Vowel Pairs: 'ai', 'ei', etc
 

I have been looking all over for a decent explaination for what to do with vowel pairs in Japanese. Every introductory piece I have read mentions mostly that all vowels are to be voiced separately; the name Aoi is Ah - Oh - Ee and so on. However, there are also words like sensei, aikido, hai, and so on that seem to blend vowel pairs into a single sound. Downloading sound clips from both this site and others, I have observed (as well as the common pronunciation in America) sensei to be said 'sen - say', hai to be 'h - aye'. The katakana for translating the 'aye' sound is also 'ai' as in 'Bai Bai' as the romanji version of the American farewell.

I have seen a few pieces describing silencing vowels between voiced consonants and the dropping of the 'u' sound from various words ending in '-su'. I have also seen that 'ii' in romanji can describe a long 'i' rather than multiple syllables, but in I have seen '-o' sounds that are lengthened with a 'u' in both romanji and hiragana, so I am curious and confused by the discepency of romanji and attempts at translating pronunciation. However, I have yet to see the reasoning for multiple and varying vowel pairs that are seemingly one syllable, nor a "comprehensive" listing of the pairs and their appropriate pronunciation.

It seems too uniform to be written off as casual speech, nor does it seem like its simply a matter of speed. Does anybody have either a guide to vowel pair pronunciation or a reason why things seem so contradictory? Thanks.

-Eric

Re: Vowel Pairs: 'ai', 'ei', etc
 

Eric,

I don't think there is anything contradictory about it. I think that is is something that your ear has to get used to hearing. There used to be a brand of rice in Japan a long time ago called "Merika" named after the US. It was because the Japanese person naming the company couldn't hear the "a" because we accent the second syllable so strongly. Culturing the ear is a part of learning any language, I guess.

Charles Hill

Re: Vowel Pairs: 'ai', 'ei', etc
 

So are you suggesting that it words like 'hai' are actually two syllables, whether my English trained ear recognizes it or not? Would it then be improper and considered "Americanized" to not treat the 'ai' interaction as separate syllables, even though it sounds that way to me?

So when I hear the word sensei pronounced (as heard from the downloaded clip from aikiweb), what am I to make of the fact that there is no 'Ee' sound at the end of the word, and it sounds more like a simple Japanese 'se', though a bit elongated, than anything else?

I appreciate the feedback, but I'm still looking for further explaination of what you mean. Thank you.

-Eric

Re: Vowel Pairs: 'ai', 'ei', etc
 

I totally hear where you are coming from. It took a long time for my ear to become sensitive enough to hear the differences. I still have problems at times. I hear an "i" sound on the end of "sensei" when spoke by a Japanese. It becomes clear when for example, my high school students complain when I make them do extra sets of forward and backward rolls. They clearly enuciate each syllable to let me know they are not happy.

Another point is with haiku poetry, with strict numbers of syllables for each line. In English, "sensei" has two syllables, but when you write haiku in Japanese, it counts as four.

As for what is proper and improper, I think it depends on who you are talking to and if they understand or not. A Japanese person will hear each syllable and if I mispronounce a word, I will end up relying on them to make an effort to understand what I am saying.

Charles Hill

Re: Vowel Pairs: 'ai', 'ei', etc
 

Thank you, that makes things more clear. I've read a bit about speaking various languages and the various techniques and terms that must be applied to sound like a natural speaker, things like where you resonate words in your mouth, various voicing and lack thereof, and I'll assume that this is just one of those cases where the enunciation of the vowel pairs are simply phenomes that translate poorly to English.

As for proper/improper, since I currently don't have anybody to talk to in Japanese (I can't afford to eat sushi on a regular basis just to learn a little Japanese) and am only just starting to read, write, and speak, I guess I'll just keep watching movies and try to discern the proper pronunciation.

