"It just looks wrong" is the best I can come up with, but I think it's that they tend to lose detail along the way.
Chinese pinyin is a good example - not only is it only an approximation of the actual sounds, there are so many homophones that you end up compressing tens of characters into one.
For languages written with an alphabet, like Russian, there are romanization systems that exactly preserve all the available information.
And there is even Serbian, which can be (and is by native speakers) written in either Latin or Cyrillic, with a 1:1 correspondence between characters.
Fun fact, since both alphabets are tought early in school and used all the time, most adults can read both of them and even switch mid-sentence without noticing a difference. This also means that we're a bit "blind" about traffic signs or directions/instructions which are in Cyrillic only - we don't notice anything strange but it's a big difference for foreigners!
This is not limited to Serbia. For instance when travelling to Greece it's much better if you know how to read Greek alphabet since not everything is presented in Latin.
While they are certainly digraphs, they are regarded as a single letter and not a 2-character sequence. They have their own sound, sort order, Unicode designation, and written orthography. For example, on advertising where a word is written vertically, 'lj' will not be separated vertically, and a hyphen never separates the 'l' from the 'j'.
However, since the letters 'n' and 'j' already exist on a keyboard, it's easier in this electronic era for people to type the letters separately instead of hunting for the 'nj' key, so the presence that you see of two character sequences to represent those letters is a consequence of the compromise of modern electronics and expediency, not innate to the alphabet itself.
TL;DR: Each letter in Serbian Cyrillic maps 1:1 to a single letter in Gaj's Latin alphabet, as each alphabet was specifically designed such that each character represents exactly one phoneme.
Adapting alphabets to languages has a long history: just look at how the Greeks butchered the Phoenician writing system with weird concepts like "vowels" and "F". This is something that makes languages unique, and it should be chosen over having digraphs or diacritics.
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.
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.
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.
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, ....
Either way, any alphabet change that leads to less correspondence between written and spoken language doesn't seem like a desirable thing...
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.
https://youglish.com/pronounce/rose/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.
(Note: One wonderfully weird thing about English is that different types of phonological assimilation happen in different directions, e.g., voice assimilation happens the other way, so "fads" is pronounced as if it had a "z" at the end and "putz" is pronounced as if it had an "s" at the end, i.e., "fadz"/"puts" rather than "fats"/"pudz")
Could you please explain how you would represent these different sound and letter combinations in Latin script in a way that would make it obvious how to pronounce them correctly:
ща(вель), ша(тун), сча(стье);
че(й), чё(рный), чье(-то);
This is just two random, very limited examples off the top of my head, not even the tip of the iceberg, but the point of the tip.
By the way, Russian also has many regional dialects and accents.
A latinized Russian should have a different character for each s-like character like it's now in cyrillic.
Also, everybody should look at English and French spelling as something definitely NOT to do.
Spanish is also good but they are actually close to latin so it's no surprise the latin alphabet also works.
German is not so great as the match between spoken and written is quite loose. Both wovels and consonants change with context.
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.
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.
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'.
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.
Your point about computer systems is true, but I think they’re slowly getting better at representing diacritics over time.
No language perfectly matches pronunciation and spelling, largely because of regional differences and changes over time. And yet you can spell the majority of Bulgarian or Serbian correctly with the Romanian spelling system.
One interesting thing is that as Cyrillic has spread, new local characters have been added depending on the language. Eg, Ukrainian has letters than Bulgarian doesn’t use. It’s pretty neat.
Isn’t this true of Latin characters, too? German has ß; French doesn’t. Inversely for ù.
This is an understatement of how complex the Vietnamese writing system is (I speak not a single word).
https://en.wikipedia.org/wiki/Vietnamese_alphabet
They don't just have diacritics, they _stack_ them to form towers of diacritics. Ok, so only 2 high, but even so.
