IGF 2021 – Day 2 – Town Hall #42 Building the wiki-way for low-resource languages

The following are the outputs of the captioning taken during an IGF virtual intervention. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid, but should not be treated as an authoritative record.



>> SUBHASHISH PANIGRAHI: I believe we can start now.  Hi, everyone.  Ruth, are you a participant, just checking?  We are waiting for two of our panelists, just in case you have a different name on your profile.

>>  RUTH:  I'm a participant.

>> SUBHASHISH PANIGRAHI: Okay, thank you.  Hi, everyone.  Thank you for joining us, and thank you, all the panelists for taking time from your busy schedule.  I will quickly walk everyone through the plan that we have for today, the schedule that we have for today, and then introduce everyone very quickly and then we'll get into some of the discussions we have planned.  I'm Subhashish Panigrahi.  I volunteer for Rising Voices.  So this event is co‑organized by O Foundation and Rising Voices.  And today we have six panelists with us, and my colleague Sailesh will be the reporter, and thank you, Sailesh, for joining and helping with that process.  That's a lot of hard work.  Probably Sailesh could go first, introduce himself and then I'll quickly introduce all the panelists and then we'll get into the discussions.


>> SAILSH PATNAIK: I'm happy to join here today to listen and learn from the work you all are doing.  I'm Sailesh Patnaik, I'm representing the O Foundation today.  I'm a comedian, community organizer, an open access advocate.  I have spent the last decade being part of the community and movement.  The interest of Wikipedia and projects and pedagogy and documentation.  Back to you.

>> SUBHASHISH PANIGRAHI: I'll quickly introduce the panelists we have with us.  I'm still waiting for two more panelists, as I said earlier.  In the meantime, I'll go to the rest of the panelists.  Eddie Avila the Director of Racing voices, it is an initiative of the citizen media organization, global voices, and global voices publishes in 40 different languages and it has been key advocate of language activism for the protection and fostering of Low‑resource languages.  Eddie will be sharing a lot of insights from his work spanning over decades in the movement and specifically from the Latin America communities that have been trying to work on several innovative approaches towards protecting the languages that are spoken in the region.  Next is Amrit Sufi, Amrit is a coordinator of the oral culture transcription project, a toolkit that will enable people to access information on uploading media, of endangered languages.  Amrit has documented folk songs in the language of Indian.   Angika is endangered from the eastern part of India.

Then next is Saldana Ivanova.  Saldana is interested in Low‑resource languages, and she also conducts research and works in the development of various language technology tools for the support of the sack what ‑‑ Sakha language in the far east.  Is Mahir Morshed.  He has contributing to the Wikipedia lexigraphy.  As I'm waiting for two more panelists, I'll introduce them just in case they come a bit late.  Omolade Abidemi is a Yoruba language scholar from Nigeria, she engaged the community for wider dissemination in Yoruba around Nigeria, next is Khadijah Abubakari, a professional teacher and Dagbani language advocate in Ghana.  She uses Wikipedia as a tool to grow and spread the Dagbani language outside Ghana.  And welcome all the panelists once again and thank you for joining.

The first thing that I would ask each of the panelists would be a quick introduction and the prompts that I have for each one of you is slightly different from each other.  My first person would be when you introduce yourself, can you also please share the language, digital activism work that Rising Voices is doing and how Rising Voices is actively engaging stakeholders for using language DOJ activism.

>> EDDI AVILA: Thanks after, I'm Eddie Avila, the Rising Voices Director of the organization global voices based here in Bolivia, in South America, a lot of our work is on language DOJ activism.  What we like to, digital activism.  Leveraging the internet, digital media, the technologies to promote and to revise lies languages in digital spaces, what does that mean, basically people around the world that see the internet and media to create new content to excite potentially younger generations of speakers and to position their language and culture in the digital spaces.  So as we know, one of the keys for language vitality is to make sure these languages are being used in new domains including the internet.  For the last, I would say, 10 to 14 years, I think officially we started using this term, language activism back in 2014 and we have seen an explosion of communities around the world that see it as an opportunity.  What we like to do with Rising Voices, create spaces for peer learning, for change, to show what people are doing around the world, but show the side that there are challenges involved, but also show that ‑‑ amplifying or highlighting initiatives that are finding innovative ways to address those challenges.

