The American Association of Teachers of Japanese has put up a PDF of "Papers from a panel on 'Less Considered Areas of Japanese Literary Translation'". It contains, among other material, two full papers: "Added in Translation: Comparative Translations of Poe's 'The Black Cat' into Japanese", by J. Scott Miller, and "Kanshi in Translation: How Its Features Can Be Effectively Communicated", by John Timothy "Handbook of Classical Japanese" Wixted.

Miller's paper is a worthwhile read but less down in the nitty-gritty than I would have liked (what, no five-way tables showing how each translator handled "mournful and terrible engine of Horror and of Crime - of Agony and of Death"?). After noting that (1) in modern Japan the word kuroneko (black cat) is probably most strongly associated with a moving company of that name, and (2) in Greece they write metaphor on the side of moving vans because metaphores means "transport" in Greek, he explains what he means by "added in translation":

I would like to suggest that Poe's "The Black Cat" became, in a way, a transcultural moving van that relocated irrational fears and emotions into fin de siecle Japan. Poe gains something in translation through the serendipity of the pre-existing, ominous metaphor of vengeful cat spirits. Let us pretend to be nineteenth-century Japanese for a moment, reading either of the then available translations for the first time, and bringing to our reading any number of encounters with stories of Japanese vengeful tales wherein the cat licks the blood of a murdered woman and turns into a demon, then sets out to avenge the victim's murder. Now, as we read Poe's story and imagine the reincarnated black cat Pluto perched on the top of the narrator's wife's corpse walled up inside the basement, we expect the obvious to happen: the cat will lick the wife’s blood. (Actually, after being in there for days, even more gruesome possibilities emerge...) And we naturally assume that the cat, transformed into a demon but trapped within the walls, will do what demons do best—SCREAM—and thereby avenge the wife’s murder. So the ending takes on more than just Poe’s original sense of irony. It gains ghoulish karmic overtones as well.

Pretending to be nineteenth-century Japanese is where I'm a Viking, but I prefer a swings-and-roundabouts model to a simple "added in translation" one. (For example, off the top of my head, if the cat's betrayal is the karmically unmarked way for the story to end, it seems to me that this undermines the eerie shifting of the boundary between sanity/rationality and madness/despair that is, after all, Poe's trademark.) I'm also not sure how to feel about the sustained moving-van metaphor, although I may just be jealous that I didn't think of it first.

Wixted's piece argues, basically, that kanshi (Chinese poetry by Japanese authors) "would be experienced differently, visually and intellectually, and better understood as being dynamic products of a Japanese literary tradition", as well as "more accessible to students of Japanese and Chinese language", if they were presented in translation with the following information:

1. the kanji for the text, essential to being able to follow and understand it;
2. kundoku readings of lines, to provide a sense of how the text is apprehended and appreciated as "Japanese";
3. both barbarized and naturalized English versions of lines, the former via English-language phrase-by-phrase parsing of lines ('English kundoku,' as it were); and
4. Chinese readings for verse-lines, to bring out the rhythm within lines (in visual blocks reflecting compounds and caesurae) and the rhyme-scheme in the poem's couplets (with highlighted rhyme-words).

An example of what this might look like is on pp 10-11. Obviously we are well into personal-taste-and-intended-purpose territory here. I wholeheartedly agree that including the Chinese text is always beneficial, but I can't really believe that there are many readers who would derive real benefit from both the "barbarized" English version and the kundoku reading, for example. I also think that the poetry suffers greatly from being interrupted by irregularly-sized notes, and would prefer all of the commentary and explanatory material gathered together as footnotes or a separate column. But really my reaction is to question the information selected for inclusion. Why include only a contemporary Chinese pronunciation for the poem? I imagine that that's how these poems get read in China today (same as most English speakers read Shakespeare without trying to reproduce Elizabethan pronunciation), but that doesn't have much to do with how these poems would have been understood in the Nara court, a medieval Buddhist monastery, or an early Meiji literary circle, for example. I have similar issues with the inclusion of kundoku readings; I don't necessarily agree, as Wixted argues on p 6, that including any reading is better than including none, because it strikes me as the sort of thing that naiive readers will mistake for objective truth (although I would not have this objection in cases where what was included was a specific kundoku reading included as part of the source text itself, rather than just the translator's preferred reading).

In short, I am down with the call for more information, but I don't think that Wixted makes a coherent case for his proposed format as the best format to adopt.