PhDniX's blog

Chinese Phonology is beautiful!

All right, I dug up a table of the different syllables that exist in the Chinese language.* And I figured now would be as good a time as any to reduce the horribly unphonemic pinyin to the truely underlying phonemes.

I was surprised to find out that most gaps are easily explicable, and that Chinese only has two phonemic vowels: a and e/o. I named e/o E, it seems that o occurs when it is near labials.

Now according to my theory, every word is made out of this phonemic structure:
C-Y-V-(Y/N)

Where C is the initial consonant.
Y is the glide y or w
V is the vowel a or e.
and N is the nasal n or ng.

All of these parts can be either there, or not, as long as there's at least one element, it'll be a valid syllable (although some are restricted)

Now there's some interesting things to be seen.
Labial+w initial clusters don't exist except when not followed by a vowel. Which is easy enough to understand due to the labial nature of that glide. Besides that, it's interesting to notice that the c, ch, s, sh, z, zh and r don't occur when there's no glide on second position.

The only problem I had was getting rid of the words with ü, like nü. Right now I have that one as /nyay/ which is almost definitely wrong I can't think of any good reason why a phonemic /nyay/ would display [ny] phonetically. Nevertheless, the  ü  glides  are much too small of a class to actually be phonemic, or so I'd like to think. It is of course possible.

I was delighted to see that j/q/x were just allophones g/k/h in front of front vowels and the y glide.

now there's the word final r. that comes from enclitic er. I think that this might phonemically be /ra/ but I'll have to look into it.

Now I'll provide a little pdf, which will need some explanation. The first three pages are pretty much the pinyin spelling, but compressed into the phonemic system which I think looks right. The next three pages contain the same table, but this time with my proposed phonemic spelling. the purple background entries are the ones I'm least sure about.

Please enjoy, and this is probably very confusing, I'll try to structure it more later on.

http://members.home.nl/marijnvanputten/chinphon.pdf

*out of this terrible habit of mine, whenever I refer to Chinese, I mean mandarin Chinese, which is the only Chinese language I'm remotely familiar with. Please excuse this perpetual misnaming.

Ninja Edit:

It seems like I'm not the only idiot who came up with the two vowel system:
http://xiang.free.fr/leiden-en.pdf
See page four of this article.

I'm not really sure whether I should be happy or whether I should cry. On the one hand, it's amazing that I found this myself in one day, without any prior knowledge of this research, on the other hand, I did not find a breakthrough new theory. Haha. Oh well. They do still analyse that darned ü to exist though. I'd like to have that out of the way. But signs indicate that I probably cant. Oh well.

I'll soon make a full sketch of the Mandarin Phonology the way I see it.

Of wolves and wloves.

This blog entry I'd like to dedicate to a rather controversial theory of mine. Namely, the cause for sporadic sound changes.

Sporadic sound changes have always been something, we historical linguists would rather ignore, because it doesn't fit in well with our theories that sound changes occur through perfectly regular set rules.

Some people are convinced that these sporadic changes aren't so much sporadic, but rather we haven't found the system behind it yet. I personally think this theory is too positive, I sincerely doubt some of your sporadic changes are actually non-sporadic.

What are sporadic changes you say? Let me give you an example.

The word for "Horse" in English and Dutch give a nice example of a sporadic change.
English has: 'Horse'
Dutch has:   'Ros'
'Ros' comes from an earlier 'hros'.
I'm not completely sure which of the two words was first, but if we can believe etymonline.com it's the 'horse' form. As you can see Dutch has then just switched around the r and the o. This does not happen in other words, and is truely a sporadic change.

A similar switch we can see in the word 'wolf' in Proto-Germanic; *wulfaz. The PIE word for this is *ulkʷos or *lukʷos.
Compare: Sanskrit vrkas, Latin Lupus, Greek λυκος.

As you can see, for some unexplained reason the root for wolf seems to alternate between 'lwof' and 'wolf' in the different Indo-European languages.

This reminded me of a rather modern form of slang from France called Verlan'.

Verlan is an inversion of the syllables in the word word 'l'envers'.

In Verlan, the point is that you change around the order of either the syllables or letters itself. Another example of verlan is the rather politically loaded 'Racaille' which president Sarkozy so 'tactfully' used to describe the people of the Parisian banlieues. The Verlan version of this word is 'Kaillera'.

