The Correspondence Library

[finlay]

Nippon and Nihon are varieties of the same (that is, I presume originally *ni-ppon and *ni-pon), I believe.

Well, no, niQ-pon and ni-pon, where niQ and ni are variations of the same thing and Q stands for a geminated consonant. (Arguably, it could also be nit-pon because the normal form of this morpheme is にち (niti or nichi depending on your preference))

In hiragana, にっぽん and にほん; the identity between the two is preserved by the fact that ぽ is a variation of ほ in the writing system. にっほん would be inadmissible, so it must turn into にっぽん.

Another example is 600, which is ろっぴゃく (roppyaku) from ろく+ひゃく (roku+hyaku); the く becomes a geminate consonant as っ and the ひ must become ぴ because ろっひゃく (rohhyaku) is against the rules.

You do occasionally see geminate h used to transcribe [x] in other languages, like バッハ (bahha) for Bach.

[clawgrip]

Nevertheless, there are certain compounds that have undergone the change. The ending -hara often changes to -wara because of this rule, e.g. Fujiwara, Yoshiwara. Also 八幡 can be Yahata or Yawata.

Are there any cases when the ending happens after /ɴ/? What then?

/ɴ/ usually turns /h/ into /b/ or /p/:
何 /naɴ/ 'what' + 本 /hoɴ/ 'long, thin object' = 何本 /naɴboɴ/ 'how many (long, thin objects)'
金 /kiɴ/ 'gold' + 髪 /hatu/ 'hair' = 金髪 /kiɴpatu/ 'blond hair'

My intuition tells me the choice between /p/ and /b/ has to do with the preceding vowel quality, but I haven't researched it.

四 /joɴ/ 'four' is a notable exception, typically leaving the /h/ unchanged:

四本 /joɴhoɴ/ 'four long, thin objects'

Where did the geminates come from?

Originally geminates come from compounding. /tu/ will reliably become Q as in finlay's post above when it comes immediately before an unvoiced stop or fricative. In front of /h/, it will cause a geminate /pp/

発 /hatu/ 'emit; release' + 車 /ʃa/ 'wheeled vehicle' = 発車 /haʃʃa/ 'vehicle departing'
物 /butu/ 'thing' + 価 /ka/ 'value' = 物価 /bukka/ 'prices; cost of living'
合 /ɡatu/ 'combine; fit together' + 併 /he:/ 'combine' = 合併 /ɡappe:/ 'merger; annexation'

/ku/ will also fairly reliably form geminates when it comes before /k/, e.g.
特 /toku/ 'special' + 許 /kjo/ 'permission' = 特許 /tokkjo/ 'patent'
白 /haku/ 'white' + 血 /ketu/ 'blood' + 病 /bjo:/ 'disease' = 白血病 /hakketubjo:/ 'leukemia'

There is a resistance to this in official word use when compounds are further compounded, e.g.
水族 /suizoku/ 'sea life' + 館 /kaɴ/ 'hall' = 水族館 /suizokukaɴ/ 'aquarium (i.e. a building for fish viewing, not a single tank)'
洗濯 /seɴtaku/ 'laundry' + 機 /ki/ 'machine' = 洗濯機 /seɴtakuki/ 'washing machine'

However, in informal speech, these are invariably pronounced /suizokkaɴ/ and /seɴtakki/.

Certain basic words also have irregular geminate-forming properties:

六 /roku/ 'six' ends in /ku/ but will form geminates with any stop;
日 /ni(ti)/ 'sun; day; Japan' and 八 /hati/ 'eight' will form geminates just like morphemes with final /tu/

Sometimes a geminate may spontaneously occur to retain the mora-length of a word when another sound is lost for some reason, e.g.
火 /hi/ 'fire' + 男 /otoko/ 'man' = ひょっとこ /hjottoko/ 'hyottoko, a traditional mask character' (reduction of /i/ to /j/ meant a loss of moraic /i/, causing subsequent gemination of /t/)
真 /ma/ 'true' + 赤 /aka/ 'red' = 真赤 /makka/ 'bright red' (loss of one moraic /a/ resulted in gemination of /k/).