Like many “interesting” philosophical questions, its grounded in a false premise; the description of alphabets derived from the Latin or Cyrillic with subsequent evolution and possibly admixture as “Latin” or “Cyrillic” is imprecise; its a letter of a particular Ukrainian alphabetic of mostly Cyrillic derivation but some Latin influence. (This can be simplified either of the ways siggested by the question for purposes like defining international character sets, but then it is more of a practical question than philosophical.)
The biggest difference is that Latin scripts tend to create new sounds by adding diacritics or digraphs while Cyrillic scripts tend to create new letters — IMO the second way is better for non-Slavic languages.
But it was invented too late, and lacked an expansionist empire, or a widespread religion, to bring it to enough neighboring peoples.
Ц -> Z
Ш -> Sch
Ч -> Tsch
one could argue that german should also have one character for the "sch"-Sound instead of 3.
The postfix-h notation was a good idea, but the fact that e.g. "ch" is interpreted differently in English, German, and French derails it instantly :(
OTOH to type fast, you might want a more ergonomic input mechanism. (But I won't dive into the rabbit hole of really ergonomic keyboard layouts; if you're curious, search for KMonad, and think about e.g. the convenience of having modifier keys right in the home row.)
I happen to write in Spanish and German, for which the us-international layout is just perfect. But I did customize my xkb for some other keys, and I used to use kmonad before.
No need to strike for DOS compatibility anymore.
Transliterating Kazakhstani Cyrillic to a Latin alphabet won't change the underlying information available in the characters, it will only change the way the characters are being presented. No detail lost.
That being said, I am wistful of the change.
~100 years ago, Turkey switched from Arabic to Latin; you can study the impact.
Even though latin Turkish alphabet is a bit problematic for software (I -> ı capitalization issues) I think Uzbec and Kazakh alphabets avoided this issue.
While in, say, Swedish, letters like Å, Ä and Ö (and, to this non-swede, arguably W) are diacritically-marked letters that became fully-fledged ones, like mitochondria becoming "part of a cell".
Actually the swedish alphabet is remarkable to me in that it preserves many otherwise useless letters like Q and Swedish orthography is quite tolerant of the use of non-letters like á ("non-letter" meaning "not part of the official alphabet").
The only diacritic used in English, the diaresis, is clearly completely optional and I suspect many people go through their entire lives without encountering it!
Sometimes they got it wrong: "island" comes from Old English and never had an 's', but it sounds close enough to the Latin "isla" that they just tacked the letter in there.
I think you have misunderstood something. Reading the pinyin without the tone markers (or worse, as-if it was English) would be an approxmiation (and tone sandhi isn't written), but otherwise it is what the characters sound like. How do you think kids in China learn Chinese today?
But rest assured, modern Chinese language having tons of homophones, itself, has nothing to do with the Latin script. They always had them.
On the other hand. No language rules forever. See lingua franca.
Depends on the source script. For Cyrillic->Latin the changes are minute. Most characters have a 1:1 mapping.
Cyrillic also has a couple soft vs hard annotation single characters, which I'm not able to explain well in terms of their cousin western European but still Indo-European languages.
So you get horrendous spellings if using Latin characters for the Slavic sounds, like you see in Polish.
(Disclaimer: i'm a native English speaker who had years of western european linguistics, also took Russian and had a Polish grandfather).
I would consider that as 1:1 mapping while technically being 2 characters in most latin languages.
I'm not sure where you get this from.
Several Slavic languages (Czech, Slovenian, Croatian, etc) use Latin-derived alphabets and managed to preserve their pronunciation with 100% fidelity. The Polish orthography is not the only solution to this, and I think it's a stretch to regard Croatian spellings as "horrendous". I am, of course, biased. But the number of characters in a word in Croatian maps 1:1 with its Cyrillic spelling, for example.
I mean they use multiple letters for a sound (or diacritics which is neater in my opinion) is all I'm saying, like modern english does for voiced and unvoiced "th" whereas older germanic had eth and thorn glyphs, and iceland still uses them.