Today, the ability to write or read in your language or use your language on the digital medium is entirely at the mercy of the operating system, says Vivekananda Pani, co-founder and CTO of Reverie Language Technologies.
The internet and digital technologies have always been uneven spaces in India, with vast amounts of content, innovation and infrastructural support being available for English, and perhaps a handful of Indian languages. However, the development of Indic languages on the digital medium has seen a resurgence in recent times, with the advent of cheap mobile phones and data plans.
We spoke to Arvind Pani, CEO and Vivekananda Pani, CTO of Reverie Language Technologies — a Bengaluru-based company that creates Indian languages solutions for businesses — on what ails the growth of Indian languages, and what the government can do about it. We also talked about what role Reverie is playing within the larger Reliance Jio family; Reliance Industries acquired a majority stake in the company in April 2019.
Below are excerpts from the conversation. Please note that the quotes are edited for the sake of clarity and brevity.
On what’s ailing the development of Indic languages: The bottlenecks
— Dependence on operating system
Vivekananda: Today, the ability to write or read in your language or use your language on the digital medium is entirely at the mercy of the operating system.
“If you’ve bought a phone or a computer that supports your language — which may be Telugu, Kannada, Malayalam, Odia, Bengali, whichever — if your phone doesn’t have that support, you have to basically accept that you won’t be able to use it in that language. And, even if there is a good way to support languages, there is no way it can be added to the device.”
—Some Indian languages will die if they don’t go digital
Vivekananda: People rejoice when Google adds a language to Translate or anything else. But the truth is far graver: India has a lot more languages, and a particular company cannot really focus on all of them. The company will not, therefore, service them and these languages will be dead in 10–15 years. These people will go digital, but they won’t be able to use digital in their own languages. Therefore they will have to forget their language.
— Prohibitive platforms
Vivekananda: Another aspect is that though platforms have been undemocratic, their technologies have also been quite prohibitive. Technology is supposed to make something easier, whereas the technologies in the use of Indian languages have been quite prohibitive.
“Let us say you are trying to compose a blog for yourself, or maybe an article, you have very few chances of selecting a beautiful font for yourself, a different style for the heading, or even do a good spelling and grammar check. This is because the technology to develop fonts has been made so complex that font designers or calligraphers are not able to develop them anymore. However, these problems will solve problems in India before the advent of open-type formats or Unicode which started coming in default in the operating system.”
— No true equality of access
Arvind: The topic of prevention of growth of Indian languages only started coming up when we forayed into digital. Out of digital, or the internet, you pick any media — print, radio, cinema or television — we never talked comparison between languages because there is true equality of access. Indian language users or English language users don’t have to do anything different to be able to access this media. Unfortunately, this is not true for digital or the internet.
“India still does not have a standardised keyboard, that we can say is there for our languages, scientifically designed and developed taking into account how kids learn to write in them in primary school. Secondly, English digital literacy is introduced in the primary school system. But nobody learns the digital usage of Indian languages in primary school.”
— India not represented well in standardisation bodies
Arvind: When you take standardisation bodies, predominantly Unicode — Indian does not even have a voting member there. This is the second largest internet user base in the world, yet we do not have any say in the largest standardization body. So it is time that India really stands up and influences the standards that we are adopting. It shouldn’t be the other way around. India should define its own standards and give it to international bodies to ensure that these are enforced like other language standards have been.
On whether businesses have incentive to push Indic languages
— Companies have no interest in supporting languages
Vivekananda: Let’s take Odia, my mother tongue, for example. Odia and Assamese rank may be ninth and tenth in terms of absolute numbers. Therefore, companies don’t find an interest in supporting such languages.
“A user becomes attractive only when the they are able to generate enough business. And the fact is that Indians themselves have lost so much respect for their languages that it has been easier for companies to downplay Indian languages. So, as long as users are able to get their job done, albeit with difficulties, companies won’t really bother. India has 22 official languages, but why would a company care about all of them if the user base is accessible with maybe just five or eight languages? Great usage by users is not the focus for companies, because there is no alternative.”
—Misplaced priorities of companies
Vivekananda: If you read Google’s blogs, they have been emphasising on language facilitation for transactional usage. They say that this is what people do and need on the internet every day. But how many transactions do you really do on the internet? On the other hand, how much do you read on the internet every day? I probably pay bills five times a month using the internet. But reading something on the internet, I am doing it almost every minute. So writing, reading and learning — this are the primary usage of the internet.