In contemporary Japanese, geminates are also common in onomatopoeic expressions, such as びっくり /bikkuri/ 'surprised' or がっかり /ɡakkari/ 'dejected'. They also occur frequently in English loanwords in place of an English short vowel, cf.:

ビート /bi:to/ 'beat'
ビット /bitto/ 'bit'

The common geminates -tta/-tte in verb conjugations come from contractions of older grammatical forms combining with /ɸ/, /t/, and /r/

Middle Japanese 思ふ /omoɸu/ + たり /tari/ = 思ひたり /omoɸitari/ -> */omoɸtari/ -> omotta
Middle Japanese 勝つ /katu/ + たり /tari/ = 勝ちたり /katitari/ -> */kattari/ -> /katta/
MIddle Japanese 取る /toru/ + て /te/ = 取りて /torite/ -> */torte/ -> /totte/

* This is theoretical. I don't know exactly how the sound changes and grammatical changes relate to each other chronologically, or if forms like */torte/ actually ever existed at all.

Also I had a personal theorem regarding the vowels, dunno whether to show it.

Share your ideas.

P.S. I'm not sure if I'm supposed to write /naɴboɴ/ or /namboɴ/ here. I figured /naɴboɴ/ was more transparent, and the resulting sound would be predictable anyway based on allophony rules.

[Bob Johnson]

P.S. I'm not sure if I'm supposed to write /naɴboɴ/ or /namboɴ/ here. I figured /naɴboɴ/ was more transparent, and the resulting sound would be predictable anyway based on allophony rules.

/naɴboɴ/ or even /naɴhoɴ/, but [nambõ] or somesuch. You don't write allophony in /fəˈniːmɪk/ notation.

[clawgrip]

Yeah, I figured /naɴboɴ/ was better. I didn't want to do /naɴhoɴ/ since /h/ and /b/ are not allophones, and preserving the base forms of morphemes would obscure the whole point the post. Thanks for confirming.

[Pole, the]

My idea was that the earliest vowel inventory we can reconstruct is something like "square" /i e : u a/; o₂ being /e/. (a)
(Or maybe it was even /i e : u o : a/, with /a/ being neutral, later collapsing /o a/ > /a/.)
The /e/ could later reduce to /ə/ (producing a "triangle" vowel system /i u a ə/) or go even more back ([ɤ]?). (b)
The next stage is the monophthongization: /ai/ > e₂, /əi ui/ > i₂, /ia/ > e₁, /au ua/ > o₁. I suggest that the fronting / palatalization is the main difference between both e₁ e₂ (/ʲe ɘ/?) and i₁ i₂ (/ʲi ɨ/?). (c)
Then, /ʲe ɘ/, /ʲi ɨ/, /o ə/ collapse.

(a) The hints can be cases of the ablaut, cf. hito < pito₂, futa < puta.
(b) The o₁ : o₂ opposition is correlated with the labialness – it doesn't occur after /p b w/. The cause could be that /ə/ labialized after these consonants first, then merged also in other backgrounds.
(c) The oppositions are correlated with the coronalness (except of /w/) - they don't occur after /t d s z n r w/. The coronals could be the first affected with the palatalization. If i₁ i₂ were /i wi/, they would likely follow the correlation of o₁ o₂.

But, you know, these are only my ideas, maybe there are much more logical explanations.

[jmcd]

One idea I have for why some inter-morpheme /ɴh/ compounds turn into [mb] and others into [mp] is when the compound gets created:

Some compounds are created
rule p/b/ɴ_ applies to existing compounds
More compounds are created
p>ɸ (except existing [mp] created from previous stage of compounding)
More compounds are created

[clawgrip]

I think it is strongly lexical in nature.

When a change takes place after /ɴ/, native Japanese roots always follow /h/ → /b/, never /h/ → /p/

Sinitic roots seem to favour one or the other individually. For example:
波 /ha/, 費 /hi/, 粉 /huɴ/, 辺 /heɴ/, 歩 /ho/　always change /h/ to /p/ after /ɴ/, never to /b/.
羽 /ha/, 仏 /hutu/, 偏 /heɴ/, 北 /hoku/ always change /h/ to /b/ after /ɴ/, never to /p/.

Some, like 本 hon, can change to both, e.g. 三本 /saɴboɴ/ but 原本 /ɡeɴpoɴ/

I can't find a Sinitic root with /hi/ that changes to /bi/, so there may be some kind of phonetic issue at work as well.

[Herr Dunkel]

Rendaku?
p > b RENDAKU
p > h NORMAL, ergo
h > b RENDAKU

[clawgrip]

Yes, it is rendaku. The question is simply what decides whether rendaku of /ɴ/ + /h/ will result in a /p/ or a /b/.

For example:
/h/ → /p/: 担 /taɴ/ 'bear' + 保 /ho/ 'safeguard' becomes 担保 /taɴpo/ 'collateral'
/h/ → /b/: 田 /taɴ/ 'rice paddy' + 圃 /ho/ 'cultivated field' becomes 田圃 /taɴbo/ 'cultivated field'

[Herr Dunkel]

Words that originally had /p/ rendaku themelves into /b/, and words loaned with /h/ fortify/somethingate into /p/ or something.

