[muddyrivers]

嚔 is indeed hard to produce, even for Chinese native speakers. It is an example of an exception, though, like entrepreneur in English (I know it is from French, so the spelling is "unusual").

First, people don't write 喷嚔 often, although they use it in spoken Chinese regularly. It seems people are reluctant to write down words that describe certain body functions.

For 嚔, the difficulty is that its "sound" component is extremely rare in Chinese characters. The majority of Chinese characters are 形声字. Each of such character has a "sound" component and a "meaning" component. Many characters can share the same "sound" component. The "sound" components are usually the "harder" component in writing, in that they have more strokes. (For 嚔, the rectangle at the left is the "meaning" component. The complicated looking part at the right is the "sound" component.) As a result, if one remembers how to write one character, one can simply remember many other characters with the same "sound" component. If one of these characters is commonly used, it is an easy task to remember all the others.

I only know two other characters who share the same "sound" component with 嚔, and I don't think people use them now. They only appear in ancient books.

So 嚔 is a character you don't often write down, and there are no other commonly used characters who shared the same "sound" component. Then you get a character who is hard for native Chinese speakers as well.

[jessriedel]

I like this analogy, but doesn't it just sort of point to the non-phonetic aspects of Chinese as the problem? If every time I had a hard time remembering how to spell something I couldn't even write down an approximation ("entrepenor") that someone or my computer could error-correct, it would strike me as a real flaw to the writing system!

(I'm assuming here that when Chinese writers can't remember the symbol for an unusual word, they also can't write down a recognizable approximation. Maybe not true?)

[muddyrivers]

See @schoen's good explanation to solve the issue with a computer.

It can be solved with a dictionary as well, by looking up the character from its pronunciation.

With neither computer nor dictionary at hand, it would mostly fall into three scenarios. First, you vaguely remember the character, but can't produce the correct one. Then you can make up a character that look closely to it, like an approximation ("entrepenor") in your example. People will get what you mean from the context.

Some people will substitute with another character of the same sound. Usually it is a character of much simpler form. This is actually an evolving process naturally occurred in history (I don't know what is the correct term in linguistics.) Many very complicated characters had been gradually replaced by simple ones of the same sound in Chinese's >2000 years' history. That is one of the reasons why one character may have multiple meanings.

Some other people will substitute with pinyin. It mostly happens in children's writing. Many children start writing diary at very early age (Chinese is indeed a difficult written language, so we have to start early :) Children like to try new words, especially the "big" words they hear from adults. But it is tiring and frustrating to keep looking up dictionary. Pinyin comes to the rescue. I find it cute, even aesthetic visually, that a child's writing has occasionally pinyin mixed with characters.

[schoen]

They could use a computer with a pinyin input method and see a list of characters that have that pronunciation. In this case they could enter "tì" or "ti4" and get a list something like

https://en.wiktionary.org/wiki/t%C3%AC#Romanization

but probably with a nicer interface. (They do have to be able to recognize the written character when they see it. Here the character Moser was trying to produce is one of the forms under item 5.) Also, there are statistical techniques to show the most probable characters in context following particular other characters, not just the most frequent characters overall.

For example, if Moser had just typed 喷 on a computer and then was looking for a character pronounced "tì", software could suggest 嚏 as the most likely possibility in that situation.

But it would probably be a fairly terrible situation when not using a computer.