OpenSpeaks/Oral Knowledge Framework

The Oral Knowledge Framework is for language archivists. It helps document how they recorded oral knowledge, how they engaged community members to discuss, disagree and reach a consensus. It is based on OpenSpeaks Archives' prior work, a pilot (2024–2025) and an implementation phase (2025–2026), and case study analysis.

Contents

• Parameter 1: FAIR and CARE principles
  • Practical steps during recording
  • Fair, equitable remuneration
• Parameter 2: Community assessment of oral histories
• Parameter 3: Multilingual transcription

Parameter 1: FAIR and CARE principles

It is written for three main kinds of documenters:

• community members recording their own language and culture
• neighbouring Indigenous documenters
• external non-Indigenous documenters

The guide follows two sets of principles:

• FAIR: Findable, Accessible, Interoperable, Reusable (helps organise recordings and metadata so that they are useful now and in the future)

• CARE: Collective Benefit, Authority to Control, Responsibility, Ethics (ensures recording process and output respect Indigenous rights, community priorities, and lived experience)

FAIR

FAIR is about making digital files and their descriptions easy to find, access, combine, and reuse.

• Findable: People should be able to discover recordings using a simple search. Each file needs a clear title, language name, and keywords.
• Accessible: When someone finds a file, they should know how to request or access it. Sometimes access is open, sometimes it is restricted.
• Interoperable: Files and metadata should use common formats and vocabularies, so that different tools and platforms (including Wikimedia projects) can work with them.
• Reusable: Files should have clear licences, details about their origin, and descriptions so that others can reuse them without guessing.

CARE

CARE is about people and relationships. It centres Indigenous and other marginalised communities, not only the data.

• Collective Benefit: Documentation should help the whole community, not only the project, researcher, or platform.
• Authority to Control: Communities have the right to decide what is recorded, how it is described, how it is shared, and who can reuse it.
• Responsibility: Documenters must act with care, follow local protocols, and support capacity-building in the community.
• Ethics: Documentation and reuse must respect dignity, safety, and local values. Not everything that can be published should be published.

Documenter roles

This guide uses three common roles. In real life, there might be more roles, and one person may fit more than one role.

• Community member: a speaker or community insider recording their own language and culture.
• Neighbouring/other Indigenous member: a person from an Indigenous community documenting an Indigenous language that might be close to their own language.
• External non-Indigenous member: a non-Indigenous person from outside, such as a researcher, volunteer, NGO worker, or Wikimedian.

The FAIR and CARE steps are similar for all, but responsibilities and power dynamics are different.

Before you record: planning with FAIR and CARE

Shared steps for all

1. Talk first, record later
   • Explain who you are, why you want to record, and how the recordings may be used.
   • Ask what the community wants from the documentation (e.g. teaching materials, archives for younger people, or citations for Wikipedia).
2. Agree on topics together
   • Prioritise topics that the community cares about. This can include traditional knowledge, biodiversity, stories, songs, local histories, or everyday conversations.
   • Avoid sensitive/private topics unless there is clear, informed consent and local approval.
3. Decide together what can be public and what cannot
   • The community does not ever want to publish some content online (e.g. sacred knowledge, restricted rituals, sexual topics or material that could cause harm).
   • Some content can be shared, but only with conditions (e.g. only within the community, only after a certain time, or only by disclosing real individuals' names)
4. Choose consent and release forms
   • Use simple, translated consent forms that explain:
     • who is recording
     • what will happen with the recordings
     • how the interviewee can say “no” or withdraw
     • what platforms may host the files (e.g. Wikimedia Commons, local archives)
5. Plan how the community will benefit
   • Decide how you will share back: copies of recordings, subtitles, local screenings, training, or community-controlled archives.
   • If there is funding, agree on fair remuneration and how a part of the project budget will be kept for community priorities.

Extra responsibilities by role

Community member

You are documenting “from within”. You may already know local protocols, but it is still important to ask elders and others about what is appropriate. Plan how other community members will be involved in reviewing, subtitling, and using the materials so that benefits are shared.

Neighbouring Indigenous member

You may understand similar histories of oppression and erasure of your own community. Still, you are a guest in a different community. Ensure local leaders and elders invite you and guide the process. Pay attention to differences in protocol, and avoid assuming that your community’s practices automatically apply.

External non-Indigenous member

You carry more responsibility because of power differences. Do not design the project alone. Co-design it with community members, share decision-making, and be prepared to step back if the community does not want certain kinds of documentation or publication.

