古文學ばむ Learn Classical Japanese

A forum for guides, lessons and sharing of useful information.
Post Reply
clawgrip
MVP
MVP
Posts: 2257
Joined: 24 Jun 2012 07:33
Location: Tokyo

古文學ばむ Learn Classical Japanese

Post by clawgrip »

I thought some people might be interested in this topic, and since I'm by no means an expert in this topic, I felt it would be a good opportunity for me to solidify my knowledge and learn more as I do it. I will follow a book I have, translating and adapting as I go to suit the audience here.

Introduction
Classical Japanese is essentially the language of the Heian Period (794-1185), a language which persisted thereafter as Japan's literary language, even as vernacular Japanese continued to evolve, until the Meiji period (1868-1912), when literature began to be written in contemporary Japanese. The Heian period was in many ways the defining era for Japanese culture. It was a time when the imperial court was at the peak of its cultural influence, laying the foundations for much of Japanese art and refined culture. It is thus of little surprise that the language of this period, technically classified as Early Middle Japanese, remained as the literary standard for the following millennium.

Lesson 1: Phonology
Classical Japanese, as a literary language, has the exact same phonology as contemporary Japanese. The characters of a text are pronounced according to their modern values. However, Early Middle Japanese had a phonology that differed somewhat from Modern Japanese. Certain peculiar vocalic features of Old Japanese had already been lost, but a number of differences remained. It's surprisingly difficult to find a concise outline of the phonology of Early Middle Japanese (or any pre-modern Japanese, for that matter),because almost every source focuses on the workings of spelling, making the issue seem more complex than it is. So here is what I believe to be the phonology of Early Middle Japanese:

Vowels
Early Middle Japanese has the same five vowels as Modern Japanese, though from what I can /u/ may have a different value (or my sources are using a broader transcription)

Code: Select all

i iː    u uː
e eː    o oː
    a aː
These will be Romanized as-is, with macrons for long vowels. The only peculiarity here is that /e/ and /o/ can only appear post-consonantally. These things are handled a bit confusingly in the script (as I will describe below), and it confounds the issue unnecessarily.

Also I'm not 100% clear on whether /aː/ and /iː/ are actually attested in Early Middle Japanese or not, and whether /eː/ is actually /ei/ or not, but I am including them for good measure.

Consonants
Early Middle Japanese had the following consonant inventory (excluding allophones for now):

Code: Select all

    m   n         (ɴ)
(p) b t d   k  ɡ
            kʷ gʷ
 ɸ    s z
        ɾ
          j    w
  • Like Modern Japanese, code /n/ was realized as [ɴ]. Or maybe it's a separate phoneme. Whichever.
  • /ɸ/ merged with /w/ intervocallically. I will Romanize it as ‹h› in all cases, as is conventional.
  • voiceless consonants can be geminated. Geminated /ɸ/ becomes /p/ and will be marked as such. /p/ may exist as a phoneme in mimetic/onomatopoeic words. Will have to confirm this.
  • /u/ cannot appear after labial consonants, (/kʷ gʷ w/), while /i/ cannot appear after the palatal consonant /j/
  • Like Modern Japanese, /s z/ and /t d/ are believed to have had varying pronunciations based on the following vowel. Based on contemporaneous Portuguese transliterations of Japanese words, we can posit the following:
    • /si se/ - [ʃi ʃe] ← (‹xi xe›)
    • /zi ze/ - [ʒi ʒe] ← (‹ji je›)
    • /ti/ - [t͡ʃi] ← (‹chi›)
    • /di/ - [d͡ʒi] ← (‹ggi›)
    • /tu/ - [t͡su] ← (‹tçu›)
    • /du/ - [d͡zu] ← (‹zzu›)
Syllable Structure
The influx of Chinese vocabulary during the Heian period altered the syllable structure of Japanese. The very strict CV pattern of Old Japanese, that forbade CC and VV sequences, was broken down. Final consonants, including final /ɴ/ and the geminate consonant, were introduced, as well as the velar labials and long vowels.