Since you mentioned syllable counting in Japanese and it seems to take into account each kana (hence the 'n' being a syllable in 'sensei'), I curious how the combo-kana work in this? I don't know the technical name, but I'm refering to 'kya', 'kyu', 'kyo' and other '-i' + 'y-'. Are they one or two syllables? How does the mini-'tsu' that connotates a break, and translates as a double consonant fit into the syllable counting scheme? I doubt I'll be writing any haiku's in Japanese soon (or English for that matter!) but I'm curious.

Thanks,
-Eric

Re: Vowel Pairs: 'ai', 'ei', etc
 

Eric,

That's a great question on the haiku. I have no idea. I asked my wife (she's Japanese) and she doesn't know either but guesses that each written character would count.

Charles

Re: Vowel Pairs: 'ai', 'ei', etc
 

Eric,

You're asking some deep and difficult questions (as you might have guessed). At the risk of getting in too deep myself, I'll try to give an outline of what's going on that will satisfy you.

These questions relate to what we (1) call "phonology". This refers to the systematic arrangement and patterning of sounds in a language, and contrasts with "phonetics", which is more about how sounds are produced and perceived in actual speech. Phonology is more abstract, i.e. there's a level of abstraction in the way language resides in your brain that organizes sounds in a systematic way. To give a computing analogy, if phonetics is like the machine language of speech, phonology is a hi-level language like Lisp or Java :-) (for another example of phonological phenomena, see my explanation in the "tori vs dori" thread).

Simply put, at the level of phonology, the "i" in "sensei" really does exist (even a 2nd-language speaker like Charles attests to it!) and the "ai" sequence of "aikido" is somehow (see below) two things, not one. However people's mouths will not always behave strictly in accordance with phonology. Thus discrepancies arise. (2)

The critical phonological unit of Japanese is the "mora", not the syllable, though the difference is beyond the scope of this treatise :) But Charles' example of how words are treated in haiku is wonderful - this is exactly the way we gather evidence about what is happening "under the hood" of language (Chuck, dude, with insights like that, you could be a linguist!) In haiku, each mora must be counted separately, and as noted the "n" that comes at the end of syllables is a mora unto itself (the only consonant to have this privilege :))

You asked about vowel "pairs". The simple answer is that in English these pairs (diphthongs) make up a single mora, while in Japanese they make two morae. If this seemed confusing to you, it's only because of so-called "L1 interference" (your mother tongue butting in while you try to learn another language). Syllables with "palatalized" onsets kya/kyu/kyo, sha/shu/sho, etc I believe are monomoraic (single mora), but don't quote me on that. Ask a haikuist if you really want to know ;)

So you can see there are no simple answers here. And to make matters worse, really figuring this stuff out requires teasing apart the way that morae, syllables, the writing system, phonetic constraints, etc all interact. I don't pretend to understand this stuff very well at all myself (not as a linguist anyway - as a speaker it's somehow mysteriously all there in my head). And indeed, Japanese phonology (simple as it may seem on the surface) is not a closed book yet by any means.

So the practical answer is: don't worry about it. Just pronounce things the way you hear other people doing, or until you get some sample data, the way that makes intuitive sense to you. Don't necessarily depend on the explanations in language textbooks, because many authors are not linguists and may explain things from a phonological perspective (because that knowledge is more accessible to the conscious mind, and is related to what they learned in school), and it isn't necessarily what you need to know about real-world pronunciation.

Ganbare!

--
(1) I'm a grad student in theoretical linguistics at the University of Utah, in case I needed to show credentials :cool:

(2) I intend to do some research on the question of long vowels (which you allude to) at some point to satisfy myself. It certainly seems that the "u" that appears in writing as the continuation of a long "o" doesn't really exist any more at all, except in writing (which any linguist will tell you is virtually irrelevant to the study of language as a human faculty). However, the "i" in "ei" sequences seems to be on its way out as well. I've many times seen "keitai" (mobile phone) written in the popular media in katakana as "keetai". I don't read manga much, but I'm sure it's common there as well. If such a thing appears even in writing (a far more conservative medium than speech), then I think it really signals a phonological change in progress, probably driven at the level of phonetics. This stuff just excites the heck out of people like me :hypno: Stay tuned!

