> I'm sure I've run into other Chinese characters that are pronounced differently depending on whether they're in a verb or an adjective.
The Japanese writing situation is much, much worse than that. Kanji get all kinds of different pronunciations on the theory that the semantics are the same. So the standard spelling of musume ["daughter"; "girl"] is 娘, but it might also be spelled 女. This gives you an alternative to the standard independent reading of 女, which is onna ["woman"]. All characters will also be pronounced differently when they represent Chinese loanwords than when they represent native words.
> And yet, in English, "made a cake" and "made no reply" have exactly the same verb. Again, one can imagine it going both ways
I agree with that.
> The real limitation here is that there's no way ready-made to indicate the "make" -> "made" transition, as you can with "makes" ("作s" ) and "making" ("作ing").
I don't agree with that; you'd do the same thing you do with every other preterite verb and write 作ed. You wouldn't pronounce that maked, but that's not a problem; we're still assuming that everybody knows English.
> IIRC, in Japanese, 他 is sometimes used for "hoka", which means "other"; as in, "Do you have this item of clothing in other (他) sizes?" The original text says, "some one or other of their daughters"
Yes, I determined something similar by looking the character up in a Japanese dictionary. I don't have the skills to determine what kinds of uses are and aren't natural, so I just decided that I couldn't label the usage wrong, but it still stood out as funny for other reasons. I was able to correctly read the English without referencing the original text.
I'm not sure what a "moderate amount of Mandarin" means; if it's on the lower end, you might be interested to know that the "other" sense does survive in Mandarin, in the words 其他 ["other"; extremely common] and 他人 ["other people"; not so common]. 他 by itself is going to be overwhelmed by the far more common use as a pronoun, I would guess.
> My expectation, which matches my (very small) experience, is that Japanese are trying to write Japanese; and that therefore they have Japanese words in mind that they're trying to represent with Chinese characters.
While I do agree that the situation could shake out in several ways, I don't think this is a total defense of the idea that "the same word" is going to be spelled the same way by people who don't know Chinese. I wouldn't expect, for example, that fire [flame] and fire [eliminate from a job] would get the same character spelling.
> One could imagine using "作" in both cases; or one could imagine using two different characters for two different shades of meaning, just as in Chinese there's 他, 她, 它, and 祂
By my understanding, 祂 is not a part of ordinary mainland usage. But this is a good example, in that there is no distinction between the words in the language, and Chinese people aren't able to make the distinction when learning a foreign language even though they do make it in writing their own. (By contrast, Spanish speakers don't have problems choosing between the English words he and she.)
You might also be interested to know that the character 做, pronounced identically to 作 and meaning the same thing, is in fact derived from 作. It (and not 作) is now the ordinary character used for the sense of making or doing. 作 is best known to me as part of the word 作者 "author". This is a pure spelling distinction that arose by some natural process within Mandarin.
> or perhaps 的 and 得 (which I'm not sure people would naturally consider different words if they were only exposed to the spoken word).
I can answer that; they don't. At least they don't for the 得 that introduces manner or result clauses; they might or might not think of the possibility infix seen in 听得了 "able to listen" as different.