The Mandarin dialect of Beijing used only 412 different monosyllables, with the result - in one quoted example - that a small dictionary, giving a total of no more than 4000 everyday words, was found to contain sixty-nine pronounced yi. The confusion is slightly modified by the famous four tones of spoken Chinese...But these are not distributed with mathematical fairness, and of those sixty-nine words no less than thirty-eight used the falling tone. Spoken in the fourth tone, yi could mean bosom, different, contemplate, wing, city, translate, a hundred thousand, hang or any of thirty other equally varied possibilities. In practice the Chinese, when speaking, avoid ambiguity by a system of duplication, tacking on another word of the same meaning just as we might distinguish between hang-suspend and hang-execute, or as schoolchildren do between funny-ha-ha and funny-peculiar.Here's the situation I often find myself in: When you're learning a language and trying to parse a spoken sentence, your ears latch on to the familiar. If the familiar part is one syllable/hanzi, but it turns out to be the 'wrong' one, your brain has gone so far down the wrong parsing tree that the rest of the sentence is a washout. You're trying to interpret what you subsequently hear based on an incorrect context.
For example, if you hear the familiar zai4 somewhere at the start, and think that it's 在, then you are going to be listening for a place, time or perhaps a verb coming next. If however it turned out that the zai4 you heard was actually part of zai4 xie2 - 载携 (to carry/to bear) - then you've already lost the thread of the conversation and it's going to be very hard to pick it up later. It would have been easier if the verb to carry had had a completely different sound. It almost seems that in order to learn a word, you have to learn all the other homophones in order to, as they say in all the best B&W detective movies, eliminate them from your inquiries.
I know that's not actually the case: nobody learns a language one lexical item at a time, and I'm not going to start a new trend! So what can we do to avoid this problem? My best guess would be to adopt the following rule of thumb: Never learn a singular syllable on its own. Always learn them in groups of two or more. The hope would be that these combinations would be what the ear will recognize as familiar.
* BTW: Ricci was the man who gave Kong Fuzi his Latin name of Confucius.
** Edit: Corrected homonym to homophone (28/03/2007)