Re: Vowel Pairs: 'ai', 'ei', etc
 

Either you're being very humble or you're wasting a lot of time and money! Oops, did I say that out loud?

On a serious note, thank you. I was mostly interested in why I saw it the way I did in writing, and you did a decent job of explaining that its really just a changing thing. Though I don't have any natives to speak with, I have heard correct pronunciation from a past sensei in the dojo setting, at least from basic things like 'hai', 'arigatou', 'sensei', etc. Things like the changing of the 'keikai' are happening in English, too, like dropping the second 'l' in 'traveller' and changing 'tonight' to 'tonite'. I'm sure you're aware of this (unless, of course, you are just throwing your money away!!), but don't feel bad, we need people like you to help us all figure out our native tongue after these damn kids butcher our language on the internet!

Domo arigatou gozaimashita,
-Eric

Re: Vowel Pairs: 'ai', 'ei', etc
 

He's being humble, another sign that he was in Japan too long.

Charles

Re: Vowel Pairs: 'ai', 'ei', etc
 

I hvnt really read all the posts in this properly, but with pronouncing japanese words the vowels are very important, its not like in English were the vowel can sound like 2 or 3 diff things... A is pronounced like the a in Aunty, I is pronounced like the E in Eagle, U is pronounced like...the u in fUr, E is pronounced like the 1st E in Elephante, and O is pronounced like the O in Octopus. If you are blending diff vowels together it is the same...like ai, say it slow like a...i and then say it quickly so it sounds like ai in aikido.
NOTE: the U on an end or in a work is sometimes not pronounced eg, arigato gozaimasU, and sometimes the i isnt as well...eg, counting itchI, shichi, hachi and the u, in roku...

I really hopes this helps...sorry if it dusnt.

ayla

Re: Vowel Pairs: 'ai', 'ei', etc
 

I'd say that the "u" vowel in Japanes is more towards the front of the mouth, and there's also usually more rounding of the lips (than in "fur"). I'd say it's more like the vowel in "food," only shorter.
The "o" in "octopus" as regularly pronounced by Americans, at least, would be closer to the Japanese "a" than its "o". The vowel is probably more like the vowel in "dough" albeit a bit shorter.
I'd say there's a bit of aspiration after the final "consonant" sound in each of the above words. Due to the location of each of them and the shape of the mouth/tongue, the aspiration takes on the characteristic of the "ending" vowel. (David, is there a technical, linguistic word for this?) So, in essene, although it's not "pronounced," it's certainly not just dropped.

-- Jun

Re: Vowel Pairs: 'ai', 'ei', etc
 

I hope I made it clear I was only poking fun at David, though I will not debate whether or not he is "turning Japanese" (in a silly reference to that horrible, horrible song!)...

While I was previously aware of certain "aspirations" of ending vowels, I have been curious as to what the rule for this is. My first guess was that it was just dropped off of common/core words like gozaimasu and desu, but that doesn't seem to hold true as I see more and more words with various vowels 'dropped.' What is the reason and/or rule for this?

Jun, in your article archived here on aikiweb about "o negai shimasu" you mentioned the 'silent' ending as a 'stop-fricative.' Is this the word you are looking for to describe this 'aspiration' or is this something different? I've searched here and on non-aikido, non-language based searches and found nothing that describes a stop-fricative in any relation to Japanese.

