No sound changes
@Suzukaze-c Are you sure about the "no sound changes occurred" part? What does that even mean? Kou could have very well been pronounced as "ko-u", and the modern pronunciation is a mere subsequent assimilation (kou>koo). This is especially true in some historical compound reductions/elisions such as imo-hito>imo-ito>imo-uto>imooto>imoto.
And if there was no sound change, why even include it there? Isn't the point of including it is to denote sound changes (which is the main point of all dictionaries used as references on Wiktionary: historical kana are used to denote historical deviations; if there is none, then no historical kana are mentioned; in fact, dictionaries tend to omit the "unchanged" bits of words and replace them with dashes). ばかFumiko￥talk 23:36, 3 August 2017 (UTC)
- I suppose that's true. (Maybe "minimal sound changes" would have been better wording.) However, the initial rationale for this type of formatting that I liked was here (User:Krun). At any rate I think it's fairly harmless overall, and largely OK since we have "by historical reading x" categories.
- (Should the romanization for historical こう be kou instead of kō? I think it should be technically possible.) —suzukaze (t・c) 23:45, 3 August 2017 (UTC)
- This might help: KSound, wonder why the ancient reading has a small katakana ゥ. There exists コウ and コゥ. 22.214.171.124 00:01, 4 August 2017 (UTC)
- @Suzukaze-c I'm not opposed to including it if necessary. The thing is that the romanization of historical kana still need more work, so maybe it's not ideal to include this instance unless it denotes the deviant part of the historical kana. How exactly should は・ひ・ふ・へ・ほ be transcribed? Pa, pi, pu, pe, po? Fa, fi, fu, fe, fo? Or just ha, hi, hu (fu?), he, ho? What does ぱ・ぴ・ぷ・ぺ・ぽ stand for? Should small kana be used at all? Dictionaries don't use ぐゑん (only げん), so should it be used here (and it has been used here)? ばかFumiko￥talk 09:42, 4 August 2017 (UTC)
- @Fumiko Take I think it is absolutely necessary to explicitly show こう as the historical reading here, and also other historical kana spellings that may be identical to the modern spelling. As @Suzukaze-c has already pointed out, it is quite possible in the case of modern こう that the historical form be かう. If we don't give any indication of the historical spelling, it is not possible for the reader to be reasonably certain whether the historical spelling is こう or かう. In many dictionaries, an indication of the historical spelling is included where it differs, but omitted otherwise. However, the perpetual work-in-progress nature of Wiktionary makes this method unusable; one never knows when a page might be edited and readings added. Most of our existing kanji entries need to be edited manually to add historical readings and this in itself is a monumental task, as it requires manual lookup in printed kanji dictionaries such as the 大漢和辞典 and careful editing and cross-linking between kanji variants. Here is an example of my attempt to do them justice in complex cases. It is quite necessary to provide for the now very common case of a reading being included in a kanji entry without the historical spelling being known. The prevalence of this will in future hopefully be much reduced but probably never eliminated. Therefore, こう in the template does not and cannot mean the same as こう<こう. However, even if it could, I wouldn't support it because I want Wiktionary to be clear and explicit. After all, this is a dictionary that is not only meant to be used by specialists, but by all sorts of amateurs as well. Even modern readings that only have one possible historical form (generally identical to the modern form), e.g. ら, benefit from their historical form being explicitly mentioned, as this lays the information out plainly for the non-specialist, who isn't to know that Japanese never had other precursors to that particular kind of reading. – Krun (talk) 15:41, 4 August 2017 (UTC)
- @Krun I feel you, but just to reiterate, I think there's still problems with the romanization rules (based on which phonology framework? Old Japanese, Middle Japanese?), and at the same time, which type of kana should be used (based on which past use, which period, which text corpus, etc.?). If those problems haven't been sorted out, maybe don't include questionable notations, don't you think? And kou or kau are just specific examples. Are you absolutely sure that the reading ku of this very entry has remained the same ever since it was borrowed (in case it was borrowed and not made up by lexicographers)? Using a format like く<く or ら<ら or げん<げん would be just fine as long as you're sure those are what they always have been, not something like, I don't know, くぅ or ら<るゎ or げん<ぐゑん, and you have come up with proper romanization as well as other preferences for readers to find out more. To be honest, I'm sorta reluctant to add historical readings to Kanji entries because what you have on Wiktionary is still an awful mess of conflicting conventions and rules. ばかFumiko￥talk 16:09, 4 August 2017 (UTC)
- @Fumiko Take The “historical” spellings are not some arbitrarily attested forms, but simply the standard spelling (正仮名遣い) used before the adoption of the 現代仮名遣い after the reform of 1946. They are widely attested, even after 1946, and is also recorded in numerous dictionaries, which generally agree on the spelling. く<く simply means that the reading is ku in today's standard and was also ku according the pre-1946 norm. It is not of much consequence whether a reading has always been the same since it was borrowed, although it is interesting to include very early forms that differ from the later norm. That is why the template provides for “ancient” forms of readings as well. These should all be standardized spellings, however. Other variations, such as various attested spellings with hentaigana, variations with small kana, etc., can be included in the usage notes below.
- Certainly, the romanization currently has problems that must be sorted out. However, I don't consider this a major issue; because the romanization is generated automatically, changes to it will automatically be propagated to all pages where it appears. I generally consider these romanizations useful and necessary, again, particularly for the non-expert who might have a linguistic interest in Japanese. I have previously suggested that such a romanization be based on Nihon-shiki, but modified, so that every vowel kana is written out (oo for おお, ou for おう, etc. instead of ô or ō; so yes, kou instead of kō, @Suzukaze-c). I would prefer the は-row to be romanized with initial f-: fa, fi, fu, fe, fo; I think that would make for good consistency while not being too modern nor too ancient and preserving the distinction between は (“fa”) and ぱ (“pa”). The biggest technical problem with the historical romanization is doubtless the lack of explicit yōon (although it only affects kun-readings and compounds). That can be remedied by assuming yōon and using dots in the wikitext to split syllables where necessary, e.g. し.よう for a word such as 使用, producing romaji siyou, but plain しよう for 承, producing romaji syou. We will have to have a short appendix page to explain it as well. – Krun (talk) 02:30, 5 August 2017 (UTC)
- @Fumiko Take, @Krun, @Suzukaze-c The current discussion regards kanji and historical on-readings based from Middle Chinese (字音仮名遣い), but there are examples in native kun as well. Understood that the small katakana in the KSound is related to the final -ng. It's rare to find such useful resources in the net. I suggest leaving out the small kana and stick to the modern ha and fu for now. Also, "くゐ", "けゑ", and related are ancient readings according to the KSound. Some historical kana usage examples If you want more useful online resources, would be happy to give you. 126.96.36.199 23:46, 4 August 2017 (UTC)