Dealing with homophone mergers

[StrangerCoug]

I had been keeping separate text files for my Wideriver language family and I accidentally overwrote the Proto-Western Wideriver to Proto-Southwoods .sc file, which I still need, with the changes for the stage after that (to Ancient Southwoods).

I put the starting word list and ending word list in this Google spreadsheet. There are two pages, one for the dictionary forms and one for the verb forms—I'm running under the mindset that everything I have helps right now. How do I use these wordlists to recreate the .sc file that I lost?

(Yes, I know that the word lists in their current form have some troublesome sets of homonyms. I'm willing to hear suggestionsbut want to focus on reconstructing the lost file.)

[StrangerCoug]

OK, false alarm—apparently I still have it. Would like some help with the homonyms mentioned, though.

[Pabappa]

I like the sound changes. As for the homophones, it doesn't look that bad, so far, compared to say Japanese which has 76 meanings for /shō/. Japanese solves the problem by basically only using those words as parts of compounds, or sometimes in placenames. But it depends how strong your urge to use compounding is. Old English had a lot of words like /eoh/ "horse", which is related to Latin equus and would probably have just become "egh" or even "ee" in Modern English, but died out early on and was replaced with another word entirely rather than being bound up in a compound with some other word. Similarly with Old English ae "law", and ae "river", neither of which survives in modern English in any form.

[Nortaneous]

Old English had a lot of words like /eoh/ "horse", which is related to Latin equus and would probably have just become "egh" or even "ee" in Modern English, but died out early on and was replaced with another word entirely rather than being bound up in a compound with some other word.

wouldn't eoh have become eigh?

anyway, homophones can be replaced or disambiguated through compounding, but if they're not confusable they might just be left alone. the only potentially confusable homophone pair i saw skimming your wordlist is the words for 'no' and 'know', which are homophones in english

[CaesarVincens]

There was study one of my professors mentioned that found that homophones occur at a rate above chance in the languages examined (I think it was Austronesian or Polynesian languages). I believe the rate of homophones varies by language, but for some languages can be fairly high. There may be a psychological or cognitive reason for this (easier memory load?).

[KathTheDragon]

Maybe it correlates with syllable complexity? IIRC, Chinese and Japanese (especially Chinese) have lots of homophones, and they both have quite simple syllables (compared to, say, English). Of course, I know of too few languages to say this with any certainty.

[Chagen]

Maybe it correlates with syllable complexity? IIRC, Chinese and Japanese (especially Chinese) have lots of homophones, and they both have quite simple syllables (compared to, say, English). Of course, I know of too few languages to say this with any certainty.

Chinese's homophones come from sound change merging different-sounding words, and Japanese I think gets most of its homophones from Chinese loanwords, made worse by the fact that words with the same vowel and different tones would be borrowed as the same-sounding words, and Chinese had many more phonemes that often were merged in Japanese--<ngw> and <m> at least initially were just borrowed as <m>.

Also, many of Japanese "native" homophones actually have different pitch accents and thus aren't actually homophonous in speech (and, thanks to kanji, writing either).