Conversation

The likelihood of lexicographers like this (104 year old G. Venkatasubiah, who spent 54 years on this) emerging in the future with sufficient frequency for the needs of a medium-sized, slow-evolving language like Kannada (35m speakers, perhaps 50% of them literate) are... low.
Quote Tweet
The man behind a 9,000-page, eight-volume Kannada dictionary that took 54 years to write scroll.in/magazine/83022
3
30
I speak Kannada poorly, with a very weak vocabulary, heavily contaminated with Tamil loan words. And I am illiterate in it. My father made a half-hearted attempt to teach my sister and me to read in the language one summer when we were kids but it didn’t stick.
2
7
My parents themselves are barely literate in Kannada, having had early education in Tamil and adult lives in Hindi areas. I’d say they’re both equally literate in English and Tamil, with Hindi a weak second, Kannada a weird third (stronger vocabulary, weaker literacy than Hindi)
1
5
They largely speak a Tamil-influenced dialect of Kannada at home and with family, consume largely Tamil television, and read a mix of English and Tamil books and magazines, with the occasional Hindi show/movie. I’d say that’s a typical multilingual mix for their generation.
1
8
My sister and I are English-first, Hindi a distant second (I rarely read anything longer than a tweet in Hindi these days), weak+illiterate in Kannada, and can fumble around in Tamil to get around. I’d say language competence has eroded from 3.5 to 1.75 in 1 generation.
1
6
In my teens I used to occasionally feel guilty about not doing my bit. In my 20s I got all Darwinist about it and decided languages that couldn’t keep up with the winners (English, Spanish, Chinese?) with tech-modern high culture + demographics on their side deserved extinction.
1
11
In my 30s, I decided some mix of computing, social media, crowdsourcing, and AI translation would give 2nd/3rd tier languages a shot at survival. I also began to believe that street/low-culture was enough for linguistic vitality and that a high-culture element was unnecessary
Replying to
I was in particular very suspicious of authoritarian high-modernist attempts to “preserve” languages and a high-culture ethos around judging linguistic value. French is a familiar example. But modern “official” Hindi is even better, though not extreme as modern Hebrew.
2
6
Now I’ve landed somewhere in the middle. Elements of high-culture work, like the OP link to a life work in dictionary building are high value for languages *if* there is also a viable low-culture living language for it to ride on. Otherwise it is dead language scholarship.
3
8
High culture as a sort of authoritarian cultural policing force on a dying language only hastens its death, as younger speakers chafe and defect to more laissez-faire languages. But high culture as a sort of high technology for doing more complex cultural things, that’s necessary
2
7
A language without an esoteric tier will soon degenerate and die in a different way: by becoming a language mostly for stupid people thinking stupid thoughts. A preliminary stage is a divide of growing mutual unintelligibility emerging between high/low versions of a language.
1
6
This divide is starkly visible in all Indian languages (a bad thing) but barely visible in English (a sign of health). A scholarly speaker speaking “shudh” (pure) Hindi sounds almost like Sanskrit and incomprehensible the way Latin does to average English speakers.
1
9
Conversely, a full-blown use of a deep-slang street version of Hindi, like Bambaiyya (think strong Geordie or AAVE in relation to English), can be almost as unintelligible to people schooled exclusively in official-sanskritized “high-culture” version, like many South Indians are.
1
7
Prediction: tailwinds of digitization, auto-translation, robust street use, and demographics notwithstanding, most 2nd tier languages are going to die of stupidification because they are below critical mass of creatives working in high culture and keeping it close to low culture
2
8
Coda: I found a translator partner for German version of Breaking Smart easily. In French, I was approached by a French digital agency. In Spanish, I had a couple of individuals interested and an incomplete translation exists it the project didn’t get done.
1
2
This tells you the limits of informal, extra-institutional translation possibilities. Traditionally published books that are reasonably successful will easily appear in 6-7 languages. Self-published ones like my book Tempo (~5k copies, so low end of “successful”) are unlikely to.
2
4
A lot of the value of existence of institutionalized high culture layers in languages is: cultural globalization. When high culture dies, books stop being translated into other languages as the translation gap between self-published and traditionally published demonstrates.
2
1
Web content is somewhere in between. It benefits from improving machine translation capabilities (I don’t know how much that relies on traditional dictionaries etc), but as Breaking Smart demonstrates, there is a limit to how much indie publishers can access quality translation.
1
2
Most stuff is barely worth reading for native speakers in the language, but the best stuff is not only worth translating globally for maximum reach, it might even be valuable enough to inspire non-native speakers to learn the language.
1
3
It’s telling that I was unable to find a good translation resource for breaking smart —> Hindi. Then I started doing it myself but then gave up as not worth the effort. Didn’t even bother looking in Tamil or Kannada (where doing it myself is not an option)
1
5
Inference: speakers of any language it’s worthwhile translating material like breaking smart into are already likely to speak English. It’s like it’s easiest to get a loan if you can prove you don’t need one. It’s easiest to get translations to languages that don’t need them.
1
4
Ironic endnote: this thread was inspired by OP profile of a Kannada lexicographer who spent 54 years writing a dictionary. Kannada is native language of Bangalore, perhaps the best-known non-western city associated with software eating world. E(P(software saves Kannada))=0.05.
1
6