What we have been trying to do around this topic is to create opportunities for that peer learning and exchange, whether through campaigns, we do a lot of social media campaigns, we return gatherings where people, for example, in Mexico, come together and see what others are doing in a kind of common response we get along the way is that I didn't know anyone else was doing this as well.  The spout they are working on different languages, just seeing and meeting others involved in the field is a really, really strong motivator and really helps sustain this movement.

With the international decade of the digital languages coming up, we have been working with UNESCO to create a toolkit, not necessarily as a step by step recipe on how to become involved in this, but, rather, you know, a broad overview on different strategies people can take, whether it's normalizing, which includes localization, just know, a lot of the wikimedia work, there their language in spaces, so it becomes more commonplace to see their language on line.  Education, we have seen an explosion of online learning classes using these tools, including messaging apps or video conferencing platforms.

So I think more and more, as people see what others are doing, it can really inspire what can be done instead of prescribing this needs to happen, having people opt into this and hopefully different networks and communities that can support them by helping them understand, you know, what's involved but give them motivation, so this continues to grow.

>> SUBHASHISH PANIGRAHI: Thanks so much, Eddie.  Next is Amrit.  As you've worked very recently with the toolkit that would help eventually for transcription of oral culture, I'd love to hear from you and others also love to hear from you, your experience as a native speaker of Angika and what was your experience around facilitating the digital activism, the language digital activism for India and how do you see open and collaborative processes helping Low‑resource languages.

>> AMRIT SUFI: Thank you.  Hello around, I'm right Sufi.  I speak Angika as my native language, I will recount something from my childhood that is  ‑‑ when I was child, my entire family, most of them used to speak Angika language, and I was comfortable talking in that, but as it came to my schooling years, three or four years old, my parents began to encourage me to talk in Hindi and English.  In public schools, it is Hindi and English.  Learning these languages, so yeah for the article culture toolkit, I talked with the speakers, academicians, when you asked them if they would like to ‑‑ most of them were like about the age of 16.  I asked them if they would like this to digitize their works, especially the writers, the difficulty with learning digital skills, they lacked knowledge about copyrights, I'm working on those aspects, including those things in the toolkit.

So, yeah, I think what can be done, the Angika language culture is brought online, not dependent entirely on education policies and laws and resources or lack of them.  Yeah, coming to the Rising Voices languages workshops, what I learned, I learned a lot from the participation of the people from all over India, and yeah, they had their own ideas on how to promoting language and how to work on them, everyone had specific skill sets.  I think during the workshops, they also learned how to promoting language on these platforms creatively, there are various ‑‑ they can participate, they can use public resources, already been done by volunteers, language activists, and that they can bring their language in online platform, take inspiration from initiatives of language activists, around the world, and they can do a lot of things maybe building apps, or promote Mr. Language by social media and create, and so on and so forth.  Whether individually or collaborating with other people, other activists.

My conclusion is language activists are ‑‑ in language activism, it will bring the languages on global platforms and also engage and include young speakers of the languages.  So that's it from me, thank you.

>> SUBHASHISH PANIGRAHI: Thank you so much, Amrit.  Sardana, you use technology a lot in your work, especially in building different tools and other kinds of technical resources that would be helpful for starting with the Sakha language community, your work in academics and from that experience, how do you see the creation of computational linguistics of language tools, making a long‑term shift in an equitable and decentralized manner.

>> SARDANA IVANOVA: Hi, everyone, my name is Sardana, I'm a third-year student in the University of Helsinki, I was born in Yakutsk, and I came to Helsinki to do my Ph.D. thesis.  The preliminary title is language technology tools to support Low‑resource languages, a case study of Sakha.  As title suggests, my thesis, my Ph.D. dissertation will be about various language technology tools which could support Low‑resource languages and my dissertation, I work on creation of tools to support Sakha.  So more concretely, those tools are, for example, morph logical analyzer, which is an essential tool in language technologies, and on top of it, cannot build other downstream applications.