I'm no expert, but this seems likely.

[Herr Dunkel]

Proto-Germanic > Proto-Norse Sound Changes

It's pretty hard to find anything reasonable on Proto-Norse on the internet, but I think I can cover most of the major changes, from what I have obseved.

*æ > ā
*ē > ā
*au > ō / unstressed
*ai > ē / unstressed
*z > ʀ
*u > o / before an approximant ?
*ww > ggw
*gj/jj > ggj
*an > ā / _C (?)
*ōn > an / _#
*w > 0 / _u
*w > 0 / _a, some other conditions because (*meðwaz > meðwaʀ)
*w > 0 / _o, probably something else since we have "Óðinn" but "voði"
*β > 0 / V_V[+back] (?)
*i > e / in the initial syllable before more than one consonant - (*þriznaz > þreʀnaʀ)
*we > i (?)


This is an incomplete list, and not really chronological (I half-tried putting them in some sort of order). It's pretty short because there isn't that much time seperating Proto-Norse and Proto-Germanic as we know it.

[Pogostick Man]

EDIT: I've fixed some additional things and it seems I may have missed some things…God willing, I would much like to fix it.

As per Theta/Tieđđá's request…let's have some PROBABLY TOTALLY AWFUL AND VERY INCOMPLETE Vietnamese/Muong sound changes! (I've been here since about 10:30 this morning. It's now 18:28 20:24, though admittedly I was in class, on IRC, &c., for a bunch of the time earlier. I've also gone through about a NOS and a half two NOSes. Can you tell?) If there's any major problems with this, Whimemsz or whoever else, feel free to slap me upside the head.

Any way you slice it, this thing's going to be a mess.

My sources for these are mainly as follows:

• Thompson, Laurence C. (1976). "Proto-Viet-Muong Phonology". Oceanic Linguistics Special Publications, 13(2): 1113 – 1203
• Wikipedia contributors (2012). "Hanoi". Wikipedia, the Free Encyclopedia. http://en.wikipedia.org/w/index.php?tit ... =509052974
• Wikipedia contributors (2012). "Vietmanese Language". Wikipedia, the Free Encyclopedia. http://en.wikipedia.org/w/index.php?tit ... =509331797

Two other things deserve honorable mention, mainly because I did next to nothing with them, but still looked at them briefly, so they're listed here for completeness:

• Gage, William W. (1985). "Glottal Stops and Vietnamese Tonogenesis". Oceanic Linguistics Special Publications, 20: 21 – 36 — I didn't really read this one so much as open it, get exasperated, and p. much ignore it.
• Thompson, Laurence C. (1979, I guess). "More on Viet-Muong Tonal Developments". — Again, I only glossed over this one, but it's mentioned here for completeness.

Now, I'm going to tackle the tonogenesis first, because it'll a) save me time and b) be less messy (but mostly a)). Proto-Viet-Muong is said here to have had four main tones, *A *B *C *D. Tone *B is thought to be related to a final glottal stop; tone *D to some other final stop; *C, from voiceless fricatives ("spirants"); and *A, some other final obstruent. Now, a tone split occurred sometime in the history of these languages, such that you got two registers of tones, the first from voiceless and the second from voiced onsets. Somewhere on the road to Vietnamese, *D and *B apparently merged. If I can read my chicken-scratch handwriting…

Code: Select all

           HANOI VIETNAMESE TONOGENESIS
 Reg |      A       |      B/D      |      C
---------------------------------------------------
  1  | mid trailing |  high rising  |   dipping
  2  | low trailing | low dropping¹ | high rising² 

¹ Tense before _S#; laryngealized else
² Laryngealized

A bit of guesswork on this one (I'm assuming the markings used in the document denote cognate tones), but…

Code: Select all

           SAIGON VIETNAMESE TONOGENESIS
 Reg |      A       |      B/D      |      C
---------------------------------------------------
  1  | mid trailing |  high rising  |  mid rising
  2  | low trailing |     low¹      | high rising² 

¹ Level before a final stop; dipping otherwise
² Laryngealized

And…

Code: Select all

                         MUONG TONOGENESIS
 Reg |      A       |       B       |      C      |      D
----------------------------------------------------------------
  1  |  mid level   |  low rising¹  | high rising | high rising
  2  | low falling  |   high-mid²   |  high-mid²  |  high-mid²  

¹ "Constricted" (Laryngealized?)
² Terminates in a glottal stop if no final stop

There was some stuff going on with presyllables and tone registers that I do not quite fully understand. For example, Thompson gives *həp- as yielding p-R1 in Muong, v-R1 in Hanoi, bj-R1 in Saigon, and β-R1 in Middle Vietnamese, but *həb- yields these same outcomes with the exception of in Muong, where the second-register tone occurs. I wrote down in my notes that apparently, when resonants occur, no presyllable was present when a first-register tone occurs, but there was when a second-register tone now is.