Spelling
Please note that I will assume that you know how to read hiragana and katakana. If you don't, please check out my thread on Modern Japanese here. Here's where the mess comes in. Early Modern Japanese had the same inventory of kana as modern Japanese but with the addition of the two letters ゐ / ヰ wi and ゑ / ヱ we. Additionally, the letter を / ヲ wo, restricted in Modern Japanese to the accusative marker and nothing else, is somewhat more frequently encountered in Classical Japanese, though its pronunciation is not distinguished from お.

The technique of using small characters to distinguish yōon from standard syllables (i.e. CjV vs. CVjV) was not established in Classical Japanese. Therefore I probably will not use them in this guide either.

The main source of confusion for deciphering the phonology is that Old Japanese man'yōgana (phonetic use of kanji found in the Man'yōshu and other old texts) consistently distinguished three types of /e/ signs in writing, in a way that suggests a significant difference from Modern Japanese:

e: 衣, 依, 愛, 榎, etc.
ye: 兄, 江, 吉, 曳, 枝, 延, 要, 遥, 叡
we: 咲, 面, 廻, 恵, etc.

Old Japanese distinguished three distinct syllables where Modern Japanese has just one syllable (e).

By the time of Early Modern Japanese, the distinction between e and ye was lost in writing, while we remained (I.e. e merged into ye, which is why /e/ can never appear first in a syllable). This is also confusingly associated with the labialized velars, which are related orthographically, but not phonemically. Plus /ɸ/ merged with /w/ intervocalically, but remained distinct in writing. Adding to the confusion is that in Early Modern Japanese or thereabouts, intervocalic we an wi merged into ye and i respectively, so the ゑ/ヱ, despite being the character for we, is associated with the sound ye.

The most annoying part of it all is that it seems people easily confuse phonemes for syllables, due to the way these sound mergers are typically discussed in the literature. When a source says that ア行のエ (e) and ヤ行のエ (ye) merged, they do not mean that /e/ became /je/ in all instances, they mean the syllable represented by エ is consistently pronounced /je/. ケ (modern ke) for example, does not become /kje/, it remains /ke/. Even Wikipedia, that bastion of truthy knowledge, has fallen victim to this, claiming that the vowel is /je/ and citing three separate sources, not realizing that those sources say $e → $ye, not /e/ → /je/. Oh well. The same happened for o, i.e. $o → $wo.

($ means syllable boundary)

The other vowels are much clearer:
a maintains the three-way distinction into MJ: a - ya - wa;
yi is phonologically impossible, and the remaining two-way distinction, i - wi, was maintained until EMJ;
wu is phonologically impossible, and the remaining two-way distinction, u - yu, remains in MJ.

So to clear things up definitively! Here are the problem syllables (and the non-problematic /k/ syllables, for comparison) and how to write them:

Code: Select all

$_     k_     j_     w_        kw_
ア a   カ ka  ヤ ya  ワ wa     クワ kwa (クヮ)
イ i   キ ki  --     ヰ wi     クヰ kwi (クィ)
ウ u   ク ku  ユ yu  --        --        --
--     ケ ke  エ ye  ヱ we     クヱ kwe (クェ)
--     コ ko  ヨ yo  ヲ/オ wo  クヲ kwo (クォ)
Just add a dakuten (゛) to the k series to get the g series.

The last column represents how it the same sounds are normally written in Modern Japanese (for loanwords, since labialized velars are gone from Modern Japanese).