As we show, we created a language learning platform called (?) where people could upload texts and do exercises, and platform would automatically create, close, fill in the gap exercises where people could exercise their language studying.

So to be on that platform, we needed morphological analyzer.  Besides that, I also work on creation of machine translation, baseline for Sakha, along with other people who work on languages, and on checking all these tools we are working on, most of them we used Wikipedia.  So Sakha Wikipedia was started in 2007, and administrator is Nicolai Pavlokalin.  I am grateful him and other people working on the Wikipedia so that we can use it in our research, because there is a look of digital resources in Sakha.  There is ‑‑ we have newspaper corpus and decided to have Wikipedia corpus and not anymore as far as I know, digital texts, it's nice that we have such a resource as Wikipedia.  Yeah, besides that, there is ‑‑ we work also on national language generation, we plan to work on AI to generate epic poems of Sakha, and also we work on template‑based generation of news.

So in modern digital world, I believe that all these language technology tools will help to support Sakha and I hope that those tools could also be useful for other Low‑resource languages.  Yeah, thank you.

>> SUBHASHISH PANIGRAHI: Thank you so much, Sardana, the next one is Mahir.  From your experience working with the wikimedia community for a long time and using wikidata in many different ways, how do you see wikimedia and Wikipedia projects in general and the initiatives helping Low‑resource languages and the speakers of those languages reclaim their space on the internet.

>> MAHIR MORSHED: Thank you for the question, first to introduce myself, I'm Mahir Morshed, I'm a third year Ph.D. student based in Illinois.  My main focus is first in the use of (?) to feature to improve recognizers for Low‑resource languages, as well as discovering units from ‑‑ for languages for which not much data is available to begin with, which as we all know is most of them.

Which ‑‑ in terms of being able to, speaking for the wick media perspective, I should notice, joint from my research but can be related, speaking from that perspective, I would say there have been a lot of initiatives from the Wikipedia foundation projects to improve accessibility to speakers of less   supported languages.  In a lot of ways, it's one of the primary pillars of the movement strategy for the next decade or so, is being able to provide support, whether it's technical, fiduciary or otherwise in terms of capacity building to these communities.

And for a lot of communities, it has worked out really well, it was mentioned in the introductions, that was a very good use case, but for other languages in general, there is the issue that there are pockets of information that tend to exist on one language Wikipedia and not in the others, it is difficult to surface that information, so it is accessible to everyone.

One ‑‑ the primary way in which this is typically done, is adding information to wikidata, which is a knowledge base, in the public domain or cc0, and is used in a lot of places as a baseline of knowledge.  You can say data form, you can download and parse the database as needed if you have a home machine capable of doing so.

For those consuming information in text, there is the issue of how to convert this information into a readable form, especially to do this across a number of different languages, for which there are a number of initiatives including three (?) data project, capturing information about the words, phrases, low cautions finish various languages, tying this to the ontology that is induced by wikidata, and then using these this information about words to construct sentences, to construct information in a readable form, which is the primary intended goal of the abstract Wikipedia, is a new project being developed right now that intends to make such a ‑‑ such a text generation happen in a sort of procedural fashion.  It's intended to be developed in the same way as other Wikis might in terms of contributing code and produces text based on that.

And I'm fortunate enough to have been active in those two realms of trying to actualize these aspects before the launch of the project.  There's a lot going on in this field.  Subhashish Panigrahi thank you so much, Mahir.  Before we move to the next as you all probably know, the universal declaration of Human Rights emphasizes equitable access.  The function Mahir was talking about in terms of accessibility, which people who have different forms of disabilities, and the challenges that you face in terms of accessing information, but also people that speak different Low‑resource languages and the and when there are different sort of technical interventions on the internet and in general on the digital spaces, not many people are able to access information, not many people are able to participate in creating knowledge and sharing knowledge, many indigenous and endangered languages that fall in the broad category of that Low‑resource languages around the world face a lot of challenges when it comes to the governance.  Internet.  Their participation or the lack of participation, is responsible or is the reason why many people don't have the agency or the ownership of many things on the internet.

