4 ms·

> For languages written with an alphabet, like Russian, there are romanization systems that exactly preserve all the available information. As a Russian speake

> For languages written with an alphabet, like Russian, there are romanization systems that exactly preserve all the available information. As a Russian speaker, I must say that while this is technically possible, in reality it's only about 80% in most cases, and there is still ambiguity and lack of clarity in the sound transcription, unless you know the original Cyrillic spelling and pronunciation. Russian has more letters. And Russian has sounds and constructions which are just not represented in Latin alphabet at all. For example, how do you differentiate between ё and йо, both valid constructions in Russian, with slightly different pronunciation, usually represented Latinized as yo? Not only that, but the soft and hard signs are also difficult to represent and differentiate between similar letters. You have to start using apostrophes for the soft sign, and then another character for the hard sign, and before you know it you might as well be writing IPA. It's a beautiful language with a beautiful alphabet and script with a lot of beautiful literature to go with them that work well together.

rippercushions 4y ago
This problem is not limited to Russian, and is in fact much worse for other languages like Thai, where the spelling of words is only loosely related to the pronunciation. You can optimize for accuracy (transcription) and lose the phonetics, or you can optimize for phonetics (transliteration) and lose the ability to reconstruct the original script, but you can't do both.
throwaway22032 4y ago
Right. At least щ, ь, and ъ I don't think have any Latin analogue (native english speaker here who has studied Russian a decent amount). A good way to illustrate the difference is to try to Cyrillicize Latin script. Дринк and Drink are just not pronounced the same way. Chinese speakers just make up names for Western stuff entirely because it can't be said. McDonald's is Maidanglao.
- Shorel 4y ago
  This also has more to do with the quirks of English pronunciation than something particular to Latin scripts. Like Mercedes having each e pronounced differently. It's madness. An example: Limonada in Spanish and Лимонада in Bulgarian sound the same, and mean the same.
pierrebai 4y ago
The flaw in your argument is the implied presumption that Latin preserve all information for languages that have always used the Latin alphabet. For some language it is closer to truth, like Spanish and German, which mostly are pronounced like they are written English and French, identical spelling represent varying sounds. English example: lose, rose, tone,... all different "o" sounds, yet they all use the same written "o". The point about Cyrillic having more letters is also misguided: you can make sounds with letter combinations. French has dozens and dozens of letter combos for various sounds, let just list a small subset: au, eau, eu, ou, ui, oi, euil, ouil, un, an, in, en, oui, ui, sh, ch, cl, cr, tr, ....
- wizofaus 4y ago
  I'm curious, I use the same "o" sound for "rose" and "tone", do many native English speakers not? But "not" has a different o sound. As does "ton" and "curious" and "book" (I'm not sure if there are any words that represent the "book" vowel with just a single "o" - though sometimes it's "ou"). Either way, any alphabet change that leads to less correspondence between written and spoken language doesn't seem like a desirable thing...
  - torstenvl 4y ago
    The vast, vast majority of English speakers will nasalize the o in tone, and the vast, vast majority of even the small fraction who don't will still pronounce the o farther back than they would for rose. Try this. Pinch your nose closed with your thumb and forefinger. Say "tone" but stop before you articulate the n. Now do the same with "rose," stopping before you articulate the s. I'm betting you felt a lot more vibration in your nose on the first one. There might be a small number of heavy-NCVS speakers for whom there is truly no difference. David Boreanaz, perhaps. Bobby Generic's mom, likely.
    - wizofaus 4y ago
      Can't say I detect a difference. Might be an Aus-E thing: https://youglish.com/pronounce/rose/english/aus https://youglish.com/pronounce/rose/english/aus https://youglish.com/pronounce/tone/english/aus https://youglish.com/pronounce/tone/english/aus But I would have thought I use -əʊ- for both, which is what's shown as the Br-E pronunciation for both words on that site.
pointlessone 4y ago
Is there really an ambiguity there or is this just a spelling quirk? Like are there any two words spelled differently in Russian that can be mixed up when transliterated? I imagine, there's an official transliteration scheme because names in international passports have to be spelled in latin alphabet for reference. Is it ambiguous, too? Even if so, it's probably not to hard to come up with more two-letter combinations to encode every letter of Russian alphabet in an unambiguous way and so that even people who doesn't know the encoding could read it close enough.
- forgotmypw17 4y ago
  Yes, there is ambiguity if you want to achieve both readability and pronunciation accuracy. If you want to just encode the "data" or just get close (but not exact) pronunciation, you can do reasonably well, though nowhere near as good as Cyrillic. If you want full parity, you just can't. Not unless you rewrite the entire language's grammar and re-teach people how to read Latin script in a way that is very different from how it is read in other languages.
- d1sxeyes 4y ago
  It's similuh to riting inglish pyoorlee fuhnetiklee. For 'two letter combinations', etc. it's worth looking at languages like Polish and Czech, which are fundamentally Slavic, and look at how the Catholic church had to bend the Latin alphabet to fit the sounds we simply don't have Latin letters for. For example, the Polish 'szcz' (as in the town name Szczecin) can be represented as a single letter in Cyrillic (щ, as another commentor has highlighted). You don't have to look far to find jokes/criticisms about Polish names being 'too full of consonants and not enough vowels'.
- yaky 4y ago
  Transliteration rules exist, but they change. Me and my sibling have a different latin last name due to this. As an interesting side effect, you can tell Ukrainians by the way their names are transliterated - it’s Iurii (not Yuri), Olha (not Olga), Hryhorii (not Grigori) etc.