Asterisk'd /t d/ indicate some weird random reconstructed phoneme that Thompson doesn't really elaborate on save to say that its's not your garden-variety /t d/. The variable W here is an approximant; K represents a velar. It seems that /w/ is treated more as a bilabial than as a velar in Thompson's paper.

Proto-Viet-Muong is reconstructed by Thompson as having had presyllables of the form *(h)ə- which were later lost, phonemicizing voicing/tonal distinctions. I assume that initial *ʔ might have had some sort of a presyllable in some cases which was then lost but it didn't really matter because you can't voice /ʔ/. Also, some of the sound changes from the Wikipedia are ascribed to either Hanoi or Saigon on the basis of the regional designations, so I'm not totally 100% sure if those changes belong to those dialects.

tl;dr – Go to Bonneville, Utah and take these sound changes with the salt flats there.

AVENGERS REALLY, REALLY MESSY AND DISORGANIZED SOUND CHANGES, ASSEMBLE

Proto-Viet-Muong to Middle Vietnamese
Initials:

• bʱ {tʰ,dʱ} {t*ʰ,d*ʱ} {kʰ,gʱ} → pʰ t tʰ kʰ (after *{kʰ,gʱ}, only first-register tones may occur)
• (h)ə{p,b} (h)ə{t,d} (h)ə{c,ɟ} (h)ə{k,g} → β dʲ ɟ g
• {p,b}j {t,d}j {c,ɟ}j → β/w dʲ ɟ
• ɓ ɗ → m n (For some reason it seems that only first-register tones can occur in this environment)
• {n̥,n,ɲ̊,ɲ}j → ɲ (Thompson appears to me to have hedged a bit on the last one; based on other evidence in the paper I'm sticking this one as a palatal nasal)
• tʃ → Ω (This is my own notation. I don't have a clue what the intermediate form was; became something else in different dialects)
• N[- voice] W[- voice] → N[+ voice] W[+ voice]
• ((h)ə)p → b / _l
• d → t / _l
• t → Ø / _ɹ (only first-register tones can occur in this environment)
• t*ʰ d g → tʰ t k / _w
• s → t(ʰ?)
• {əkʰ,əgɦj} → ɟ (I know, right? I think Thompson implied this was just a bit of a kludge anyways)

Finals:

• l → Ø / {i,e}_
• l → j / else
• c ɲ → t n / ! E_ (apparently the precursor to Vietnamese short *a was treated as a short vowel here)

Middle Vietnamese to Hanoi Vietnamese
Initials:

• pʰ → f
• kʰ → x (only seems to have occurred with first-register tones)
• {β,w} {dʲ,ɟ} → v z
• ɹ → z (only seems to have occurred with first-register tones)
• bl → z
• ml → mɲ → ɲ (Thompson seems to indicate that this may have become [l] as well; only seems to have occurred with second-register tones)
• Ω → s
• cʰ → tɕ

Vowels:

• ɨ → i / _(ə)w
• ɛ → a / _C[+ palatal]

Miscellanea:

• w → Ø / tV_wk (Conjectured)

Middle Vietnamese to Saigon Vietnamese
Initials:

• pʰ → f
• kʰ → x (only seems to have occurred with first-register tones)
• {β,w} → bj ~ vj ~ v
• {b,t}l → ʈ (?! Thompson seems to have denoted two sounds using a particular symbol…this is a mess)
• dʲ ɟ → z j
• ml → l
• Ω → ʂ
• cʰ → ʈʂ
• ɹ → ʐ (sometimes?)

Finals:

• c ɲ t n → t n k ŋ / a_ (not when long)
• c ɲ → t n / {i,e}_
• {c,ɲ} → Ø / else

Vowels:

• ə → Ø / {i,ɨ}_{p,m,w}
• ə → Ø / ɨ_j
• ə → Ø / u_{m,j}
• The contrast between short a and short ə is neutralized when _w{k,ŋ}
• a → aː / _{j,w}
• ə(ː) ɛ → ɨ ɛə / _K
• ɛ → a / _C[+ palatal]

Miscellanea:

• w → Ø / tV_wk (Conjectured)

Proto-Viet-Muong to Muong Khen
Presyllables don't seem to have affected Muong much.

