Linguists divide Chinese into about ten varieties that are not mutually intelligible, although some count with fewer and some with many more. They are not “dialects”, they are more comparable to the Romance languages – they have a common origin, but are no more similar than Portuguese and Romanian. Mandarin and Cantonese are the most well-known, but there are many more.
And that they have a common written language is both a stroke of genius and a curse.
Written Chinese builds on ideograms. One sign corresponds to one word, with no regard for pronunciation. The words were originally small pictures of the word. If English had a similar system, the word “bicycle” could have been written with a small circle and a T-like shape representing a pedal.
And the stroke of genius here is that if you do it that way, you can use the same written language to write German and Swedish! It would simply be pronounced “Fahrrad” in German and “cykel” in Swedish. So you could use one written language for three very different spoken languages.
That’s what they did in China. The same written language is used everywhere, which means that someone speaking Cantonese can write to someone writing Mandarin and they’ll understand each other perfectly, even though they might be unable to communicate by speaking. That’s what even made it possible to have one unified administration for a country the size of Europe with twice the population, and twice the ethnic diversity. (If you think “Chinese” is one single ethnicity, it just means you don’t know anything about China.)
The downside, of course, is that it’s going to be a nightmare to learn, since you basically need to learn one character per word (it’s not quite that simple, but basically). And another nightmare to design a keyboard for, although that came later. I’ve seen a Chinese typesetting machine; I worked at the only translation agency in Sweden that could do typesetting in Chinese at the time (early 1990s). It was the size of a large desk and required some strength to operate the big lever that picked up the characters, and 30 words per minute was considered lightning speed.