Sunday, November 18, 2012

Konjaku Monogatarishū Vol. 2, Tale 1: Śuddhodana's Death #1 (part 1)

I'll be using a slightly different format for Konjaku than I have been using for Umegoyomi. Konjaku is written in a style known as wakan konkōbun (和漢混淆文), or "mingled Japanese and Chinese writing". It is defined on Wikipedia as "Sino-Japanese composition written with Japanese syntax and mixed on'yomi and kun'yomi readings". As indicated on the linked Japanese Wikipedia page, Konjaku is one of the most famous early examples of wakan konkōbun, while later examples include TsurezuregusaHōjōki, and Heike Monogatari.

There are also some other quirks that make the text more difficult to understand from just a transcription of the scans, so a modified transcription will be more prominently displayed. Also, there are a few things I've learned through the Umegoyomi translations about the whole process, so hopefully this approach will be more transparent.

The main changes in the modified transcription will be:
  • Any traditional characters (旧字体) will be changed to simplified characters (新字体). The traditional forms will still be used in the figures and their captions.
  • Any kana that lack dakuten or handakuten will have them added. As noted on Wikipedia, these diacritics were not considered standard until the Tokugawa era, hundreds of years after this manuscript was produced. Edit: as suggested by yudantaiteki on the RtK Forums, dakuten usage can be more accurately characterized as having gradually developed over time, and there are cases in which they can be found in Heian-era writings. Konjaku doesn't appear to be one of them, though.
  • The replacement of certain Chinese grammatical patterns with their Japanese equivalents. This will become more apparent after starting to read the translations.
  • HTML furigana will be frequently added to parts of the text. The irregular furigana usage, frequent use of Indian names written in ateji (in the case of this tale, the kanji transcriptions were originally created by the Chinese), lack of furigana in the original work (unlike Umegoyomi), and continual use of obscure terms makes this quite useful. Furigana should render correctly by default on the latest versions of Chrome and IE; for Firefox, there's the extension HTML Ruby, but many reviews claim it will cause serious performance issues when you have lots of tabs open (an issue I've run into myself). Similarly, there's an Opera add-on; it worked for me in my brief testing, but I don't have extensive experience with it (however, Opera's add-ons are architected quite differently from Firefox's extensions, so there being performance issues with the Opera add-on as well isn't a foregone conclusion).
I'm starting with the very first tale in the Indian (天竺てんじく) part of Konjaku that is available in the Suzuka Manuscript. The first volume is missing (at least from Kyoto University's scans), so this will be tale 1 from volume 2 of the collection.

Fig. 1: 佛御父
First up is the title of the tale:
 The story of when the Buddha's father, King Śuddhodana, died, part 1 (fig. 1)
So the first thing you're probably wondering is where all those okurigana in the modified transcription came from. The truth is that I don't really know. They're not in the original text, but they're in the transcription from Kyoto University, and certainly make sense here.

One might think that the lack of kanji indicates that the title is actually in Classical Chinese, which certainly could be the case. However, from my experience with kanbun, it would be very rare for a string of characters that long to be in the same order in both Japanese and Chinese, given their wildly differing grammars. My theory is that all kana are omitted, but the kanji are in the Japanese order. Therefore, the reader is required to supply the kana by himself. The Heian nobility was certainly obsessed with all things Chinese, so such a title would have made the text look more sophisticated.

The verb 「給フ」, pronounced 「たもう」, is an honorific supplementary verb suffixed to verbs to show respect towards the subject of the sentence ― King Śuddhodana, in this case. Note that 「給フ」 requires the preceding verb to be in the continuative form (連用形). This is likely how it was determined that the okurigana for 「死」 should be 「ニ」. The use of katakana, rather than hiragana, is explained below.

Fig. 2: 今昔
One interesting thing here is that Śuddhodana's name doesn't appear to be written using ateji, unlike many of the other names we will encounter. In modern-day Mandarin and Cantonese, the pronunciations of 「浄飯」 are "jìng fàn" and "zeng6 fang6", respectively. Even looking at the meaning ("clean" and "food"/"rice"), doesn't yield much useful information.