Let me know if you're interested in me continuing this.
Last edited by clawgrip on 21 Sep 2016 01:42, edited 2 times in total.
clawgrip
MVP
MVP
Posts: 2257
Joined: 24 Jun 2012 07:33
Location: Tokyo

Re: 古文學ばむ Learn Classical Japanese

Post by clawgrip »

Lesson 2: Nouns
This will be a short and simple lesson. Words in Classical Japanese can be classified broadly into four categories based on two variables: boundedness (basically, whether it's an independent word or a suffix) and capacity to conjugate. In this lesson we will focus on the main type of unbound, non-conjugating word: namely, nouns.

As a broad class, nouns in Classical Japanese can be divided into 5 types:

Proper nouns, e.g.:
かぐや姫 Kaguyahime (legendary princess)
京都 Kyōto (a city, maybe you've heard of it)
隅田川 Sumida-gawa (a river in Tokyo)
仁和寺 Ninna-ji (a temple in Kyoto)

Regular nouns, e.g.:
haru "spring"
sakura "cherry blossom"
楽しみ tanoshimi "pleasure; enjoyment"
舟遊び funaasobi "boating"
もの思ひ monoomohi "meditation; anxiety"

Both proper nouns and regular nouns function much the same as in Modern Japanese. Those familiar with modern Japanese know the core particles は wa (topic marker), が ga (nominative/subject marker), and を wo (accusative/object marker). These exist in Classical Japanese also, but they are used much more sparingly. In standard, non-colloquial Japanese, these particles always written, but in Classical Japanese, like modern colloquial Japanese, they are frequently dropped.

Numerals, e.g:
一つ hitotsu "one"
二人 hutari "two people"
三年 sannen "three years"
四時 yoji "four hours"
五番 goban "fifth"
六代目 rokudaime "sixth generation"

Grammatical nouns:
These are regular nouns that also have secondary grammatical functions, or nouns that only carry grammatical meanings but function identically to nouns, e.g.:
こと koto "thing; affair", (intangible nominalizer)
とき toki "time", (conjunction "when")
もの mono "thing; object", (nominalizer and other functions)
ゆゑ yuwe "reason" ("by; for")
ほど hodo ("amount"; "way"; "during"; etc.)

Pronouns:
Pronouns include personal and demonstrative pronouns. There are a number of them, and are rather different than modern Japanese.

Code: Select all

          1P          2P          3PROX     3MED      3DIST     INDEF
Personal  あ、あれ    な、なれ    こ、これ  そ、それ  か、かれ  た、たれ
Pronouns  a,  are     na, nare    ko, kore  so, sore  ka, kare  ta, tare
          わ、われ    なんぢ                          あ、あれ  なにがし
          wa, ware    nanji                           a,  are   nanigashi
          おのれ                                                それがし
          onore                                                 soregashi
The indefinite pronouns indicate meanings such as "someone", "whoever", "what's-his-name", etc.

Demonstrative Pronouns

Code: Select all

          1P          2P          3PROX     3MED      3DIST     INDEF
objects   --          --          こ、これ  そ、それ  か、かれ  いづれ
                                  ko, kore  so, sore  ka, kare  idzure
                                                      あ、あれ  なに
                                                      a,  are   nani
locations --          --          ここ      そこ      かしこ    いづこ
                                  koko      soko      kashiko   idzure
                                  これ      それ                いづく
                                  kore      sore                idzuku
                                                                いづら
                                                                idzure
directions--          --          こち      そち      あち      いづち
                                  kochi     sochi     achi      idzuchi
                                  こなた    そなた    あなた    いづかた
                                  konata    sonata    anata     idzukata
                                                      かなた
                                                      kanata
You can also see that there is significant overlap between 3rd person personal pronouns and demonstrative pronouns.

There are various kanji for these pronouns:
a, are (1P)
wa, ware, a, are (1P)
onore
汝、爾 na, nare, nanji
ko
so
ka, kare, a, are (3P)
彼処 kashiko
nani, idzu-
-ko, -ku
-nata

As you can see, writing the pronouns kanji may in some cases make it unclear which version of a particular pronoun is meant.