So when we think of Wikipedia, that was born from the larger movement of citizen, science and federated knowledge sharing processes, that was essentially sort of a way to envision how the internet could be an open and shared and collaborative space.  So Wikipedia is not just the only project, but there are many projects in the open knowledge movement that were intended for that purpose.

As we have moved many, many years now, almost two decades after the foundation of such projects, or almost two or three decades of foundation projects, we need to think about how the participation of many people are still missing on the internet and how that is affecting the content that we see on the internet when we think of the algorithm sick basis on the internet.  That is the lack of participation of many communities.

We wanted to understand from the perspective of each of our panelists, because that you all bring really unique and nuanced perspective to share with us how that you see communities taking ownership and how low resource communities taking, and exchanging information online, what kind of strategies they are using to do that, and what does the future of the internet look like when it comes to low resource languages and comes to many low‑resource languages, one of the challenges we see particularly with the article languages that don't have any writing system is that there's no way for such languages to have a journal or something that's an authoritative source of information.

How do communities document their oral history and how those oral hurts are sort of verified and are treated the same way you would treat many journals or books and so on.

So I am thinking from that perspective, and the challenges that many communities face, many communities also don't have the affordability to have resources that ‑‑ languages that are dominant languages have access to or communities that speak those dominant languages have access to.  You think that also affects a lot how different communities and different stakeholders make decision on the internet, and I think when we think of a federated structure that we envision to bring to the internet and the communities taking challenge of many decision‑making in that internet space, we have to think about how communities could be equipped with tools and resources and even funding and other resources that could have them sort of become the owner of their own content, and make decisions in just being consumers when they are using the internet.

So I think I want to from the next set of questions around two areas.  One is broadly around the Human Rights, equity and inclusion by using technology and different kinds of tools, the first one is how, say, wider gee graphic collaborations are happening on the internet and the second is how decentralization is happening in terms of mobilizing and sharing and disseminating knowledge and also ensuring participation of different communities, particularly the low‑resource language speakers.  So thinking from that perspective, my first question would be for you, Eddie, how do you envision the language digital activism moving the needle around, furthering access to linguistic rights and to knowledge.

>> EDDI AVILA: Thanks, Suba.  That's an ongoing question whether it is moving the needle, what sorts of impacts it's having.  I think it depends on the change, the type of impact that each community wants to have.  I think some are mutual easier if you want to increase the amount of digital content in your language, you can count the number of pages, counting the number of data available in that language.  When you talk about potentially want to change attitudes and use digital activist to stay with their language, it's harder to measure that.  Obviously when there's so many different factors involved, when there's public policy, whether it's about some of the linguistic considerations that have to take place, whether it's about bilingual education or education in the language, all those factors contribute.  To say that language activism is definitely moving the needle, it's hard to prove, to measure that.  It is part of a wider, you know, shot of thing and attitudes because of anecdotes, I think that's much easier to do.

I think people can say, I've gotten comments to my social media account where people say I'm so glad you're doing this, I'm a heritage speaker, my grandmother spoke it, I'm more motivated to learn it and stay with it.

I think those are things you can observe, but I think it goes back to really defining what it is you want to change and what it means for you.  For some people, if it's creating Tik Toks and Instagram videos in their language, if they can reach one person, it's well worth it.  For some other people, they may want a bigger change, they want to see something much more.  I think it's not so much seeing change and maybe demonstrating impact out of a desire to show part of a development world where you need to have strategic objectives, to prove that it's what you say is everything the effect it is, but rather to maximize one's time because I would say the vast majority of communities and individuals we work with or know are doing that because of their own individual passion and love of their culture and language.  And so it's not necessary to prove something, but rather because resources and time is limited, being able to maximize that limited time, to have as much reach as possible.  I think it's much more about that, not about proving to funders or proving to anyone else that what they are doing is worth their time, but rather for them to feel like what they're doing is having the intended change so that you can potentially as much as possible, avoid burnout or avoid the limited time that they have.  I think that's, for us, is our interest, to help activists reflect on that, but I think anecdotally, yes, we could say it has had an impact, but I think it needs to go beyond anecdotes, a strategy and approach where the activists themselves define what it is to be successful and moving the needle.  I think hopefully we can facilitate the conversations, facilitate highlight examples of other projects that may be using some sort of measurements or impacts approach, but you think at the end of the day, it is up to each activist to decide for themselves what it is they like to plush and whether they feel like what they're doing is part of that solution.