Other examples are keufli for flics 'coppers' and ouam for moi 'me'. No lets hypothetically say that one of these words becomes so popular that we stop using the non-Verlan word, and if even later all accounts of Verlan were lost, you'd get the impression that without an apparent reason a word had just completely turned itself inside out.

And I think this is exactly what causes these bizarre sporadic changes, playfulness in language, like Verlan. Other people have already previously suggested such things. Like how 'wolf' can turn itself inside out, and how some languages have a -p- rather than a -kʷ-. This was apparently done because there was a taboo on the word 'wolf', maybe because the early Indo-Europeans found this creature so scary that it was considered bad luck to pronounce its name correctly.

Now we get to the more far out part of my theory, which you may readily disagree with, because I have no good way to prove it.

Me and a very good friend of mine, Leah, have in the past years created a sort of idiolect which combines Dutch and English with internet language and our own little changes. By now people have lost all track of our conversations, but we generally understand each other (onderstad00!).

Another feature of this little dialect of ours is that we randomly transpose letters. We pronounce words like 'wolf' for example as 'wlof'. Why the hell do we talk about wolves? Well it's a sort of religious object, the so called aardappelwolf or aardappelwlof, don't ask.

But all of a sudden when I started saying 'wlof' I realised that this disturbingly much like what happened to the Indo-European language. Now my theory is, that some cultures have a certain predisposition to certain sporadic changes. In Germanic languages you should never be surprised to see resonant-vowel inversions. In Romance languages you shouldn't be too surprised to see inexplicable vowel changes. So maybe, this is part of our cultural heritage, a sort of inherited language game, much like Verlan is passed on from person to person.

Who knows? It doesn't really help us predict sporadic changes, but understanding which 'sporadic change games' are more prone to happen in which language family will greatly help us with reconstruction of words that have undergone these changes.

Rendaku in Japanese

I'm going to break my promise, because my Tangut text is giving me severe troubles to translate. Therefore I'll now give a small, English summary of what my paper on Historical Linguistics was about.

Namely about Historical Causes of Rendaku in the Japanese language. I'll also attach a link to the actual paper, which will only be worth reading if you speak Dutch, but who knows, I might have millions of non-replying Dutch readers. (right.)

First off what is rendaku?

Rendaku is the process in modern Japanese in which the first consonant of the second element of a compound becomes voiced.

For example
koi 'love/passion' + hito 'person' = koibito 'lover' (h is historically p, so b is the voiced variant of h).

In modern Japanese this process is fairly unpredictable. My original hypothesis was that historically there must be some logical explanation for this. Further research showed that it isn't one 'cause' which caused all rendaku, but actually multiple processes working simultaneously and after each other, while undergoing vast amounts of analogy to 'fix' some words with or without rendaku again.

I won't go into all these processes, for it would take me about 20 pages to properly write them down. I'll only discuss one cause for rendaku, which I researched in my paper.

It's the rendaku that occurs through the simplification of the genitive particle "no". My theory is that many of the rendaku that has occurred between the Old and Middle Japanese period is through the elision of "no".

Originally the voicing of rendaku wasn't so much voicing, but prenasalisation. So it's logical to think that a syllable with a nasal is the cause of rendaku. But just having this hunch, of course isn't enough. You're also going to need proof, and proof is what I found.

There are very very many surnames in Japanese that undergo this rendaku. Japanese names are often place names.
And example of one of those surnames undergoing rendaku is "isozaki" which translates to "Beach of the Cape".

Even in modern Japanese, if you would actually want to talk about the physical place "beach of the cape" you would not use "isozaki" like the surname but you would use the genitive particle "no", like so: "iso no saki".

If you then have a look at the older Japanese, you will not find such surnames which are different from the place names. The above example would both be "iso nö saki", whether it's the place name or the surname.

There is many many examples of this. Thus my theory is that, in colloquial speech, in environments where "iso no saki" was seen as a unit rather than 2 nouns and a particle, the vowel was often elided, and then remaining n caused prenasalisation.

This makes it especially easy to understand that this especially happens in surnames, where the meaning isn't really important. Later on this colloquialism was lexicalised as the way said surname was to be pronounced.

It'd be a bit odd to conclude this is what happened purely on surnames though. Luckily there's also several words which have undergone the same kind of "no" elision as in surnames.

