother tongues. 1,474 of them are still unclassified [11]
• Data: Adi Vaani has 48,000+ downloads but only about 6,500 monthly users [1]
• Data: SPPEL's first phase covers 117 endangered languages and mother tongues (those spoken by fewer than 10,000 people) [11][12]
• Report/Committee: Working Group on Tribal Languages, MoTA (first meeting 2026). Its framework has six areas, and Chamu Krishna Shastry set out the Seven Pillars of language development [1]
• Law/Case: 92nd Constitution Amendment Act, 2003. It added Santhali and Bodo to the Eighth Schedule, and there are still no objective criteria for adding languages [7]
• Scheme: SPPEL (Ministry of Education, run by CIIL, since 2013). It is the older endangered-language archive that Adi Vaani's repository should take in [11][12]

14. Mains Relevance

• GS-I: Salient features of Indian society and diversity; Indian culture (languages, oral traditions).
• GS-II: Welfare schemes for vulnerable sections (STs); education and human resources; government policies and interventions.
• GS-III: Science and technology, specifically AI applications and indigenisation of technology.

Possible question stems:

1. "Language is the first gateway to inclusive governance for tribal communities." Discuss with reference to recent initiatives for documenting and digitising tribal languages. (250 words)
2. Examine the potential and limitations of AI-based tools such as Adi Vaani in preserving India's endangered tribal languages. (150 words)
3. Mother-tongue-based multilingual education can close early learning gaps among tribal children. Critically analyse the implementation challenges. (250 words)

15. Related Topics to Study Next

• Eighth Schedule and the demands for inclusion: Santhali, Bodo, Khasi and others; the absence of objective criteria [7].
• UN International Decade of Indigenous Languages (2022–2032): the global framework behind this initiative [9][10].
• Tribal Research Institutes (TRIs): the State-level data partners for Adi Vaani [1].
• Bhasha Samiti and the IKS Division (Ministry of Education): the parallel language and knowledge-systems effort [1].
• Janjatiya Gaurav Divas: MoTA's cultural outreach, where Adi Vaani was showcased [5].
• Constitution in Santhali / Ol Chiki script: an example of official-language recognition for a tribal language [8].
• Article 350A and NEP 2020 mother-tongue instruction: the constitutional and policy basis for MTB-MLE. These were not verified in the retrieved sources, so check them from the bare Act and policy text.
• AI for Indian languages (e.g., Bhashini): the comparable MeitY language-technology effort. Also not verified here.

16. NCERT Concepts to Revise

• Eighth Schedule (Polity): Santhali and Bodo were added in 2003; no objective criteria exist for including tribal languages.
• Census 2011 language data (Geography): Census groups mother tongues into main languages, so small tribal languages stay hidden in a national baseline.
• Cooperative federalism (Polity): State TRIs supply data while the Centre builds the platform, which is the Working Group's partnership model.
• Census (Economy): Explains why Census enumeration cannot reach small, unclassified mother tongues, so fresh fieldwork is needed.
• Community (Polity): Community-led documentation and consent over oral traditions are central to the proposed approach.
• Technology (Economy): Adi Vaani applies AI technology to languages, and data scarcity limits what it can do for small languages.

17. Common Errors / Trap Areas

• Ministry confusion: The Working Group and Adi Vaani belong to MoTA. Bhasha Samiti and the IKS Division fall under the Ministry of Education [1].
• Adi Vaani language count: Releases differ: 4 live languages (Sep 2025 and Mar 2026), 6 including Garo and Kui (Nov 2025 explainer), and 9 (Oct 2026) [2][4][5][1]. Always attach a date to the number.
• 9 vs 10: Adi Vaani translates text in 9 languages, but the TRI-contributed corpus covers 10 languages [1].
• Lead institution: Adi Vaani was developed under an IIT Delhi-led consortium. IIT Dharwad and IIT Kanpur are Working Group members, not the developers [1][2].
• Eighth Schedule date: The amendment is the 92nd, enacted in 2003 but gazetted on 8 Jan 2004. The 2004 date is often mistaken for the year of inclusion [7].