>> SUBHASHISH PANIGRAHI: That's a really important aspect you mentioned about how the time is limited and doing a lot of work from their passion and ask exploring to communities that are already volunteering for limiting time, that is a lot of ask, that human factor is often something that we don't discuss actively.

Talking of that factor, the human factor that goes into sort of revitalizing a language or even actively using a language on the internet, my question to you would be your work around building the transcription toolkit or recording the culture as audio or video or the recent workshop we did together, what do you see as a long‑term shift as a result of these initiatives.

>> AMRIT SUFI: Thank you, as a result of the initiatives, people can have access to resources from other pluses in the world.  So even, you know, speakers of low‑resource languages can work on language federation on their own and activists can utilize the specific skills to contribute to language activism, which I don't know what the response is.  They won't be entirely at the mercy of social, economic circumstances.

For instance, the participants in the languages workshop we are informed that some language activities are using the tactic of normalizing the digital spaces, localization of software and platforms, like Firefox, Wikipedia, and wikimedia projects, some language activists have built articles in their own languages, a lot of articles, and they have got the language successfully out of the incubator phase.  Me and my colleague are working on the transcription cool kit.  We are localizing wikimedia commons and WIKI source by uploading oral history, and oral history videos and folk songs, and transcribing them and uploading them on WIKI of course, that is the ‑‑ sort, WIKI library, and in doing that, diversifying the library.

I think my conclusive statement is see the public being engaged directly in deciding the fate of their language.  So I might sound very optimistic after the input, but that is the hope in the future, in the coming years of these initiatives.  Thank you.

>> SUBHASHISH PANIGRAHI: Thank you, so much, Amrit.  Just the aspect you mentioned about normalizing the use of languages that are otherwise not used publicly, that's really important.  Also, sort of a building block, a starting point for many low‑resource language digital activists.  As you worked for quite a long time with the wikimedia community, Mahir, what in your opinion would be the sort of building blocks for nature of speaker community and the other stakeholders of low‑resource language, and what are the sort of entry level barriers you see in the wikimedia sphere for such communities.

>> MAHIR MORSHED: The barriers will differ depending on, of course, the role within this group of possible members of the language community and others who may have something to gain from proliferating and thriving.

One that I think is common to a lot of individuals is just the technical knowledge that's needed to get around the media WIKI university and how it pervades the projects in a lot of ways, that's one of the main stumbling downs, which is slowly being addressed in a lot of ways.  A there have been toolkits made to make or navigating these a lot easier.  It's successful with the Kurdish communities and a number of other Indian language communities in terms of helping them get around.

For the, I guess, other alternative types of knowledge recordings, such as oral histories, video histories, et cetera, things which are not committed to paper in a published forum, there could be a lot more support for organizing these recordings and making sure they are uploaded, categorized, sometimes transcribed, WIKI source, language WIKI source is attempting to build tools to make transcribing videos easier, it's worked pretty well for many of their use cases, and they are trying to expand that outside.  I think this will be used for Dagbani.  I mean, I mentioned that there has to be some sort of a baseline knowledge of the language itself, not so much as a collection of sentences of which things can be statistically driven, a collection of words with information about them and connections among them such that some meaning can be induced by, you know, going to WIKI and finding a concept and construct the name of the concept from words in that language, whether it's Dagbani, or Angika.  That is one thing that is slowly being improved.  Some other languages have been enthusiastic about getting vocabulary into wikidata system so that Wikipedia later can use it.

There's still, again, this issue of linguistic knowledge, maybe not of the language, but enough to be able to guide communities which speak these languages to better coordinate and direct their efforts such that something meaningful can come out of it.