Examples of multi-character pronouns:
此れ kore
其方 sonata
其処 soko
何処 idzuku
Last edited by clawgrip on 14 Aug 2020 16:04, edited 2 times in total.
clawgrip
MVP
MVP
Posts: 2257
Joined: 24 Jun 2012 07:33
Location: Tokyo

Re: 古文學ばむ Learn Classical Japanese

Post by clawgrip »

Lesson 3: Verbs
This is where things begin to get a bit complicated. Classical Japanese functions similarly to Modern Japanese in that verbs employ compounding of morphemes in order to convey meaning, but Classical Japanese is a bit more systematic and complex. This lesson will cover unbound and to an extent, bound, conjugable words.

First it is necessary to define what exactly conjugating a word means. Simply put, a word that conjugates is a word that can change into (up to)six separate stems (irrealis, conjunctive, conclusive, attributive, realis, & imperative), which can then accept further morphemes (or not). Bound, conjugating morphemes are those morphemes that are added to a stem, and then themselves conjugate into the six tems in order to accept more morphemes (or not).

Conversely, all morphemes subordinate to the main verb are bound to specific stems, to which they must attach. To make this clear, let's use some practical examples.

Let's take these three morphemes:
~ず -zu negative marker
~けり -keri past tense marker
登る noboru "to climb"

The negative marker ~ず zu is quite understandably bound to the irrealis stem, so if we want to negate the verb 登る noboru, we must first conjugate it into the irrealis stem: 登ら~ nobora-, and now we can add the negative ending, resulting in 登らず noborazu "do not climb".

The past tense marker ~けり keri, on the other hand, is bound to the conjunctive stem, so we must conjugate 登る noboru into the conjunctive stem, 登り~ nobori-, and we can now add the past tense ending, resulting in 登りけり noborikeri "climbed".

If we wanted to combine these, that is, make a past tense negative, we need to conjugate the negative marker into the conjunctive stem, i.e. 登らず noborazu → 登らざり~ noborazari-, and then add the past tense marker, resulting in 登らざりけり noborazarikeri "did not climb".

I hope this explanation makes the basic concept clear, because understanding this type of chaining is essential for understanding Classical Japanese. If you need more or better examples, let me know.

These conjugable suffixes are called 助動詞 jodōshi, and from now on, I will use this term, because "conjugable suffix" is a bit too unwieldy to use every time.

Now let's learn how to conjugate into the six stems.

Based on the way they form their stems, normal verbs are classified into five standard classes and four irregular classes. One of the standard classes and three of the irregular classes comprise only one verb each, while the remaining irregular class has two verbs in it. Thus, the vast majority of verbs belong to one of four standard classes, meaning once you learn the rules for those four classes, you will be able to conjugate the majority of verbs.

Let's look at the classes.

Yodan/Quadrigrade
The class gets its name due to the stems being formed by adding one of only four vowels to the root. It is notable as the only regular verb class that takes no stem suffixes. Here are the six stem endings, and the verb above, 登る noboru, a yodan verb, conjugated into those stems:

Code: Select all

irrealis     -a  登ら nobora
conjunctive  -i  登り nobori
conclusive   -u  登る noboru
attributive  -u  登る noboru
realis       -e  登れ nobore
imperative   -e  登れ nobore
As you can see, only four vowels, a, i, u, e, are used to conjugate all six stems, and no suffixes are added after them. Some of the stems are identical in form.

The other four regular classes only use one or two vowels, meaning they must additionally employ suffixes to distinguish stems.

Kami-Nidan/Upper Bigrade
Bigrade verbs, as I suspect you have inferred, use two unique vowels in the formation of their stems, in this case /i/ and /u/. The reason it is "upper" bigrade is because in a chart of kana, イ i appears physically above エ e, the "lower" vowel. Here is a chart of the upper bigrade stems, using the verb 伸ぶ nobu "to stretch/lengthen"

Code: Select all

irrealis     -i    伸び    nobi
conjunctive  -i    伸び    nobi
conclusive   -u    伸ぶ    nobu
attributive  -uru  伸ぶる  noburu
realis       -ure  伸ぶれ  nobure
imperative   -iyo  伸びよ  nobiyo
You can see that the bigrade verb employs suffixes on the last three stems to disambiguate.

