An experiment with phonetic writing

[ololololo]

As promised, we’re experimenting with Chinese . I’ve chosen the Cantonese dialect. The original text is one of the SBS [Shennong Bencao?] recipes that AI provided me with (I’m also sorry about that). ... I turned to conventional phonetics. Overall, this script is much closer to the sounds than the previous one.  ... I can suggest that you work with writing, and based on the results, find the appropriate recipe

I see that in your transcription the second word is repeated two more times near the end of the recipe.

So, assuming the transcription is consistent for that word, one should scan a digital file of the Shennong Bencao (in Chinese characters or Mandarin pinyin, it should make no difference) looking for an entry with the same repeat pattern.

For curiosity, I did that with my transcription of the Starred Parags section, treating commas like word spaces.  My little prog says that parag f104r.43 has 31 tokens and the second token qokaiin is repeated exactly twice at indices 18 and 29 (counting from 0). Maybe...

But probably not, because the VMS Author apparently omitted certain standard fields of the SBJ recipes, including some that occur at the end of most recipes.  If that recipe is typical and complete, the last of the three repeated tokens (third one from the end) is probably 生 = shēng in Mandarin = saang¹ in Cantonese, which means "Grows in" or "Habitat" or "Provenance".

All the best, --stolfi

In fact, the trick is only in the characters. You could say it’s the same as pinyin, just with the letters replaced.
There may not be a recipe. If there isn’t one, I’ll provide the text I used. Well, most likely it won’t, since AI gave it to me.