I disclose I've been trying to put out proof of concept.  I recognize it was difficult depending on how much information about meanings that cannot derived from the concepts that exist algorithmically, on wikidata.  There are difficulties that can and should be addressed and I think will be in due course.

>> SUBHASHISH PANIGRAHI: As we don't have ‑‑ thank you so much for sharing that, Mahir.  As we don't have enough questions from the audience, I was wondering if I could ask all of you one set of questions each, and then I think we can close the session with some thoughts from you, particularly around your recommendations.  Before we do that, I would go in the same order.  The question that I have in mind, when we think of digital language ‑‑ sorry, language digital activism as a domain that includes many kinds of activities.  What do you see sort of the barriers for many communities to actually participate in this event or in these activities, Eddie, particularly from your experience with many low‑resource languages that live in places that might be disconnected from the rest of the world in terms of access to connectivity in their communities that might be speaking the language at home, but outside the home, it would be an official language, the dominant language of the region?  All those kinds of challenges that they face in general, in life, while using the internet, and how does that ‑‑ how do those barriers affect their participation when they have to participate in an online event where the language of medium would be probably a language that is understood by a majority, but are there specific barriers that you think that is responsible for stopping people from participating and taking part in the language digital activism.

>> EDDI AVILA: Sure, thanks.  I think obviously there's about 7,000 different languages around the world and each has their own unique context.

I think generally, in our mapping process, we have identified a number of challenges, probably one of the biggest is accessibility.  I've seen an interesting map, I don't remember who it was from, but they overlaid, you know, where indigenous communities here in Latin America live, and where, you know, stronger internet connectivity is available.  I think it coincided with those places  where these languages are spoken, are usually places where there's poor, little or expensive or relatively expensive internet connection.  I think that's not a coincidence but a really telling sign, this is a really important issue, not just internet connectivity but affordability, how much of one's, you know, monthly wage goes towards communication internet.  Again, obviously there's some ways around that.  People are creating intranets where it doesn't need to be connected to the outside connection, but people can connect through blue tooth or some other means to access information and upload, whether it's educational books in those languages, other communities are creating their own community networks, there's lots of common needs where digital activists and community activists can work together to help address those issues.

You know, slowly but surely, I think a lot of these places will be connected, whether this is next year, ten years from now or 20 years from now.  I think a lot of activists are working towards that goal.  So that when they're coming back home or in the rural parts are connected, they don't want to find an absence of information.  They are working towards making sure that once that day comes, that they have access to information.  Barriers include linguistic barriers, as you mentioned, Suba, a lot of languages are oral based.  So potentially some of the different tools are off limits or not as accessible.  Maybe that's a barrier.  Also consensus about the route way to write it or consensus about a way to write it.  We have seen lots of examples where maybe there's communities that want to take a more pragmatic approach, it's about mixing the dominant language with the native language because more people understand that, even though it may not be as pure where some other people think, yeah, we need to find a word for every ‑‑ a new word for every word that exists.  Technology terms, for example.

Those competing visions may manifest itself in different ways, you know, but also, you know, political and social challenges in some parts of the world, even recognizing that there are indigenous communities is against the law.  So in those kinds of environments, it's already a big uphill battle to communicate publicly, but also online.  Those are the types of things we are seeing.  Some are, you know, easier to maybe address or there's some other recommendations to address those, maybe the linguistic may include two different versions of the same side, one more mixed with the dominant language, another one ‑‑ just recognizing that it's okay that there's different ways to see things, but obviously some of the political implications may be harder to address from an activist perspective, but I think all of those are just having the conversation around those challenges and showcasing ways that communities are addressing those challenges is the first step to go.

>> SUBHASHISH PANIGRAHI: Thank you so much, Eddie.  Those are really important points just made about political challenges being a huge barrier in many places where activism is considered as an illegal way to advocate for languages.  I think we all have seen many forms of that, starting with censorship of a specific kind of content, the voices of many people are marginalized, and people are targeted based on their activism because language being a tool for many activities, and not just promoting the language or culture, it's beyond that, tied to the survival of many, many communities.