Practical steps during recording

Ethical recording practices (CARE)

Explain again before you record

repeat the key points: topic, purpose, where it may be shared, and the right to stop at any time.

Respect pauses and refusals

If someone does not want to answer or seems uncomfortable, change the question or stop. Not answering is always an acceptable answer.

Co-create the questions

When possible, ask the interviewee or community members what questions they want to be asked. Allow them to suggest topics, corrections, or different angles.

Include context

: Record short segments about who is speaking, where, when, and why this topic matters. This helps with future interpretation and reuse.

Technical practices (FAIR)

Use clear, easy-to-understand file naming

Example: LanguageCode_Village_Topic_SpeakerInitials_YYYYMMDD.extension

Record good audio and video

Use a stable device, minimise background noise when appropriate, and check that everyone is clearly visible or heard, unless they ask not to be.

Capture basic metadata while recording

Note language(s), location (at the level the community is comfortable sharing), date, and key themes. This can be in a notebook, form, or app.

Think about accessibility

If possible, record in a way that later allows adding subtitles, transcripts, or translations. Shorter segments are often easier to subtitle and reuse.

After recording: describing, reviewing, and deciding access

Community review (CARE)

1. Community-first review
   • Before uploading anywhere, share the recordings with interviewees and, where appropriate, community reviewers.
   • Ask: “Is this accurate? Is anything missing? Is anything you want removed or kept private?”
2. Record decisions
   • Note which recordings are:
     • public
     • restricted (for community only)
     • delayed (to be released after a certain time or event)
     • never to be shared
3. Discuss how to handle sensitive material
   • Agree on how to treat names, faces, sacred sites, or personal stories. This might include blurring faces, changing names, or keeping only audio.

Metadata and licences (FAIR)

Write rich but respectful descriptions

Include language, content summary, keywords, and any relevant cultural context that the community approves.

Use agreed licences

For public materials, choose licences that reflect community decisions. For Wikimedia, this often means Creative Commons ShareAlike (CC BY-SA), but only if the community understands and accepts what that implies. Community might allow the release of some, but not all, recordings under CC BY-SA licenses. Respect and follow their decision.

Connect to other data

Where appropriate, link to related recordings, Wikipedia articles, or external references. This increases interoperability and reuse.

Keep metadata even when access is restricted

Even if the file itself is not public, a minimal description (with community approval) can help others know that the resource exists and how to request access.

Using recordings in Wikimedia and other platforms

For community members

• Decide together which recordings can go to Wikimedia Commons or other open repositories.
• Use subtitles, transcripts, and translations to make content usable in different Wikimedia projects:
  • Wikipedia articles
  • Wikidata items
  • Wikisource (where appropriate, for peer-reviewed texts or transcripts)
• Think of Wikimedia as only one layer. Keep or create community-controlled copies and archives first.

For neighbouring Indigenous documenters

• Recheck that publishing on a global platform matches the wishes of the recorded community.
• Where you act as a bridge to Wikimedia, keep local leaders involved in decisions about article content, image choice, and contextualisation.
• Share back: show which pages and entries were created or improved because of the recordings.

For external non-Indigenous documenters

• Ensure that local learning resources, language classes, or community archives are part of the project design.
• Do not rush to upload everything online. Prioritise items that the community explicitly wants to see in Wikimedia, and that they feel are correctly framed.
• In academic or public writing, cite the community and interviewees appropriately, and share drafts with them for feedback before publishing when possible.

Role-specific checklists

Community member checklist

• ☑ Ask elders or leaders – what to document and what to avoid.
• ☑ Use simple consent explanations in your language.
• ☑ Plan how recordings will be used to teach younger speakers.
• ☑ Keep copies in safe, community-controlled storage.
• ☑ Decide together which files go to Wikimedia and which stay internal.

Neighbouring Indigenous member checklist

• ☑ Ask to be invited; do not assume you should document.
• ☑ Learn and follow local protocols, even if they differ from your own.
• ☑ Share equipment and skills, not only files.
• ☑ Ensure that decisions about access and licences rest with the documented community.
• ☑ Be ready to hand over archives if the community prefers to manage them directly.

External non-Indigenous member checklist

• ☑ Co-design the project with community members from the very beginning.
• ☑ Budget for fair compensation, training, and community-controlled infrastructure.
• ☑ Use consent and governance tools that are readable and translated.
• ☑ Plan for exit: how will the work continue without you?
• ☑ Be accountable: report back, share findings, and be open to critique.