On a related topic, how does this relate to foreign loan words in katakana? There are many words that borrowed or translated that do not end in either a vowel or the letter 'n', but the final sound must still be reproduced by a Japanese kana, vowel and all. I have seen that 'u' is often used between two consonants and to end the word. It makes sense that these added vowels would be pronounced mid-word, as that would be logically easiest for native Japanese speakers, but what is the regularity that the final vowel is pronounced? For example, my katakana name would be Eri(k)ku Supineri. As an Italian last name, the final vowel is not only pronounced but distinctive and key; however, the first name Eric is pronounced the common way in English, no tricks here, but how would the Japanese pronounce it? Given an inability to speak any non-Japanese language, would they try to pronounce my name as close as possible to my "actual" name, regardless of spelling, or would they pronounce the ending 'u' ?

Thanks,
-Eric

Re: Vowel Pairs: 'ai', 'ei', etc
 

Sorry but im from New Zealand so i might pronounce different words/sounds differently then most people, so my saying Octopus and things like that to me is like that and i couldnt really think of ny other words at the time so my bad!!! lol SORRY!!!
Tha japanese might pronounce the name as close to english as possible and sometimes the U is silent so for Eriku the U might be silent but if they do pronounce it they will just add the U on the end of K...not much of a difference.

ayla

Re: Vowel Pairs: 'ai', 'ei', etc
 

I'd like to make comments on many of the follow-ups in one message, but it's too much trouble, so I'll do them individually :)

No, the bit about a change in progress is just my personal hunch. I suspect that /sensei/ has been pronounced [sensee] for hundreds of years. The key is that the underlying abstract representation of the word (in people's heads, indicated by slashes) has always been /sensei/ even though what comes out of their mouths (in brackets) is [sensee]. If a change is really happening, this means that the underlying form stored in people's heads will become /sensee/ as well.

Not to be overbearing about it :), but actually since these changes are not accompanied by changes in pronunciation (i.e. they're just spelling conventions), they are of no interest to linguistics whatsoever :D

If you want an intriguing example of something happening in American English, listen to the way some people pronounce L (at least the L that comes at the end of syllables - like "full"). I've noticed recently (since coming back from Japan, actually) that for some people it sounds more and more like W (these two sounds are actually phonetically quite close, although to avoid boring everyone too much, you'll have to take my word for it :)). I can see the effects in my own speech to some extent, but for some people it's very pronounced. This change happened in Polish fairly recently (meaning, maybe within the last few hundred years), which is why sounds spelled with an L-like character are pronounced as W. Can you imagine if this change runs its course, and it ends up like it did in Polish? It's hard to conceive of your language changing that much, but this kind of change is continually happening in all languages.

Again, pardon my geeky enthusiasm, but I really do love sharing this stuff with people.

You forgot to put a smiley after that :)

Dave

Re: Vowel Pairs: 'ai', 'ei', etc
 

I don't think you mean "aspiration", Jun, I think you just mean that *something* is pronounced, even though it might appear that the vowel has been deleted. No?

If there's a technical term for what you're describing, the closest I can get off the top of my head is "assimilation". IOW, the mouth begins to anticipate the coming sound, and then gets the rug of voicing pulled out from underneath it, leaving a kind of a "ghost" vowel behind. So the linguistic explanation for the point Jun's making is that high vowels (/i/ and /u/) are devoiced in certain positions in Japanese, not deleted. And as with most analog processes, devoicing is not an all-or-nothing proposition. In some cases total devoicing will occur, and in other cases it's only partial. Many factors will contribute to how much voicing is removed in a given case, including a speaker's personal speech habits.

So here's the answer to Eric's question about the pronunciation of his first name as well: I think it's safe to say that those Japanese who have some ability in English or other foreign languages will devoice the final vowel almost completely, so that it sounds like it should, and those who have never been able to speak a foreign language to save their lives and failed English in school will add a fair amount of voicing to the /u/. I think the latter group are really shrinking in number quickly, though, as the years go on.