Edit: Matt's comment below has cleared this up. "Śuddhodana" in Sanskrit is "शुद्धोधन", which can be broken down as "śuddha" ("शुद्ध"), meaning "pure", and "odana" ("ओदन"), meaning "rice" or "food". This fits the meaning of the kanji 「浄飯」, so this can be considered an 「意訳」, or meaning-based translation.
Once upon a time, the Buddha's father, the Great King Śuddhodana of the country of Kapilavastu (fig. 2), was confronted with old age, suffering from disease while some days passed and worrying things weighed heavily upon him, with no limit (fig. 3).
Fig. 3: 老ニ
First of all, we are encountering the use of kana in the original text for the first time. Note how only katakana are used, and how they are (often) set to the right side of the text and are in smaller print than the kanji. This was quite common in the Heian era, when hiragana were considered feminine, and therefore inappropriate for men to use (Edit: also suggested by yudantaiteki, it's more that the male-dominated areas of writing used katakana. Men could also write in styles that used hiragana, like waka). Since these works are of foreign origin (an area of literature that was, at that time, male-dominated), it makes sense that only katakana are used. Moreover, katakana were originally devised by Buddhist monks, and this work consists of many Buddhist tales.

Obviously, there is no actual 「ハ」 between the first and second characters in fig. 2. The decision to add a 「 ハ」 (the modern particle 「は」) was made by me. Clearly, 「今昔こんじゃく」 can have the same meaning as 「今ハ昔」 (and it does), but this makes it easier to read/understand.

Our first example of ateji also appears in fig. 2, with 「迦毘羅」, which has a reading clearly created from the Japanese pronunciation of the first half of "Kapilavastu".

Edit: as explained by a commenter below, the primary Japanese reading of 「迦毘羅」 is "kabira", and not "kapira", as listed here, despite the original word having a "p". This is probably because the "name of the city must have entered China with a Prakrit or non-Indic source, which voiced the medial stop". However, according to this page, it appears that "kapira" is a valid alternate reading. This is interesting, considering that the modern readings of 「毘」 in Mandarin and Cantonese are pi2 and pei3, respectively.

There are several interesting things to note about fig. 3. First is the verb 「臨テ」. This kanji is encountered in verb form in modern Japanese as 「のぞむ」, meaning "to look out on" or "to confront". As is often seen, the meanings stick to the kanji much longer than the readings do, which change with the frequent shifts that occur in any spoken language. I chose to assign the reading 「み」 based on the fact that (a) it fits the okurigana, where 「のぞ」 does not, and that (b) it is still used as a reading for names, or nanori (名乗り). I have noticed that nanori sometimes preserve archaic kanji readings. Edit: as Matt as suggested in the comments, it's more likely that 「臨テ」 is read as 「のぞみて」. This is indeed a grammatically valid construction, and is equivalent to 「望んで」 in modern Japanese, which is just a minor sound shift.

Also of note is the irregular okurigana for 「受」 ― this is something encountered previously in Umegoyomi, but it'll probably come up much more often with an older text like Konjaku.

The word 「日来」 has two possible readings ― 「にちらい」 and 「じつらい」. In fact, as suggested in this comment, when Konjaku was originally written, such words may have actually been pronounced in Chinese. Since both have the same meaning, I just went with the first one listed. Edit: as Matt has suggested in the comments below, the meaning is likely 「日頃」, and not 「ふだん」. In classical Japanese, the phrase 「日頃を経る」 means "some (i.e., a few) days pass".

Note how the kana right after 「悩乱」 looks much more like the hiragana 「し」 than it looks like the katakana 「シ」 (Edit: yudantaiteki, in that same post, said that in his experience, this way of writing 「シ」 is standard). One downside of these Kyoto University scans is that the quality isn't that high ― zooming in doesn't help much, given their low resolution. In any case, it appears that in the top left corner of that kana, there are two strokes, as seen in 「シ」, so I selected the katakana version. Moreover, it would be rather odd to see a random hiragana interspersed in a Heian text, even though we've seen that semi-arbitrary switching back and forth is quite commonplace in Umegoyomi.