For example 'spring mist' harugasumi. which comes from haru 'spring' + no + kasumi 'mist'. Even now you could probably still describe spring mist as 'haru no kasumi', but the word with this meaning is 'harugasumi'. In Old Japanese though, there's no such thing as harugasumi, only "paru nö kasumi".

There's also some pairs of surnames and nouns, which were originally the same, but in which the surname lost its "no" particle.

For example ama no kawa (litt. river of the sky) "milky way" also occurs as a surname, but then it is written "amagawa".
Another one is ko no ha "leaves" which also exist as a surname but then written "koba".

I could still write a lot more about this interesting thing, but you'll just have to see for yourself in my paper. I consider it as proven that the genitive particle "no" is indeed the cause of some of the rendaku that has occurred in between Old Japanese and Middle Japanese.

This doesn't explain though, the already existing Rendaku in Old Japanese. Whether we'll ever be able to explain that, I don't know, I  don't think we will.

(I'll include a link to the paper later on, if anyone is interested.)

P.S. Just for the record I received a 8/10 mark for this paper, so I guess whatever I'm saying makes enough sense :P

Edit:

I've decided to upload my paper I wrote, it's a shame it's in Dutch, but hopefully those who are interested wont experience this as too big of a barrier.

http://members.home.nl/marijnvanputten/rendaku.pdf

And Bodhisattva Inexhaustible Intent said:....

Just to keep up the pa ce in talking a bit more about Tangut, I have decided to do a small sentence analysis from a passage of the Lotus Sutra. It's not a direct translation from Sanskrit, but rather a translation from Chinese, which will soon become apparent.

This is what the (short) Sentence looks like:

Sentence

This sentence consists of 3 phrases. I'll discuss them induvidually.

Phrase1

This is the subject of the sentence. It means 'Bodhisattva of Inexhaustible Intent' This is a direct translation of the name that the Chinese use:
無盡意菩薩 non-exhaustible-intent-bodhi-sattva. The Tangut use a slightly different order and have intent-non-exhastible-bodhi-sattva. Note that there is no case marker, although technically it could have one. Since it's a transitive sentence this would be the Ergative case. But the Ergative is not often marked, and especially not when the only other noun phrase in the sentence already has the Gentive-Absolutive (which it does).

The following is the second phrase: (sorry for not getting the pictures in uniform sizes)

Phrase2

This simply reads Buddha+absolutive.  The absolutive also functions as a marking of the genitive, and sometimes even as a marking of the dative.  Tangut only marks this case marking and the ergative, where the ergative even seems to be constructed (The characters of the ergative mean something like: "He who does the action"). So the use of the limited amount of case markers is quite wide.



The last phrase is the verb phrase, this is also the typical order of words: SOV, which shouldn't come as a surprise to you if you're known with Sino-Tibetan languages, unless you only know Chinese, which is SVO (though there are dialects with a stronger SOV tendency than Mandarin if I'm not misaken).

This is the last phrase:

Phrase3

This reads word-say `to say words'. Although technically 'word' would be the the Object of the sentence, here it is definitely to be taken as part of the verb, rather than two loose words. Note that there is no agreement. In the 3rd person there's never any agreement in number or person. Or rather, the absence of a suffix marks the 3rd person, while suffixes mark the first and second person.

There is some indication of the subject agreement though. There are two characters for the verb `speak' in Tangut. One is the so called `Theme 1' and the other one is `Theme 2'. This is in fact the same word, but with a different vowel, or rhyme. For `speaking' Gong reconstructs tshjij-1 for the first theme. And tshii-2 for theme 2. These vowel changes also have to do with the verbal agreement. The 3rd singular/plural agreement used Theme 1.

Altogether this sentence means: "The Bodhisattva of Inexhaustible intent said to the Buddha:".

That's an open ending. Tomorrow I'll probably discuss what will follow.

Tangut

When I started this blog, I intended to write a lot about my pet project: the Tangut language. After I started my blog, I found out that I suddenly had a lot of idea's on Indo-European, and due to my study this overshadowed anything I was doing with Tangut at the time.

But now I decided I would finally like to at least make a small start on telling a bit about this incredibly interesting language. All of this is self study, with my sources mostly taken from the internet, some inaccuracies might occur, although I doubt you'll notice much of it in this introduction.