Fair, equitable remuneration

This compensation model is grounded in FAIR-CARE principles, Equity-Centered Participatory Compensation Model (EPCM) (centring marginalised voices and fair compensation), reparative justice (acknowledging historical oppression and systemic barriers), and intersectionality (compound marginalisation across multiple identities).

Compensation structure

Note that the below table has an amount to give an example. You need to fix this amount based on your local labour system.

Role	Base Rate (USD/hour)	Rationale
Native speaker interviewee	$25–35	Knowledge holder; primary contributor
Subtitle creator/translator	$20–30	Skilled linguistic work; bilingual/multilingual
Community reviewer	$20–25	Quality assurance; cultural verification
Language coordinator	$25–35	Project management; community liaison
Basic technical support	$18–25	Basic technical help

Intersectional adjustment factors

Historical Marginalization (caste/race/ethnicity)

Category	Multiplier	Examples
Scheduled Caste (Dalit)/other severely marginalised group	+30%	Dalit and other severely marginalised group
Scheduled Tribe (Adivasi)/Tribal/Indigenous/Aboriginal/Racial minority (Black/Afro-descendant)	+25%	Historically marginalised Indigenous group
Other Backward Classes (OBC)	+15%	OBC communities in South Asia
Religious minority (persecuted/historically marginalised)	+15%	Muslims in India/Myanmar/United States or Christians in Pakistan or Tamil Hindus in Sri Lanka, etc.

Gender-based marginalisation

Category	Multiplier	Examples
Trans/non-binary persons	+25%	Additional barriers to economic participation
Women/Femme-identifying persons	+20%	Gender pay gap and barriers to participation
Widows (in patriarchal contexts)	+15%	Economic vulnerability in South Asian contexts

Disability and health

Category	Multiplier	Examples
Physical disability	+25%	Mobility impairments, chronic pain, etc.
Sensory disability	+25%	Blind/low vision, Deaf/hard of hearing
Neurodivergence	+20%	Autism, ADHD, learning disabilities
Chronic illness	+20%	Long-term health conditions affecting work capacity

Applying the Matrix (calculation)

Multipliers are added to reflect compounding marginalization.

Formula:

Final rate = base rate × (1 + Sum of all applicable multipliers)

Example 1: Dalit woman interviewee

• Base rate (interviewee): $30/hour
• Dalit (SC): +30%
• Woman: +20%
• Total multiplier: +50%
• Final rate: $30 × 1.50 = $45/hour

Example 2: Adivasi woman with disability, subtitle creator

• Base rate (subtitle creator): $25/hour
• Adivasi (ST): +25%
• Woman: +20%
• Physical disability: +25%
• Total multiplier: +75%
• Final rate: $25 × 1.75 = $43.75/hour

Parameter 2: Community assessment of oral histories

Community review of recordings of Indigenous languages can match the principles of peer review in scientific publishing. A community can have a protocol to check, correct, and approve what has been recorded.

Considerations for the assessment process:

• which community members reviewed the material
• what concerns or corrections they raised
• how disagreements were discussed and resolved
• what final consensus the group reached

Documenting the above can act as evidence of verification, similar to editorial review notes in a conventional peer publication. The documentation can show that the recording was not only created but also checked and approved.

The assessment process needs to follow local community protocols and ways of knowing. This can include who is allowed to review certain materials, what must remain private, and how authority and expertise are recognised in the community.

Parameter 3: Multilingual transcription

The goal of this parameter is to prepare oral knowledge recordings for citing. If resources allow, each recording should have a transcription and subtitles in at least one widely used language, preferably a neighbouring official language that native speakers are comfortable with. This helps others outside the community understand the content.

Transcription quality also needs review. Community members can check the transcript and subtitles, just as they review the recording itself (see Parameter 2). This creates multiple layers of verification.

For citations, time-coded transcripts can serve as page numbers in printed texts. Each statement can be linked to a specific time in the recording. This reduces the perception that oral sources are “imprecise” and gives editors and readers a clear way to verify what was said and where it appears in the media.

How to use this framework

This framework is for use while documenting languages and sharing the recordings.

Bibliography

• BBC Subtitle Guidelines, Version 1.2.5 (March 2026)
• BBC Academy: How do I create subtitles?
• EBU-TT-D Subtitling Distribution Format (Tech3380)

Contribute & discuss

This is an open educational resource and is open to all who want to contribute. Please edit this page to contribute. If you want to discuss anything, please do so in the Discussion page.