Fig. 4: 无限
(just an
example, not
from the text)
At the end of fig. 3, we encounter the first example of rewriting a Chinese grammatical pattern into its Japanese equivalent. One might think from looking at the original manuscript that the kanji 无」 should actually be 「元」, but the overlain version of the text provided by Kyoto University has it clearly marked as 「无」. The differences in the handwritten versions of the two kanji are certainly minimal.

Whereas the original text says 「无限シ」, if there were kunten ("guiding marks for rendering Chinese into Japanese") included, they would likely indicate that this should be read as 「限り无し」. This is also how it is transcribed on Kyoto University's site, albeit as 「限り無し」. 无」 is just an alternate kanji for 「無」, which can be seen in 「い」 (although it is usually left in kana form in modern Japanese in in this context). I chose to leave it as 「无」, to reduce the number of unnecessary changes.

The proper method in kanbun of indicating the way such a pattern should be read can be seen in fig. 4. The use of a kaeriten (the 「㆑」 symbol on the left side of the figure) between the two kanji indicates those two kanji should be reversed when being read in Japanese. The hiragana are the readings of the kanji, and the katakana are their okurigana. Thus, we get the 「限リ无シ」 for the kakikudashibun (書き下し文) - the equivalent text when rewritten in classical Japanese.

Fig. 5: 身ヲ迫ル事
The way this pattern was written at the end of fig. 3 is probably just one of the idiosyncrasies of wakan konkōbun. As Chinese grammatical patterns go, this is a fairly simple one, so it was probably assumed that readers would be able to parse the text without the aid of a kaeriten or complete okurigana.
Compelling the body [to do anything] was like pressing oil (fig. 5).
This was an odd sentence because although it was short and the grammar was straightfoward, the meaning eluded me. It appears to be some sort of figure of speech I'm not aware of, so I could use some help here.

There is one other interesting point ― we see the first example in Konjaku of how the addition of dakuten to the text is left up to the reader. That is, the 「ガ」 is left as 「カ」.
Fig. 6: 今ハ限リ

Thinking that now [he had reached his] limit (fig. 6), [Śuddhodana] lamented that he would probably die without [first] seeing his sons the Buddha and Nanda, his grandson Rāhula (fig. 7), his nephew Ānanda, etc. (fig. 8)
Fig. 7: 御子ノ
The first point of interest in fig. 6 is 「思シテ」 ― it comes from the verb 「おぼ」, which is just a polite form of 「思う」. In fact, the modern Japanese translation given is just 「お思いになる」.

Another thing to note is how the last two kana are not vertically aligned, as would be expected. This is a little reminiscent of togaki, which we saw in Umegoyomi, but I think that it might also have been to make the kana fit into the space of one kanji, so they don't stand out that much or waste space on what might have been expensive paper.

Fig. 8: 甥ノ阿難
Fig. 7 is full of name ateji. First we have another name for the Buddha, 「釈迦しゃか」, which comes from his Sanskrit name Śākyamuni ("शाक्यमुनि"), meaning "Sage of the Śākyas", where the Śākyas were the tribe that the Buddha was born into.

We see similar ateji for the Buddha's half-brother (Nanda, or 「難陀」), the Buddha's son (Rāhula, or 「羅睺羅」), and for the Buddha's cousin (Ānanda, or 「阿難」). The kanji for Rāhula are particularly interesting, for two reasons. First, 「羅睺羅」 is often written as 「羅ご羅」, including in the Kyoto University transcription and in the title of his Japanese Wikipedia page. For whatever reason, use of the kanji 「睺」 is not very common/popular. Second, notice how the first and third kanji are the same ― once again, there's clearly no meaning to be drawn from the kanji (which, in other contexts, can mean "gauze" or "net for catching birds").