Here's an example of where final /u/ will be almost totally voiced: imagine a young woman who's ticked off because you're suspicious of some claim she's making, and she wants to emphasize that she's telling you the truth (dammit): "Datte atashi ga itte'ru koto ga hontou nan deSU". I can see Jun, Charles and the other fluent speakers nodding and smiling :)

Dave

Re: Vowel Pairs: 'ai', 'ei', etc
 

That's the way it was taken :)

The vowels /i/ and /u/ (called "high vowels" because they are pronounced at the top of the mouth) are devoiced when surrounded by voiceless consonants, or at the end of a word preceded by a voiceless consonant. That's off the top of my head, but I think I got it right :)

Dave

Re: Vowel Pairs: 'ai', 'ei', etc
 

<axe action=grind>This is why, in our increasingly globalized world, we need to be teaching basic phonetics in primary school! If everyone understood even just the basics (and there really isn't that much to it), and how to read International Phonetic Alphabet, we wouldn't have these kind of misunderstandings!</axe>

Thank you :)

Re: Vowel Pairs: 'ai', 'ei', etc
 

Probably a much better description of the phenomenon than my use of the word "aspiration." Thanks, David.
Ah, yes. Thanks for the reminder of that term, too. It's been a while since my former life in linguistics...
Indeed! There's always also the interjection, "Usooooooooo!" when you tell someone something they don't believe...

-- Jun

Re: Vowel Pairs: 'ai', 'ei', etc
 

Here's something I came across while further researching both the questions and your answers (thank you, by the way). I don't understand it, but I'm going to pretend I'm smart anyway and quote it. :)

Now, as I said, I'm not exactly sure what, if anything, all this means. Well, lacking numbers, variables, and mathematical operations it doesn't mean anything, but that's a whole nuther arguement.... :P I think I got through the (above) Abstract and halfway through Part 1 before I punched myself in the face and went and proved that e^(i*pi) + 1 = 0 just to wash the bad taste out of my mouth!!

Anyway, perhaps these are the correct terms and labels?

To clarify, is this a rule of the language, a rule-like practice of modern speech, or something that is irregular? I only know American English and Latin, so I'm used to irregularities as well as silly things (like accents) that are 'supposed' rules, but were rarely actually used by native writers/speakers. I was once told by a Latin professor that of the thousands of accents we were to read in any given modern text, only two cases actually mattered, and even then they still didn't really matter, as they were mostly left out and dealt with by context by the common Roman.

Again, thanks for your time.
-Eric

Re: Vowel Pairs: 'ai', 'ei', etc
 

Sorry, after some more thought I decided it's not assimilation after all (just glad no one else got a chance to call me on it first :)) So I'm left naked, with no fancy-sounding jargon, however it seems that Eric's follow-up may shed some light in this area...

Re: Vowel Pairs: 'ai', 'ei', etc
 

Dou itasimasite (quiz: how many of those /i/s are devoiced? :))

So you're gonna throw formants in my face, eh? :) Is that how it's gonna be? :D

It means that people working in serious acoustic phonetics have actually researched this, and can provide a much more technically accurate analysis of it than my overly-simplistic MA-student drivel :) Maybe I'll go read that paper... I'm sure I'll learn something useful from it. I think I can say with confidence, though, that you don't need to know this stuff to be able to speak Japanese well :)

Yes, I'm not sure what you mean, and no. It is an absolutely regular, although not obligatory, phonological rule (1). Of course, this is not the kind of rule you learn in school (most if not all of which are antithetical to the scientific study of language), but the kind you learn by listening. Since until now Latin is the only foreign language you know, it's not surprising that you might be less in the habit of thinking about phonology, since we have no samples of Latin speech by native speakers to study :)

I studied Latin as well, many years ago, and I've forgotten almost all of it, but I must admit I have no idea what you're talking about. It sounds interesting, though. Can you clarify?

No need to thank me... I'm here because I like to be, like you :)

Dave

(1) Non-obligatory means that you can voice all of your high vowels and you won't be speaking wrongly, people will just think you're odd, or a foreigner. Compare this to an obligatory phonological rule like the palatalization of /s/ before /i/, i.e. you cannot pronouce the syllable /si/ in Japanese like the English "see", it must be pronounced like "she", no ifs, ands or buts.