Shimo-Nidan/Lower Bigrade
The lower bigrade verb class is nearly the same as the upper-bigrade class, the only difference being that i is replaced with e. Here is the verb 食ぶ tabu "to eat"

Code: Select all

irrealis     -e    食べ    tabe
conjunctive  -e    食べ    tabe
conclusive   -u    食ぶ    tabu
attributive  -uru  食ぶる  taburu
realis       -ure  食ぶれ  tabure
imperative   -eyo  食べよ  tabeyo
Upper Monograde
Monograde verbs, of course, use one unique vowel in the formation of their stems, in this case /i/. The -i is in fact a part of the root itself. For the sake of clarity, however, I have included the final -i in the chart below. Here is a chart of the upper monograde stems, using the verb 見る miru "to look at"

Code: Select all

irrealis     -i    見    mi
conjunctive  -i    見    mi
conclusive   -iru  見る  miru
attributive  -iru  見る  miru
realis       -ire  見れ  mire
imperative   -iyo  見よ  miyo
As you can see, it employs suffixes in all but the irrealis and conjunctive stems. In fact, the -i is part of root itself.

Lower Monograde
This class functions identically to the Upper monograde, but with /e/ in place of /i/. There is only one verb in this class: 蹴る keru "to kick":

Code: Select all

irrealis     -e    蹴    ke
conjunctive  -e    蹴    ke
conclusive   -eru  蹴る  keru
attributive  -eru  蹴る  keru
realis       -ere  蹴れ  kere
imperative   -eyo  蹴よ  keyo
These are the five standard verb classes. You will notice that none of these verb classes distinguish all six stems. Now we will look at the irregular stems, some of which do distinguish between all six stems.

N-irregular
There are only two verbs in this class: 死ぬ shinu "to die" and 往ぬ/去ぬ inu "to leave". The N-irregular class distinguishes all six stems:

Code: Select all

irrealis     -a    死な    shina
conjunctive  -i    死に    shini
conclusive   -u    死ぬ    shinu
attributive  -uru  死ぬる  shinuru
realis       -ure  死ぬれ  shinure
imperative   -e    死ね    shine
R-irregular
There are also two verbs in this class: 有り ari and 居り wori, and they differ from the quadrigrade only in the conclusive stem, which is identical to the conjunctive stem, rather than the attributive stem:

Code: Select all

irrealis     -ra  有ら ara
conjunctive  -ri  有り ari
conclusive   -ri  有り ari
attributive  -ru  有る aru
realis       -re  有れ are
imperative   -re  有れ are
K-irregular
There is only one verb in this class: 來 ku "to come". It is highly irregular, but easily learned:

Code: Select all

irrealis     來       ko
conjunctive  來       ki
conclusive   來       ku
attributive  來る     kuru
realis       來れ     kure
imperative   來/來よ  ko, koyo
As you can see, the pronunciation of the stem is not indicated orthographically.

S-irregular
There is only one verb in this class: す su "to do". Like ku above, it is highly irregular, but also easily learned:

Code: Select all

irrealis     せ    se
conjunctive  し    shi
conclusive   す    su
attributive  する  suru
realis       すれ  sure
imperative   せよ  seyo
This verb frequently appears after other nouns to verbalize them. Sometimes it is voiced, taking the form -zu, but other than the consonant being voiced, it conjugates identically.

We have also mentioned two dependent jodōshi, ~ず -zu and ~けり -keri. These also conjugate into the six stems, but these suffixes are often somewhat irregular, and often defective, that is, many do not conjugate into all six stems.