rippercushions 4y ago

This problem is not limited to Russian, and is in fact much worse for other languages like Thai, where the spelling of words is only loosely related to the pronunciation. You can optimize for accuracy (transcription) and lose the phonetics, or you can optimize for phonetics (transliteration) and lose the ability to reconstruct the original script, but you can't do both.

throwaway22032 4y ago

Right. At least щ, ь, and ъ I don't think have any Latin analogue (native english speaker here who has studied Russian a decent amount). A good way to illustrate the difference is to try to Cyrillicize Latin script. Дринк and Drink are just not pronounced the same way. Chinese speakers just make up names for Western stuff entirely because it can't be said. McDonald's is Maidanglao.

Shorel 4y ago
This also has more to do with the quirks of English pronunciation than something particular to Latin scripts. Like Mercedes having each e pronounced differently. It's madness. An example: Limonada in Spanish and Лимонада in Bulgarian sound the same, and mean the same.

pierrebai 4y ago

The flaw in your argument is the implied presumption that Latin preserve all information for languages that have always used the Latin alphabet. For some language it is closer to truth, like Spanish and German, which mostly are pronounced like they are written English and French, identical spelling represent varying sounds. English example: lose, rose, tone,... all different "o" sounds, yet they all use the same written "o". The point about Cyrillic having more letters is also misguided: you can make sounds with letter combinations. French has dozens and dozens of letter combos for various sounds, let just list a small subset: au, eau, eu, ou, ui, oi, euil, ouil, un, an, in, en, oui, ui, sh, ch, cl, cr, tr, ....

wizofaus 4y ago
I'm curious, I use the same "o" sound for "rose" and "tone", do many native English speakers not? But "not" has a different o sound. As does "ton" and "curious" and "book" (I'm not sure if there are any words that represent the "book" vowel with just a single "o" - though sometimes it's "ou"). Either way, any alphabet change that leads to less correspondence between written and spoken language doesn't seem like a desirable thing...
- torstenvl 4y ago
  The vast, vast majority of English speakers will nasalize the o in tone, and the vast, vast majority of even the small fraction who don't will still pronounce the o farther back than they would for rose. Try this. Pinch your nose closed with your thumb and forefinger. Say "tone" but stop before you articulate the n. Now do the same with "rose," stopping before you articulate the s. I'm betting you felt a lot more vibration in your nose on the first one. There might be a small number of heavy-NCVS speakers for whom there is truly no difference. David Boreanaz, perhaps. Bobby Generic's mom, likely.
  - wizofaus 4y ago
    Can't say I detect a difference. Might be an Aus-E thing: https://youglish.com/pronounce/rose/english/aus https://youglish.com/pronounce/rose/english/aus https://youglish.com/pronounce/tone/english/aus https://youglish.com/pronounce/tone/english/aus But I would have thought I use -əʊ- for both, which is what's shown as the Br-E pronunciation for both words on that site.

pointlessone 4y ago

Is there really an ambiguity there or is this just a spelling quirk? Like are there any two words spelled differently in Russian that can be mixed up when transliterated? I imagine, there's an official transliteration scheme because names in international passports have to be spelled in latin alphabet for reference. Is it ambiguous, too? Even if so, it's probably not to hard to come up with more two-letter combinations to encode every letter of Russian alphabet in an unambiguous way and so that even people who doesn't know the encoding could read it close enough.

forgotmypw17 4y ago
Yes, there is ambiguity if you want to achieve both readability and pronunciation accuracy. If you want to just encode the "data" or just get close (but not exact) pronunciation, you can do reasonably well, though nowhere near as good as Cyrillic. If you want full parity, you just can't. Not unless you rewrite the entire language's grammar and re-teach people how to read Latin script in a way that is very different from how it is read in other languages.
d1sxeyes 4y ago
It's similuh to riting inglish pyoorlee fuhnetiklee. For 'two letter combinations', etc. it's worth looking at languages like Polish and Czech, which are fundamentally Slavic, and look at how the Catholic church had to bend the Latin alphabet to fit the sounds we simply don't have Latin letters for. For example, the Polish 'szcz' (as in the town name Szczecin) can be represented as a single letter in Cyrillic (щ, as another commentor has highlighted). You don't have to look far to find jokes/criticisms about Polish names being 'too full of consonants and not enough vowels'.
yaky 4y ago
Transliteration rules exist, but they change. Me and my sibling have a different latin last name due to this. As an interesting side effect, you can tell Ukrainians by the way their names are transliterated - it’s Iurii (not Yuri), Olha (not Olga), Hryhorii (not Grigori) etc.