Initials:

• s → h
• cʰ → s
• tʰ → h (Only seems to have occurred with first-register tones)
• {kʰ,gʰ} → x (Presyllables don't seem to have affected this much)
• m n → b d (Only seems to have occurred with first-register tones)
• {p,b}j {t,d}j {c,ɟ}j → b d j
• ɓ ɗ → b d (Only seems to have occurred with the first register)
• {n̥,n}j {ɲ̊,ɲ}j → ɲ j
• N[- voice] W[- voice] → N[+ voice] W[+ voice]
• (h)ə{p,b} → t / _l
• m → Ø / _l
• tɹ → tʰ

Miscellanea:

• w → Ø / tʰV_k (Conjectured)

Miscellanea

There's a couple of changes that are listed as apparently having occurred some time with effects in Modern Vietnamese that I'm not quite sure how to handle above, so I'm dumping them down here:

• a → ɨə
• ɔ → uə

The underlining is Thompson's.

There was some interesting occasional alternations given between Muong and Vietnamese. In some cases, Thompson posited a conditioning due to a preceding *ə (like maybe due to a presyllable or something). I'm not sure which two languages/dialects are being compared here due to the presence of the asterisk on the right-side forms, which might make it Middle Vietnamese; his default A : B format is Muong Khen : Hanoi Vietnamese.

• -o : *-əw
• -u : *-əw
• -i : *-əj
• -e : *-əj (IIRC this one was commented on as being rarer than the above)
• a : ɨa

This post has been edited a few times, mostly to clarify things but also to take some things out.

[Ser]

WTF does the capital omega even mean?

[WeepingElf]

WTF does the capital omega even mean?

An ad-hoc notation for a phoneme whose precise nature is uncertain.

[Hakaku]

Yes, it is rendaku. The question is simply what decides whether rendaku of /ɴ/ + /h/ will result in a /p/ or a /b/.

For example:
/h/ → /p/: 担 /taɴ/ 'bear' + 保 /ho/ 'safeguard' becomes 担保 /taɴpo/ 'collateral'
/h/ → /b/: 田 /taɴ/ 'rice paddy' + 圃 /ho/ 'cultivated field' becomes 田圃 /taɴbo/ 'cultivated field'

Could it be that you're looking at the wrong element? 担 historically ended in a final -m; while 田 ended in -n. Same with 三 (-m) in 三本 /saɴboɴ/ and 原 (-n) in 原本 /ɡeɴpoɴ/. This last example is probably a little more conclusive given that the second morpheme is the same.

Thus:
/m+p/ → /mb/ → /ɴb/
/n+p/ → /np/ → /ɴp/

[clawgrip]

Yes, it is rendaku. The question is simply what decides whether rendaku of /ɴ/ + /h/ will result in a /p/ or a /b/.

For example:
/h/ → /p/: 担 /taɴ/ 'bear' + 保 /ho/ 'safeguard' becomes 担保 /taɴpo/ 'collateral'
/h/ → /b/: 田 /taɴ/ 'rice paddy' + 圃 /ho/ 'cultivated field' becomes 田圃 /taɴbo/ 'cultivated field'

Could it be that you're looking at the wrong element? 担 historically ended in a final -m; while 田 ended in -n. Same with 三 (-m) in 三本 /saɴboɴ/ and 原 (-n) in 原本 /ɡeɴpoɴ/. This last example is probably a little more conclusive given that the second morpheme is the same.

Thus:
/m+p/ → /mb/ → /ɴb/
/n+p/ → /np/ → /ɴp/

An interesting idea, but unfortunately your two examples contradict each other:
担保 /tam + ho/ = /tampo/
三本 /sam + hoɴ/ = /samboɴ/

Admittedly, 本 seems to be a bit anomalous though, probably because it is a classifier in this instance, so it may not be a good example.

However, Initials with /-ɴ/are prone to variation, e.g.
原: 原本 /ɡeɴpoɴ/, 原版 /ɡeɴbaɴ/
難: 難破 /naɴpa/, 難波 /naɴba/
金: 金髪 /kiɴpatu/, 金歯 /kiɴba/
印: 印譜 /iɴpu/, 印旛 /iɴba/

As we have seen, aside from 本, which has probably gained a /b/ pronunciation through analogy with other classifiers such as 匹 and 杯 (though oddly, not 品), finals with /h-/ do not tend to vary, which suggests it's the specific initial-h morphemes that determine the specific realization of rendaku.