Thinking from that perspective, and just going a little bit into the technical solutions, the technical tools that are created, Sardana, as a Ph.D. student, your work is primarily around building different kinds of technical tools, for the Turkish languages.  How do you see different communities being involved actively and taking the leadership role in creating those tools or managing those tools, or sort of taking over the control of those tools?  Do you see that, or do you see technical community being sort of a separate and isolated community that is building different tools and then their users for native language speakers and those speakers are use nothing tools?  What is your sort of experience?

>> SARDANA IVANOVA: Thank you, yeah.  So I think that the people who develop language technology tools should closely work with communities who speak those languages.  As Steven Berg pointed out in his conference, it is very important to work on language technology together with people who speak the language, to ask them what kind of technologies they would need and not take responsibility of managing language of the community and making tools and not asking if they need the tools or not.

So I think that language technology developers and scientists should work closely with communities.

>> SUBHASHISH PANIGRAHI: Thank you so much for sharing that, Sardana.  I think that underlines a very important point about being inclusive and being respectful, but also being welcoming and ensuring that if a technology is being built to support a particular language, then the community has to be involved and the community has to take a front seat, and they have to lead the whole process, whereas the research community or developer community could be in a supporting role, rather than driving that process.  I think that's a very important point that he just made.

My next question would be to Mahir, and based on your work, this could also be your recommendation, what do you see sort of the existing platforms that could include Wikipedia and wikidata, but also beyond that.  As sort of really low entry barrier platforms that communities can start with because it's always overwhelming for any new community member, right, if I'm a new community member interested to start working for my own language, and that is low‑resource language, then it is always very intimidating for me to figure out where do I start from, right, and as you've worked with different communities in different spaces in that spectrum, what do you think in your experience the platforms that have a low‑entry barrier that are easy to understand, someone who has some level of understanding of the internet, of course, that's sort of the basic literacy that is required, but beyond that, what are those platforms that you see as low entry barrier platforms.

>> MAHIR MORSHED: On the wikimedia projects, there's a couple of ways to go about it.  It depends on the perspective that the speaker has, in terms of what they want to do with their language there.  If they want to bring the rest visit the world to their language or make what is organized in information, accessible in the fastest way possible, I think one of the ways to do that is to go on wikidata.  Every concept on wikidata has some name assigned to it in the language and possibly a textual description for those on the go and who want to understand what a concept is just looking at it, and just translate the names into your language.  It's a pretty straight forward activity in a lot of ways.  There are some individuals from Cameroon who have managed to add support for many, if not most languages that are spoken in that country.

So it's possible now to do this translation in all of those languages.  And alternatively, if you're interested in bringing your knowledge as a speaker to the rest of the world, then one way to do that, as I mentioned with oral histories or video histories, record yourself or someone else, or the information you have or produce a text, depends on your circumstance, because, for example, Bali has a lot of texts in different printed forms, and those have been digitized ‑‑ manuscripts, sorry.  They are being proofread at the moment.  If you have an oral history, there are efforts to make transcribing those easier.  That's one way to do it.

Actually, the folks behind the video and audio transcription are very friendly, they can help someone who is interested in getting set up, getting that to work.

Those are, I think, the two main ones that, from my perspective, at least, speaking as someone who doesn't frequently edit Wikipedia proper, but have seen other communities take advantage of that for the better, it depends on your objective and what you're comfortable with from the beginning.  In terms of the written word, the article word, et cetera, or if you want to translate things very quickly.  There's a number of wases to start.

>> SUBHASHISH PANIGRAHI: That so much, Mahir.  Going to Amrit, what would your recommendations when it comes to some of the examples that actually Mahir gave, oral history, documentation, but also from your work, what are your recommendations that worked for you, and you think it could also work for many of the low‑resource languages, particularly languages that don't have enough people to contribute, so it's not just other kinds of resources, financial resources, of course is one of the biggest barriers for contribution, but oftentimes, it's the lack of number of active contributors, and oftentimes, when it's a volunteer‑led project like Wikipedia, there is a challenge for even established languages to have enough number of volunteers.