On what went wrong on policy front
—India never focused on its languages
Vivekananda: I think it was a huge policy decision that got misplaced on our side. Our country never focused on our languages, which is the reason why, although India itself created its solutions for Indian languages in the digital platform — way earlier and very mature ones, the Indian government just did not mandate that the software that are being sold in India should follow the Indian standards or at least be compatible with the Indian software industry.
“The Internet that came in was very happily accepted and kept completely dependent on the Americans of the world to decide when and how they would want to facilitate Indian languages. This was something that the industry couldn’t have done. This was entirely dependent on policy because Indian government itself has been the largest customer for these companies.”
On what can be done now
—Make platforms more democratic
Vivekananda: What can be done today is basically to make operating system platforms more democratic, to be able to add language features on to them. That is a basic ask. It can’t be forced by the market, and can only happen through policy. If that happens, only then will a large number of languages find better usage on these platforms.
—Introduce digital literacy in primary schools
Arvind: Introduce digital literacy in primary schools in other languages, as has been done in English. Second, standardise character sets for Indian languages. Whatever characters are taught in the schools, take those as out character set. Primary education is possible in Indian languages. The reason why we are finding this difficult is because we never allowed that ecosystem to flourish, whereas the English ecosystem was allowed to. So obviously, there is a gap. There must be a starting point. Third, we need to standardise the input method. And, finally we need to democratise the operating systems.
COVID effect on Indic languages
Vivekananda: We see that there has been a significant focus on providing solutions online. We have had unprecedented requests for language solutions, especially voice. There has been a huge move to move to online services, which were not ready for local languages and hence posed a huge challenge. There were a lot of people who kept thinking that Indian languages aren’t necessary. The pandemic has forced them to change this thinking.
How Reverie fits into Jio family
—’Synergy’ between the companies
Arvind: It is a strategic alliance that made sense for us. Reliance is also an Indian company solving a problem for the same user base that we are targeting. And, in a way, inclusion is what they are working towards.
“There are three major pillars for digital inclusion. One is infrastructure, which is the pervasiveness and affordability of data. Second is the channel, which is affordable touch mobile phones using which internet experience could be rich. And third, the language. Jio has done a lot of work on the first two, and Reverie has been working on the third pillar and that is where the synergy stands.”
— Set top boxes and mobiles
Arvind: There are various channels, such as set top boxes and mobiles. We are working with Jio on their roadmap for these applications. If you take a typical set top box, we’re improving the entire experience of using a set top box to be able to, say, change your channel. May be at some point, they [Jio] decide to integrate that entire shopping experience into the box. So all of those things, Reverie can do that over voice.”
Vivekananda: Reliance has a pretty large ecosystem, and language usage in these mediums — reading, writing and speaking. We are developing solutions in all three experience areas.
Focus areas for 2020 and beyond
—Speech and sentiment analysis
Arvind: Speech is one technology that we have been working on and our ability to address more number of use cases. Even for English, there are two challenges. One is sentiment analysis that we are working on. Apart from that, Anuvadak is a new product that can handle dynamic data conversation.
—Translation industry
Arvind: We have also come up with product called Prabandhak, entirely targeted to translation industry. Translation industry in India is very fragmented. It faces three challenges; [it is] unorganized, unscalable and unreliable. If you are a business and you want to get to a scale where you need to translate 10,000 words a day, you may not find a single source to get started. And you have to put together your own processes in order to do that. The Prabandhak as a platform is a marketplace backed by our technologies that takes care of both the supply and demand sides.
— Goals: 500 million daily interactions
Arvind: Our aspiration is that in the next one-and-a-half years, our technology should reach about 500 million people on a daily basis. I am not talking about unique users; if a user has interacting with banking and government services, these would be considered two instances. We don’t really have the means to track individual users.
Also read:
- #NAMAindic: What’s Holding Up Travel And E-Commerce Companies From Working Indic Languages? (2017)
- #NAMAindic: Why Isn’t There Enough Indic Language Advertising Online? (2017)
- #NAMAindic: Why And How Indic Languages Can Thrive On The Internet (2017)
***Correction (12:26 PM, February 9): Changed name to Vivekananda Pani (instead of ‘Vivekanand Pani’) and name of company to “Reverie Language Technologies”. Originally published at 12:22 PM, February 8.