In fig. 8, there's another Chinese grammatical pattern ― 「不見スシテ」, which is how it is in the original text. This time, interestingly enough, it is dealt with in the opposite way ― it is overdefined, rather than part of the interpretation being left up to the reader.

The Chinese grammatical pattern 「不見」 simply indicates negation of the kanji 「見」, and would be written in premodern Japanese as 「見ズ」, which is exactly what we see in fig. 8, with 「見スシテ」 (remember that insertion of dakuten is left up to the reader). What's interesting is that both the kanji 「不」 and the okurigana 「ス」 are included, when just one would have sufficed.

We once again see the placement of multiple kana (this time, three of them: 「スシテ」) in the space for one kanji.

Also of interest here is the verb 「死ナム」, which can be parsed as the imperfective form (未然形) of the verb 「死ヌ」 (which is 「死ナ」), followed by the auxiliary verb 「ム」, in its attributive form (連体形), which is also 「ム」. The auxiliary verb takes on the meaning of appropriateness ― i.e., "should not die". Edit: as Matt pointed out in the comments, in this context, 「ム」 more likely has the meaning of "was apparently going to". He defined this as the "hypothetical" meaning of 「ム」, but I see another "hypothetical" meaning for 「ム」 in my textbook, used for "If..." sentences. It would be more appropriate to define this as being speculation/conjecture about the future (推量).

The verb 「給フ」 is seen here in the perfective form (已然形), as 「給ヘ」, followed by 「リ」, which is an auxiliary verb with the perfective function ― it indicates the completion of an action or process.


  1. Man, this is long. I admire your stamina!

    One interesting thing here is that Śuddhodana's name doesn't appear to be written using ateji, unlike many of the other names we will encounter.

    Yes, this is an 意訳 (translation of meaning). The original etymology of the name is suddha (pure) + odana (rice), thus 浄飯. (That's the Pali version, anyway; I'm on much shakier ground with Sanskrit.)

    The word 「日来」 has two possible readings ― 「にちらい」 and 「じつらい」. Since both have the same meaning, I just went with the first one listed. In fact, as suggested in this comment, when Konjaku was originally written, such words may have actually been pronounced in Chinese.

    In combination with へる though I think "ひごろ" is more likely. "After a few days..."

    Re 望, I think actually "nozomite" might be it; at least in the Heian period, "X ni nozomu" was a pretty common way of saying "face X" (not sure how much earlier it was used).

    Re 死なむ, I agree with your breakdown of the parts but not your interpretation; I think here it is used as a sort of "hypothetical case", thus "lamented that he should [=was apparently going to] die without seeing his sons [...]". In modern Japanese this would just be 見ないで死ぬこと, even in a hypothetical not-yet-happened case, but back in the day there was a distinction. (Just don't ask me to define it rigorously in theoretical terms.)

    Another thing to note is how the last two kana are not vertically aligned, as would be expected.

    Yeah, this is interesting. I observe that in the images you have here, the only okurigana that appears shifted to the left is テ. If this is a pattern throughout the text then logically we can conclude that the scribe thought of テ as unique among okurigana, which is the doorway to a whole new room of intriguing questions.

    By the way, this is just a technical suggestion, but how about rubifying words as words rather than individual characters? e.g. not 羅(ら)睺(ご)羅(ら) but 羅睺羅(らごら). I can see the argument for both ways of doing things (especially for ateji) but for us Firefox users the latter is much easier to read (since the ruby specification is not supported properly in Mozilla).

    1. > I think here it is used as a sort of "hypothetical case"

      Yeah, I just realized that there is no negative auxiliary verb in 「死ナム」, so it can't be "should not".

  2. The mu from shinamu:

  3. The characters for Kapila is read "kabira" not "kapira" ( Not too many a Sino-Japanese word (those with a ん or っ basically) contains a "p" anyway. Why was a character containing "b" (in Chinese and Go-on) used for an Indian "p"? The name of the city must have entered China with a Prakrit or non-Indic source, which voiced the medial stop.