Soooo, only read the more detailed answers if you really want to. It's mostly historical info and linguistic technicalities that can be interesting and helpful if you are interested in that sort of stuff or not convinced by the explanations you've heard. I apologize for the length, but the stuff I got out of my Japanese linguistics classes in university don't often come in handy in my day to day life, so sometimes I just have to dump that knowledge somewhere...
księżycowy wrote:From what I understand, the dotted circle is where the vowel is whispered/dropped.
Simple answer:The dotted circle shows that the vowel is basically not pronounced.
More detailed answer:The dotted circle shows that the vowel is devoiced which means that the vowel is produced without the vibration of your vocal chords. In practice, the vowel can be dropped pretty much entirely, leading to what sounds like consonant clusters that aren't normally permitted in Japanese like /fk/ or /st/. However, it's not entirely correct to say that the sound disappeared entirely: it leaves behind a few traces. For one, the mora where the vowel is "dropped" still retains its "rhythmic weight" so to speak. That's to say, if you clap out each mora in a word, the sound where the vowel is devoiced still counts as a mora and will be perceived as such.
As an example, the word スター (star) is a three mora word (su-ta-a). When spoken, however, it won't exactly sounds like the idealized [suta:]. It'll likely sound more like [sta:]. However, rhythmically, /su/ will keep it's status as a mora regardless of whether the /u/ is voiced or devoiced. You could sort of imagine /s/ as its own mora: (s-ta-a). This is particularly tricky for words like 進む (すすむ susumu) where the first す contains a devoiced vowel and so the word might sound, to an untrained non-Japanese ear, like two syllables: "su-mu". In reality, you have something more like "s-su-mu".
The second sign is that the devoiced vowel will leave its residue on the consonant that precedes it. This is particularly perceptible with い段 characters (characters that end in
i) since /i/ has a tendency of causing the preceding consonant to be palatized (which means that the tongue is brought closer to the hard palate [the hard part of the top of your mouth between the soft palate at the back and the alveolar ridge near your front teeth] when the sound is produced). So in theory, a devoiced く and き still sound somewhat different.
This sort of devoicing tends to only occur once in a word and only occurs when the vowels /i/ and /u/ are surrounded by two voiceless consonants (consonants that do not cause your vocal chords to vibrate: /p/, /k/, /s/ (/sh/ [ɕ]), /h/ (both /f/ [ɸ] and [ç]) and /t/ (both [ts] and /ch/ [tɕ]) in Japanese) or when the vowel is preceded by a voiceless consonant and occurs sentence-finally (so like です at the end of a sentence).
księżycowy wrote:But what is the handakuten on ユキスキ?
Simple answer:The NHK dictionary uses handakuten with the か行 (all characters starting with
k) to indicate that the consonant /g/ is pronounced [ŋ] (
think English ng, the link has a recording), but you can just consider it the same as if it were written with a regular dakuten (i.e. が, ぎ, ぐ, げ, ご) and pronounce it [g].
More detailed answer:The handakuten marks when か行 is pronounced as [ŋ] whereas the regular dakuten marks when か行 is pronounced as /g/. In Tokyo Japanese, [ŋ] used to be and still is a possible realization of /g/ word-internally (i.e. not at the very beginning of a word). In fact, back in the 40s, a researcher named Kindaichi found that about 30% of all Tokyo Japanese speakers used [ŋ] consistently word-internally, another 40% used both [g] and [ŋ], and the remaining 30% used only [g].
So, for example, you have words like 'mirror': 鏡 (かがみ kagami) which would be pronounced as /kagami/ or [kaŋami] because the /g/ occurs in the middle of the word. However, a word like drama: 劇 (げき geki) would be pronounced by everyone as /geki/. */ŋeki/ is not an option because the /g/ occurs at the beginning of the word.
However, these word-internal [ŋ] pronunciations don't apply to 5 kinds of words:
1. Onomatopoeic words: がらがら (garagara) which is always [garagara] and not [garaŋara]
2. The number 5 and words containing it: 十五 (じゅうご juugo) which is always pronounced [d͡ʑu:go] and never [d͡ʑu:ŋo]
3. /g/ directly preceded by the honorific 御 (お o): お元気 is always pronounced [ogenki] and never [oŋenki]
4. When /g/ is the initial position of the second element of a compound word: 世界銀行 (world bank: せかいぎんこう sekaiginkou) would always be produced as [sekaigiŋko:] and never [sekaiŋiŋko:].
5. The /g/ in loanwords like ヨーグルト (yogurt: yo-guruto) which would always be pronounced as [jo:guruto] and not [jo:ŋuruto].