[Nortaneous]

requests: Scottish Gaelic, anything Germanic

[Qwynegold]

I think it is strongly lexical in nature.

When a change takes place after /ɴ/, native Japanese roots always follow /h/ → /b/, never /h/ → /p/

Sinitic roots seem to favour one or the other individually. For example:
波 /ha/, 費 /hi/, 粉 /huɴ/, 辺 /heɴ/, 歩 /ho/　always change /h/ to /p/ after /ɴ/, never to /b/.
羽 /ha/, 仏 /hutu/, 偏 /heɴ/, 北 /hoku/ always change /h/ to /b/ after /ɴ/, never to /p/.

Some, like 本 hon, can change to both, e.g. 三本 /saɴboɴ/ but 原本 /ɡeɴpoɴ/

I can't find a Sinitic root with /hi/ that changes to /bi/, so there may be some kind of phonetic issue at work as well.

Hmm, I looked a little into the modern Mandarin pronunciations of these characters.
波 - bo or po
費 - couldn't find this in my Japanese wordprocessor, and for some reason I can't paste in kanji from the internet
扮 - bàn
辺 - bian
歩 - bù

羽 - yû
仏 - fó or fú
偏 - pian
北 - bêi

So it seems like all the ones that become [p] in Japanese are also [p] in Chinese, while those that become in Japanese have all sorts of different pronunciations in Chinese.

[finlay]

費 is fèi in Mandarin.

also it's perhaps disingenuous to equate [p] in Mandarin and Japanese, since in one it represents the fortis member of a pair (ie, the 'p' sound) and in the other it's the lenis member (ie, the 'b' sound). Besides, the Japanese pronunciations tend to come from an older version of Wu or Cantonese pronunciations, not Mandarin.

[Wattmann]

I'm thinking of something for Icelandic, to dispell the misconceptions of its invariability!

Ø > u / C__ʀ#
ʀ > r
w > Ø / #__r
w > v
f > v / !!#__

V[+nas] > [-nas]


y(:) > i(:)
i: > i > ɪ
u: > u > ʏ*
ɔ(:) > ø*

aʏ > øʏ

o: > ou
a: > au
e: > je

ø: > ɛ:
ɛ: > ai

P: > ʰP [when stressed]

g > ɣ / V__V
P[-voice] > [+aspir]
P[+voice] > [-voice]

V > [+long] / __S**

A > [-voice] / #h__

*Those two changes seem to have a common origin, and the first looks like it was shared with Swedish (AFAIK only partially tho)
**S = short consonant

These are those changes which we find across (maybe almost?) all of Iceland's dialects. The northeast and south differ somewhat in some consonantal things, and I believe vowels just as much. Feel free to correct me!

[dhok]

Proto-Central-Algonquian to Shawnee

Taken from An Outline of Shawnee Historical Phonology, by Wick R. Miller, Vol. 25, #1, International Journal of American Linguistics. Shawnee is closely related to Sac, Fox and Kickapoo and a bit further down the road to Miami-Illinois. Shawnee has some bizarre shit going on with preglottalized clusters; there was also a syncope process which made things even hairier, and is not entirely understood because "Shawnee morphophonemics are yet to be stated". There wasn't any chronology presented, but here's what I could get out of it:

(using variables S= short vowel, L=long vowel)

θ -> l
s -> θ
Ø->h/#_
we, wa -> ho/#_ (paper said that probably we, wa -> o first, then the above change happened)
θ, s, l -> θ/[qh]_
nl -> n
N -> Ø /_C (Miller: "leaving no trace in vocalic syncope")
h, q, x, t, č -> ʔ/_C
s->š/_C
ʔ -> Ø /[-voiced]_C
ʔ -> Ø /#[+voiced]_C
Ø -> ʔ/[+voiced]_šC
Ø -> ʔ/[+voiced]_h (except in initial position)
S -> Ø/_šC (except when preceded by *w)
V -> Ø/_h
S -> S, Ø /[+voice]_ʔC, #_ʔC (unless preceded by *w; there seems to be an "alternation" between the short vowel and zero)
S -> Ø /[-voice]_ʔC
L -> S/_ʔC, _šC
e -> i, e /#_ (we have only two examples of initial *e, one of which became e, the other i, with initial h- added of course)

[Whimemsz]

Some updates/corrections to the PA > Shawnee changes. First, note that Miller's paper is quite old, and some of the information is outdated (e.g., there's no such thing as Proto-Central Algonquian, Shawnee isn't necessarily closest to Fox-Sauk-Kickapoo, and some of the reconstructions are inaccurate or use confusing notations).

