[yorwba]

Having few unique syllables doesn't mean tones are required, since syllables can be combined. Most Mandarin words are disyllabic or longer, and 400×400 = 160k is enough combinations for a quite large vocabulary.

Unique characters being required to distinguish homophones in modern written Mandarin is mostly a circular effect due to the characters already being available, so people use them in ways that would be ambiguous when read aloud (as intentional puns or simply to be more concise.)

If there had been no preexisting writing system and written Mandarin was a simple transcription of spoken Mandarin, introducing characters would be about as helpful as indicating the Indo-European roots of words in English writing, which is to say that some people might get a feeling of epiphany after realizing the connection between seemingly disparate words, but it would hardly be practical for everyday use.

[Umofomia]

Evidence that Chinese can be perfectly understandable written without the use of characters can be seen in the Dungan language (https://en.wikipedia.org/wiki/Dungan_language), which can be considered a dialect of Mandarin Chinese, but is written in the Cyrillic alphabet.

> Unique characters being required to distinguish homophones in modern written Mandarin is mostly a circular effect due to the characters already being available, so people use them in ways that would be ambiguous when read aloud (as intentional puns or simply to be more concise.)

Indeed, because of the way Dungan is written, it ended up evolving differently with respect to how new vocabulary is derived, often borrowing words phonetically from Russian instead of constructing them from Chinese morphemes that might otherwise be considered ambiguous when used individually.

[carlmr]

>Most Mandarin words are disyllabic or longer, and 400×400 = 160k is enough combinations for a quite large vocabulary.

While true, I'd bet that some combinations dominate because they sound better/are easier to pronounce.

Also just because you can technically differentiate 160k sound pairs doesn't mean you can do it in a noisy environment.

Japanese and Korean have a similarly limited number of syllables and have very long words compared to English. I'm guessing because they don't have tones.

If you look at communication theory you don't only need distinct sounds, you also need error correction. Which requires extra bits of redundant information.

Tones just make it possible to carry extra bits.

Longer strings of syllables like in Japanese and Korean do the same.

More complex syllables, like in English, too.

It's just multiple different ways of carrying enough bits in speech to work in a noisy environment.

Another analogy could be password strength. You can have a very long numeric password (Japanese & Korean), A password with a mix of a-zA-Z0-9 of medium length (English). A password with weird special characters but shorter (Chinese), and they all end up having the same entropy (given that the password rules are known to the attacker).