There are also some minimal pairs (words that differ only by one sound like English
lit and
lot where only the vowel is different) that are said to be differentiated by word-internal [ŋ] or [g]:
千五 (one thousand and five: せんご sengo) as [seŋɡo]
戦後 (postwar: せんご sengo) as [seŋŋo]
There is a historical split in Tokyo between these pronunciations. Yamanote (the western part of Tokyo which was traditionally more middle/upper class) heavily favoured [g] whereas as Shitamachi (the eastern part which was more of a lower/middle class area) favoured [ŋ]. Nowadays, Yamanote [g] has been mostly imported into Shitamachi speech, likely due to the prestige and the more abundant number of Shitamachi people who worked in Yamanote than vice versa. As a result, Tokyo Japanese as a whole seems to be moving towards word-internal [g] rather than [ŋ]. In fact, young speakers are very unlikely to produce [ŋ]. You will still hear this pronunciation, especially in older speakers, but also in NHK newscasters because it is considered by some to still be the more prestigious way of speaking.
In other words, the handakuten marks a historical and prestigious [ŋ] pronunciation of /g/ that is good to be aware of (since it might confuse you when you hear it if you're unaware of it), but is not something that you need to concern yourself with unless you wish to adopt this pronunciation. You can basically consider it a regular dakuten for your learning purposes.
vonPeterhof wrote:Basically the horizontal line represents high pitch and the small vertical one represents the downstep. Unaccented words like 行き過ぎ have no downstep, so they start low and continue high. アンケート has two permissible pronunciations, an atamadakagata one where you start high only to drop the pitch right after the ア and continue low, and a nakadakagata one where you start low, go up after ア and drop after ケ. Another way of expressing those pitch patterns would be 行き過ぎ [0] and アンケート [1][3], where the numbers indicate the accented mora, i.e. the mora right before the downstep.
To develop on this point, you can assume that all Japanese words "want" to have a high pitch unless the accent comes along to bring it down. They start low and then immediately go high because that's what they like. Then, when the accent happens, the word is no longer allowed to go high again. If you wanted to see this in terms of "rules" and "definitions", it would look something like this:
1. The accent is a change from high to low pitch. It technically occurs between two morae, but the high-pitch mora that precedes the accent is said to be accented.
2. Words can be accented or unaccented.
3. If the first mora of a word is unaccented, it will be low.
4. The pitch of a word is naturally high leading up to the accent, even if there is none.
5. The pitch of a word is naturally low after the accent.
This is pretty hard to grasp in writing alone and I realize you have to look at three different things at once, but I'll try anyway. It's also just a lot at one time. You might need to reread it a few times or find a better explanation (hopefully with sound and visuals).
Let's take a look at the first pronunciation of アンケート.The one with the line over ア.
- Rule 2 tells us what kind of word we're dealing with. In this case, アンケートis an accented word and the accent occurs on ア.
- Rule 4 tells us why there is a line over ア. Since the drop occurs after ア, everything before that accent (which is just ア) needs to be high.
- Rule 5 tells us why there is no line over the rest of the word (ンケート). The accent has already occurred, so the pitch must stay low for the rest of the word.
- Rule 1 explains the little downward tick after ア. It denotes the fact that the word is experiencing a switch from high to low pitch.
- Rule 3 is irrelevant because the first mora is accented.
We can do the same with the second pronunciation of アンケート. The one with the line over ンケ.
- Rule 2 tells us what kind of word we're dealing with. In this case, アンケートis an accented word and the accent occurs on ケ.
- Rule 3 is why you see no line over ア in the second pronunciation of アンケート. The accent falls on ケ and not ア, so the first mora (ア) is low.
- Rule 4 explains why the pitch goes up on ン and stays up until ケ. The word naturally wants to have a high pitch, so it rises between ア and ン to achieve this.
- Rule 1 explains the little downward tick after ケ. It denotes the fact that the word is experiencing a switch from high to low pitch between ケ and ー.
- Rule 5 explains why there are no lines over ート. The accent occurred after ケ, so the pitch has to stay low for the rest of the word.
We can even use this for 行き過ぎ.
- Rule 2 tells us what kind of word we're dealing with. In this case the word is unaccented, so we don't even need rule 1 because there is no accent.
- Rule 3 tells us why イdoesn't have a line over it. There is no accent in the word, let alone on the first mora, so the first mora must be low.
- Rule 4 tells us what happens for the rest of the word: there is a line over キスギ because the pitch is high until the accent. In this case, there is no accent so the pitch stays up for the entire word.
- Rule 5 is also irrelevant because there is no accent to bring the pitch back down.