So I think from your experience, what would be the recommendations.

>> AMRIT SUFI: Yeah, they are a number of volunteers for Angika, there is an article page for it.  As well, but yeah, I think, you know, it's enough ‑‑ there's of course a look of contribute earns in even the major languages, English and other spoken languages.  So languages instead of trying ‑‑ my subjective opinion, instead of trying to build Wikipedia in their own language, they should upload ‑‑ maybe they can upload folk songs of their language or pictures of utensils or ornaments, dresses.  So I was training one of the volunteers in this project, so she is from the community, so there are very small number of people, like there are only 7,000 speakers visit the language.  That volunteer got really interested in that ‑‑ she can do that, she can present her culture on WIKI platform.  She has been uploading pictures of the dresses that they wear on marriages or their various ceremonies, or things like those.  She has also interviewed her grandmother, you know, and speaking her language.

I think representing your culture and representing videos and media from your culture is a good way to increase the visibility of your language on digital platforms.

>> SUBHASHISH PANIGRAHI: Thank you for sharing that.  Eddie, from your perspective, what do you recommend going forward for oral languages or low‑resource languages or people that are working for such languages?  And what has worked?  What is sort of your experience when it comes to some of the stories and learning from those stories?

>> EDDI AVILA: I think as we head into the international decade, you think this is really a golden opportunity to amplify the work of activists, to really showcase what they're doing on a local, national and international level.  You know, slowly there are more opportunities for them to connect with policy‑makers, I think especially around some of the committees or groups around the international decade.  Still a lot of work to be done so there's not that gap between grassroots activists and policy‑makers that cannot make decisions to make change.

It's a slow process, I think this is a good opportunity to may be catalyze that, some of the people we worked with, six or seven years ago have taken on leadership positions in their community or even at a local government and national level.  So I think those types of things are fewer examples of that, but hopefully as more people develop not only skills with digital activism, but about organizing, about understanding the challenge, hopefully more opportunities for them to be in a position to make decisions that could affect more people.

Finally, I think the downtown create spaces for peer learning and exchange, I think it's very powerful when another activist from another language community shares their work, and has someone come up to them saying I want to do the same for my language, how can you help me, or how can I do that and have a place to direct them, hopefully the toolkit project we are working on provides a place for them to get more information and more places where they can go.  Finally, that in this work, it's really important that these are activities that the community themselves, they want to see instead of an outside body can say, you need to do this in your language, create this in your language, but rather really comes from the bottom up or from a grassroots up, and hopefully there's systems in place or communities in place, networks in place that can be supportive so that you cannot really build off their excitement and enthusiasm, and it doesn't lead to anywhere, but hopefully there is a path or many paths that can really help them achieve what they want to achieve.

>> SUBHASHISH PANIGRAHI: Thank you so much, Eddie.  As we are almost at the end of the session, I'd like to first thank all the panelists for sharing your inputs and such valuable insights around your work and the challenges that many of the communities that you work closely with are facing.  Just to summarize, we started with the different kinds of communities that are facing challenges on the digital activism front, and we discussed about how some of the communities have managed to create resources and then able to share information with each other and share learning with each other and support each other.  So the peer learning part is something Eddie highlighted a lot.  We learned from Sardana about the Sakha community and Turkish language community’s researchers are working together to build technical resources.  Particularly around NLP, the natural language processing or artificial intelligence and so on and so forth.  We learned from Mahir, a Wikipedians, about several new initiatives, several ways to participate, especially communities that are smaller communities, that don't have enough people to contribute for the protection and growth of their language, how there are different ways to contribute using Wikipedia and wikimedia platforms, such as contributing for translation of description on wikidata, or creating a dictionary using lexi.  These are low‑entry barriers.  And I think ‑‑ all these inputs are really valuable and thank you also, Amrit for sharing about your own experience with oral culture documentation.  Thank you all of you again, hopefully it was useful for the participants.  But if any of you have questions, please reach out to us and we would love to share from the learning of our panelists.  Thank you so much.  Bye.