Re: Vowel Pairs: 'ai', 'ei', etc
 

Hrm. I was explicitly taught this rule in the first term of my Japanese language classes. /u/ and /i/ were referred to as "weak vowels", rather than "high vowels," but it was still explicitly taught

In my experience, the amount of devoicing is region and gender dependent, with more devoicing being characteristic of masculine speech.

Re: Vowel Pairs: 'ai', 'ei', etc
 

Zero. I'm a crude, crass person who rarely does anything nice enough to be worthy of thanks, let alone responding to such thanks with further kindness. In fact, I don't voice any of the letters! :)

Well, I know the last /i/ is devoiced and can see no reason for the first one to be, as its preceded by vowels, and I cant even think of any English words silly enough to start with silent vowels (though consonants seem to run amok). I have a feeling that the middle /i/ is also voiced because if I'm understanding things correctly, it is the past tense, very polite form of 'suru' and other examples of pronouncing various forms of 'suru' you voice that /i/. Am I being too logical about all this? :)

Anyway, my final answer is one.

You might not have understood exactly what I was asking, but you answered my question nonetheless. And yes, I do tend to think in a rule-based way. I think it has to do with being able to conjugate 200+ forms of a verb without knowing how to really pronounce it or even knowing what it means!!

Well, there are lots of modern text books that have decided to "enhance" the readers comprehension of the language by adding a multitude of accents to passages of Latin. However, the only ones that have anything to do with specifying the form and case of a word are thus:

1st Declention Singular Ablative (long 'a' in puella) to distinguish from the Singular Nominative and Vocative and

[I believe, its been 5 years...] the base form of 2nd Conjugation verbs (long first 'e' in -ere) to distinguish between 3rd Conjugation verbs (with a short first 'e' in -ere).

So there are all these accents that don't really mean much in many cases, because they refer only to abstract notions of pronunciation and emphasis that it doesn't seem anybody really knows for sure about. And then, on top of that, most historical texts simply forego the use of any accents - including the two specific cases I listed above. It seems Romans simply weren't interested in conveying pronunciation in writing. So it seems modern texts try to make it analogous to modern Romantic languages like Spanish and French by adding accent marks, when it was used much more commonly like English, where it was devoid of any such markings. At least I've never seen it taken to the rediculous level the French took modern/Romanized written Vietnamese. That makes the Chinese ideograms of old look easy!

-Eric

Re: Vowel Pairs: 'ai', 'ei', etc
 

I have a feeling that this is the sort of thing that is explicitly explained to non-native speakers rather than to Japanese children. I mean, here I am, asking this very question looking for a spelled out answer. However, I don't think I've ever asked or been explained something like the 'th' sound in my native English. I knew how to say it long before I ever saw it in writing and simply accepted its written representation. I think it has to do with 1) learning to speak a language before learning to read it and 2) the more adult nature to question rather than accept.

If I'm speaking out of line, David, please feel free to smack me back down. :)

-Eric

Re: Vowel Pairs: 'ai', 'ei', etc
 

Misunderstanding... what I meant by "rules learned in school" was the sort of things we all learn in primary and secondary school about our mother tongues, that warn us not to split infinitives, use double negatives, or end sentences with prepositions. Knowing and using these rules is important for social reasons (we want people to respect our intelligence and level of education), but they have nothing to do with true grammatical correctness. The Japanese have their version of these rules too... the only one I can think of now is a proscription against verb forms like "tabereru" (can eat). The "socially acceptable" form, of course, is "taberareru".

Indeed, as I think I said earlier, the amount of voicing given to high vowels in the "devoicing environment" is highly variable and controlled by many different factors, some of which may be related to the individual speaker. Regional dialect may be part of that, too - I never noticed myself. As for gender, I think it's not quite that men are more likely to devoice, per se, but simply that there is sociolinguistic pressure for women to speak more clearly and enunciate better.

Dave