2

I just learned today really how prevalent homophones are in Chinese, some of which are listed here. I asked a similar question about how to say Chinese sentences using all the variations of meaning of a homophone here.

We dried the sweet smelling anhydride into a liver powder.

It would be like:

We gān the gān smelling gān into a gān powder.

Now I'm curious. If there are only 400-something Chinese syllables, times 4 tones, so about 1600 possible 1-syllable words. Yet there are at least 10k characters in use, meaning roughly 5 meanings are possible per phonetic syllable. I am not sure I am doing calculations right, but that is at least true for gān from my other post, which has at least 5 unrelated meanings.

My question is, how far can you take this? What other languages have such a high degree of reuse of phonetic words? Natural or constructed languages. English has some, but IMO to a much lesser extent (I know of a handful of words with 2 distinct meanings, like "bear" the animal vs. "to carry", but 99% of words have 1 general meaning, as opposed to Chinese which has way way more homophones). How do you not create confusion when devising a system like this?

English avoids confusion when using the multiple meanings of a single homophone by the use of affixes, as in:

The bear beared the bear bearing bears.

But if there were 5 unrelated meanings of the word "bear", then I don't see how it could not create confusion.

I have ported 4,000 words to the conlang I'm working on so far, but they are all highly generic words, plus animal names (so far). In Chinese, they have about 100 1-character animal names, then the rest of the animals are derived from these 100 combining them together or with adjectives (hippo = river horse, panda = cat bear, etc.). But now I am thinking of how to port things from these categories, each which have thousands if not way more "types" of things:

body parts (tongue, bones, muscles), diseases (malaria), rocks (emerald), occupations (ballet), material (cotton), food (vegetable, bread, etc.), building type, land forms, flowers, trees, fish, molecules, etc..

I was also looking at the Organic Chemistry nomenclature in Chinese, which has seemingly random and meaningless names for various organic compounds, with barely any phonetic resemblance. In essence, they are just reusing existing sounds/homophones, adding one more meaning to the sound/word. If I could do this, and do it well, then I could basically pick a word like "wolf" and call a specific animal a "wolf", a molecule a "wolf molecule", and a building a "wolf building", basically creating new unrelated meanings for the same underlying word, and theoretically there would be no confusion. But this is a very foreign concept to me, so (a) I'm not sure it would even work out in the end, or (b) how confusing it could be come, and (c) how to proactively avoid creating confusing scenarios.

How did Chinese not make things confusing when creating these homophones? Coming from a programming/software background, the only way I could see this reliably being done is to have millions of sentences in a database, and every time you add a new homophone, you would check if it would create confusion against the millions of sentences. Even that wouldn't be enough, you would have to try and construct sentences from each homophone to make sure no confusion would take place. But that seems like an insurmountable challenge, so curious how a natural language like Chinese has accomplished this. And also curious if there are other constructed languages that have created a high degree of homophones like this. But more generally, wondering how I could take advantage of this sort of high-homophone-density system in the conlang I'm working on.

1
  • 1
    "...have millions of sentences in a database, and every time you add a new homophone, you would check if it would create confusion against the millions of sentences." This is basically what natural languages do. A word is missing and someone starts saying X to represent it. If it creates confusion in to many or too important contexts it's less likely to catch on.
    – Edvin
    Oct 20, 2022 at 7:05

1 Answer 1

3

Coming from a programming/software background, the only way I could see this reliably being done is to have millions of sentences in a database, and every time you add a new homophone, you would check if it would create confusion against the millions of sentences.

That's basically how it works. Remember that the purpose of every natural language is to communicate, and every natural language is under enormous evolutionary pressure to fulfill this purpose.

Suppose I want to bring a new word into English; perhaps I want to talk about the purviews of ancient Sumerian deities, so I borrow the Sumerian word me. It turns out that the purviews of ancient Sumerian deities are very cool and come up in conversation all the time, and soon everyone in my social circle wants to talk about them.

Will the borrowed word me catch on, or will the homophone with "may" (or the homograph with "me") be too confusing? Well, speakers will find out pretty quickly. If it's causing problems and impeding communication, they'll switch to "purview" or "domain" or borrow another word like partzum that's less ambiguous. After all, their goal is to convey some particular meaning, and they'll choose their words to achieve this goal. If a word doesn't help with this purpose, it won't be used.

We see this happen all the time in language evolution. If a word is too ambiguous to be useful, it will be replaced. In most varieties of Occitan, the word for "rooster" is a descendant of Latin gallumexcept in Gascony, where it's been replaced by a word for "vicar". Why? Because sound changes in Gascon caused the descendant of gallum "rooster" to sound the same as the descendant of cattum "cat" (compare castellum > Gascon castet). This ambiguity was apparently unacceptable to the Gascon farmers, who started calling their roosters "vicars" instead; the context makes it clear enough if you're talking about a farm animal or a priest, so this homophone wasn't an issue.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.