Let's look at these two first of all:

-zu
As I described above, this is the negative morpheme, used simply to negate verbs. It is always bound to the irrealis stem.

Code: Select all

irrealis     ざら/(ず)  zara/(zu)
conjunctive  ざり/ず    zari/zu
conclusive   ず         zu
attributive  ざる/ぬ    zaru/nu
realis       ざれ/ね    zare/ne
imperative   ざれ       zare
行く iku "go"
行かず ikazu "do not go"

You can see that it is somewhat similar to the n-irregular class, but a number of stems have two forms. The difference is this: the first form used used when it is followed by another jodōshi, and the second form is used when it is when it is phrase-final, or when it is followed by a 助詞 joshi (non-conjugable suffix), the latter being indicated by parentheses.

-keri
This is the standard past tense. It is always bound to the conjunctive stem.

Code: Select all

irrealis     (けら)  (kera)
conjunctive  --      --
conclusive   けり    keri
attributive  ける    keru
realis       けれ    kere
imperative   --    --
This one is missing a few stems, and also the irrealis cannot be phrase-final. You can see that it otherwise resembles the r-irregular class. The reason is that I am pretty sure it is a merger of earlier -ki + ari. Basically anything in Classical Japanese that conjugates like ari does so because it is a contraction involving ari.

行く iku "go"
行きけり ikikeri "went"

Let's throw in one more, for fun:

-nu
This is a past or perfective form that indicates that something is done, or something that the speaker is certain about. It conjugates identically to the n-irregular class. It is bound to the conjunctive stem.

Code: Select all

irrealis     な    na
conjunctive  に    ni
conclusive   ぬ    nu
attributive  ぬる  nuru
realis       ぬれ  nure
imperative   ね    ne
行く iku "go"
行きぬ ikinu "has gone; is gone; will surely go (i.e. something like "is as good as gone")"

Note on verbs ending in -yu:
As I mentioned above, there is no difference between e and ye, so you can write either e or ye. If you need to turn yu to yi (not sure if this happens), you would just drop the y, i.e. yu→i

For those of you familiar with Modern Japanese, you should see some similarities, though in some cases this is partially obscured by sound changes.
The base form in Modern Japanese is the same conclusive form as Classical Japanese (or in some cases the attributive, but that's another story). The standard negative morpheme has changed from -zu to -nai, but the same irrealis stem is used, e.g. kiku: kikazu→kikanai. The past tense has undergone significant sound changes, but signs of it are still there, e.g. kiku: kikikeri kiita (the original /k/ has been elided). So, the concept is basically the same, but it is a little more complex but systematic in Classical Japanese.

Now it's time for exercises!

Exercises
Here are a few words for you to work with, with their classes listed:
4 = quadrigrade, U2 = upper-bigrade, L2 = lower-bigrade, U1 = upper-monograde, L1 = lower-monograde, N/R/K/S = irregular
行く iku "go" (4)
伸ぶ nobu "stretch" (U2)
食ぶ tabu "eat" (L2)
見る miru "look at" (U1)
蹴る keru "kick" (L1)
言ふ ihu "say" (4)
聞こゆ kikoyu "be audible/heard" (L2)
御座す ohasu "to be present; to arrive (honorific)" (S)
始む hajimu "to begin" (L2)
nu "sleep" (L2)
落つ otsu "fall" (U2)
信ず shinzu "believe" (S)

Transliterate and then translate the following into English (use whatever subject pronouns you feel like for the English translation):
1. 始めけり
2. 寢ず
3. 見ざりけり
4. 言ひぬ
5. 御座せず
6. 來ぬ

Translate the following into Classical Japanese (pronouns need not be included):
7. (I) I will not eat (remember there is only past/nonpast)
8. (I) did not kick (it).
9. (It) stretched.
10. (It) has fallen.
11. (It) cannot be heard/is not audible.
12. (I) did not believe (it).
Post Reply