So, my best effort, incorporating information from Towhid Bin Muzaffar's MA dissertation, Computer Simulation of Shawnee Historical Phonology and with some other corrections based on a few other papers plus my limited knowledge of comparative Algonquian:

(O = obstruent)

we → o
θ → r (but not when before an obstruent)
r → s /{h,ʔ}__
nr → n
N → Ø /__O
{h,s,ʧ,θ} → ʔ /__O
r → ʃ /__O
e → i /#(C)__ (but remains e in a few cases?)
iː → i /__j
j → Ø /C__i(ː)
je → i /C__
j → Ø /{ʧ,ʃ}__eː
j → Ø /{ʧ,ʃ,w}__aː
w → Ø /t__i
wa → o /#__
V[short] → Ø /__{ʃp,ʃk}, C__ʔC, __hV (but PA initial *we does not syncopate before ʃp/ʃk)
Vː → V /__#, __{ʔC,ʃp,ʃk}, __hV
ʔ → Ø /C[vls]__C, __CC
Ø → ʔ /C[vd]__{ʃp,ʃk,hV}
Ø → h /#__V
s → θ
r → l
Ø → i /#C__jVː (for some speakers)
ʃ → s (for many speakers)

[Whimemsz]

Proto-Algonquian > Menominee

From Hockett, C. F. (1981). “The Phonological History of Menominee.” Anthropological Linguistics 23(2): 51-87; and Miner, Kenneth L. (1979). “Theoretical Implications of the Great Menominee Vowel Shift.” Kansas Working Papers in Linguistics 4(1): 7-25. The two papers don't agree on all points, and I've rearranged a few changes in an attempt to come up with what seems to be the most reasonable merger of the two.

The reconstructed Proto-Algonquian phoneme inventory is here.

In the notations below, "O" = "obstruent," "%" = "separated by one syllable", and "G" = "glide" ([w] or [j]). "Glottal word" means any word with a ʔ+O cluster following the first vowel, and a "nonglottal word" means all other words. "Even" refers to every other syllable counting rightward from the beginning of the word or from the most recent long vowel. Other vowels are termed "odd."


we, je → o, i /C__
we → o /#__
θ → s /__O
θ → r
V[short] → Ø /__# [does not apply in disyllabic words containing two short vowels]
Ø → h /V[short]__#
Hm → m
s, r → h /__O
w → Ø /h__V
a → o /m__w [in the suffix *-amaw]
V[short] → [long] [when V is the second vowel of a word and follows a short-vowel syllable. Does not apply in glottal words]
e → i /V[long]%__ [does not apply when /__ʔ, h]
N → h /__O, __r
e → i /#(C)__ [does not apply when /__ʔ, h]
e → i /__k, m [when in the second syllable of glottal words]
w, j → Ø /C__#
C → Ø /C__#
wi(ː) → o(ː) /C__w
ʃ, ʧ → s, ts
V[long] → [short] /CC(G)__C{V,#} [i.e., when following a cluster but not followed by a cluster. Only applies “after the first long vowel of a nonglottal word, and everywhere in a glottal word”]
V[short, even] → [long] /__CC
V[long, even] → [short] /__C{V,#} [does not apply when in second syllable of a non-glottal word]
e → æ
eː → æː
i → e
iː, oː, oʔ → eː, uː, uʔ [blocked when iː or a C+G sequence follows anywhere in the word, but does apply if æ(ː) intervenes before any following iː or C+G]
wiː, jiː, weː, jeː, wæː, jæː → iː /C__
wi, ji, we, je, wæ, jæ → i /C__
æ[odd] → e [except /__h, w, j, ʔ]
r → n
wa, ja → uə̯, iə̯ /C__

[dhok]

PIE to Ancient Greek

Taken from Voyles and Barrack, An Introduction to Proto-Indo-European and the Early Indo-European Languages, a great volume with full sound changes for a representative of most of the branches of IE, with the notable and lamentable exclusion of Baltic and Tocharian. I included examples for most of these changes but then got lazy towards the end. I'll probably start on Sanskrit tonight.

1. Early PIE -> Late-Stage

The authors take most of this information from Sihler 1995 and Beekes 1969. Under the account presented. PIE at this stage is conjectured to have four short vowels /i e o u/, plus h₁, h₂, h₃; the symbol <h> in the sound changes presented designates any laryngeal.