Tangut is the Russian name (Тангут) used for the language which is otherwise known by it's Chinese name Xī-Xià (西夏). For simplicity's sake I'll stick with Tangut. The language was spoken in the Tangut Empire, which was situated North West of China.

The language is a Qianic-Tangut language, which is a branch of the Tibeto-Burman languages (Ignoring any of George van Driem's rather controversial theories on what the Sino-Tibetan language family looks like). The language in itself isn't that remarkable, it's much like it's related languages, and anyone familiar with Tibeto-Burman languages wont be surprised to find it's a Ergative-Absolutive language, with a SOV word order and, typical to its branch of Languages, marks the direction of action as a prefix to the verb.

What is  very exciting though, is its script. Tangut has a completely ideographic script* much like the Chinese script, and it has over 6000 characters to write the language. To me, Tangut is one of the most elegant and beautiful scripts that I've ever seen, although it's also a nightmare, due to lack of clear definition of the phonetics behind Tangut, and because the grammar isn't too well suited for an ideographic script like this.

My Icon that I use actually is a Tangut word, meaning 'I', and is also a suffix in verbal agreement. I've been working on a dictionary, which isn't coming along very quickly, but I am progressing bit by bit.

The reason why I've never posted about this language before, is also because it is hard to find example sentences. Not easy ones, or ones without any problems in explaining which character means what. Nevertheless I hope to post up some interesting information about a couple of sentences in Tangut soon.

For more information see:
http://www.omniglot.com/writing/tangut.htm
http://en.wikipedia.org/wiki/Tangut_Empire
http://en.wikipedia.org/wiki/Tangut_Language

*I've called it Ideographic Script, though ideo-phonetic, or semanto-phonetic script might be better.

Futuristic Thematic Vowel Switch

I've just returned from my holiday to Brittanny, France. Of course I picked up a Breton grammar, a dictionary and a book, I wouldn't recognise myself if I wouldn't put time and money into a minority language.

So here I am, writing a small entry on Breton. First of all, it's a very fun language, it's like welsh spoken by French people (which is more or less the case).

What I found incredibly striking when I was reading about the verb, was the formation of the future tense, it has to do with the thematic vowels.

First of all I shall explain what thematic vowels are.

In Proto Indo-European, there's two types of verbs, thematic and athematic ones. An athematic verb is simply the stem followed by its endings. With thematic verbs though, in between the stem and the ending there's a placement of a vowel (either e/o).

So the thematic paradigm looks something like this:

1st singular: *-o-h2
2nd singular: *-e-h1i
3rd singular: *-e
1st plural: *-o-mom
2nd plural: *-e-th1e
3rd plural: *-o

These are the primary endings for the present tense. (and maybe previously only the Subjunctive.)

These thematic endings always have this distribution, and also occur in the secondary (aorist/imperfect endings) and also in the forms of the middle voice.

The distribution is: o goes with 1sg, 1pl and 3pl. While e goes with 2sg, 3sg and 2pl.

Breton more or less follows this pattern:

scriv-a-n `I write'
scriv-e-z `you write'
scriv- `he writes'
scriv-o-mp `we write'
scriv-i-t `you (pl.) write'
scriv-e-r `they write'

both a and o go back to *o. and i and e go back to *e. scriv is quite remarkable, since there's no ending, this is probably due to some law which cuts of word final e's though.

Remarkable is the -e- in the 3rd plural, we'd expect the -o-. Also the ending is remarkable, it seems to be the ending of the perfect/medium rather than the present. but this discrepancy wasn't what I wanted to talk about, the really remarkable thing is when you look at the formation of the future tense:

scriv-i-n `I will write'
scriv-i- `you will write'
scriv-o `he will write'
scriv-i-mp `we will write'
scriv-o-t `you (pl.) will write'
scriv-o-r `they will write'

Can you see what happened? The distribution (except for the 2nd singular) of the e's and o's switches around! So the distribution now is 1st, 2nd singular and 1st plural take *e, while 3rd singular, 2nd and 3rd plural take *o.

What does this mean? That the Bretons (or maybe the Kelts? I'm not to familiar with the Celtic languages) made an interesting innovation which made use of the seemingly meaningless existence of the thematic vowel to form the future tense. It's unlikely that this has anything to do with how Indo-European handled it's future, but it's still a very fun quirk to see in a language.