Character dictionaries are sorted by stroke order, not by sound.
So if a Chinese/Japanese person encounters a character that they don’t know (and since there are, by a conservative count, more than 50,000 of them, this will surely be the case), they examine the character for familiar radicals (stroke patterns; there are 216 of them, in order of number of strokes), then look under that stroke pattern and find the character based, again, on number of strokes.
Voila. The character comes up, complete with pronunciation possibilities (often, there is more than one), root meaning, and then popular character combinations.
There are also special dictionaries for person’s names, etc.
I once bought a character dictionary in Korean, because it had a more complete list of characters than any I had seen in China or Japan. Also because it had pronunciation guides for Chinese, Korean, and Japanese, making it most useful for comparing the three languages.
I still have it. Bought it at a Korean-language bookstore in San Francisco CA. The bookseller was rather bemused by my enthusiasm.