The sound changes here are standard late-stage PIE reflexes:

[h₁ h₂ h₃]e -> [e a o]

e[h₁ h₂ h₃] -> [ē ā ō]

Syllabic [h₁ h₂ h₃] -> e a o

ih uh -> ī ū
ph th kh -> pʰ tʰ kʰ

In addition, the PIE palatovelars *ḱ ǵ ǵʰ became k g gʰ at this stage. The book I have does not include palatovelars (its authors believe they derive from clusters), but if you accept their existence, this is what happened.

2. Late-Stage to Greek

a. V: -> V /_RO where R is a sonorant, and O is an obstruent.
The above is known as Osthoff's Law. By the time Greek was written down, this rule failed to apply (so inherited ἔγνον from *egnōnt, but later-innovated φέρωνται, not *φέρονται.) Moreover, this rule did not apply when the long vowel derived from a contraction of a long and a short vowel, so e.g. τιμῶντες from earlier τιμᾱ + οντες.

b. bʰ dʰ gʰ gʰʷ -> pʰ tʰ kʰ kʰʷ
Devoicing of aspirated stops.

c. e: o: -> ɛ: ɔ:
"After this change...the rule was dropped from the grammar."

d. o -> u / X_X where X is either a labial or a sonorant (both must be present).
This is sometimes called Cowgill's Law, and "its exact environment is not clear"; it changed some lexical items and then dropped from the grammar. For example, PIE *nokʷts became *nukʷts, producing Greek νύξ. This occured before the next change, as, for example, the PIE labiovelars counted as labials in this change.

e. ṃ ṇ -> am an / _V, _y ; -> a / _C, _w, _#
ḷ ṛ -> al, ar OR la, ra
Here we have changes in syllabic sonorants. These should be straightforward: as an example, *gʷṃtis -> gʷatis -> βάσις. Syllabic ḷ ṛ changed as above, but whether the /a/ appeared before or after seems to be unpredictable: *mḷdʰakos -> μαλθακός, but *kṛdiā -> Attic καρδία, Homeric κραδίη.

f. X(X) -> Ø /_# where X is a stop.
Basically, one or two plosives, at the end of a word, drop, eg *pawid -> παῖ. Note that word-final stops of proclitics are unaffected, since they are phonologically part of the following word.

This is followed by a sketch of the Greek pitch-accent system. There is a lot of diachronics, but they're more complex than I can get assed to get into. The only major change was the recessive rule: the earliest the stress can occur in Greek is the antepenult.

g. Changes with a consonant followed by y:
[t tʰ]y -> s(s) / [+sonorant], #_
[p pʰ]y -> pt
[k kʰ kʷ kʷʰ]y -> s(s)
[g gʷ d]y -> dz
ly -> ll
my -> ny
Sihler believes that the [t tʰ]y change probably occured first. Examples: medʰyos -> μέσος, *skepyomai -> σκέπτομαι, bʰulakyō -> φυλάσσω, etc.

h. y -> h and sometimes dz /#_
y -> Ø /V_V
Examples: *yos -> ὅς, *yugom -> ζυγόν, *dweos -> δέος.

i. VC[y w] -> V[y w]C -> V[iu]C
Metathesis of y and w, leading to some lexical restructuring. Example: *morya -> μοῖρα.

j. m -> n /_#
Word-final m becomes n; easily seen when you compare the accusative singular to Latin or Sanskrit.

k. s -> h / #_[+sonorant]
-> h / R_[+sonorant] where R is a non-nasal sonorant.
The famous change of s to h.

l. h -> Ø / #_[+sonorant], with the possible exception of /r/
-> Ø [+sonorant]_[+sonorant], including between vowels.
The subsequent deletion of h between vowels and in many other cases. There were some cases where the s was re-restored; for example, the s in the aorist of vowel-stem verbs, like ἐλύσα, was by analogy with consonant-stem verbs where the s was not deleted.

m. Ø -> a, e or o /#_[r l m n w]V
Basically, before a word initial liquid or nasal followed by a vowel, a vowel- a, e or o- is inserted. Which vowel appeared seems to have been unpredictable- there was a tendency for /e/ to occur _[r l]B, and /o/ to occur _[r l]F. This change did not occur in all cases. For example, PIE *nomṇt produced Latin nomen, but Greek ὄνομα.

n. t -> s /[+segment, not /s/]_H [where H means a high vowel]
-> s /#_w

/t/ became /s/ before a high vowel, and initially before *w: *tu -> σύ, *gʷṃtis -> βάσις.

o. Changes in the labiovelars. Basically, [kʷ gʷ kʷʰ] become:

-> [t d tʰ] / _
-> [p b pʰ] / _[a o], _N
-> [k g kʰ] / _u, all other _C
Examples: PIE *penkʷe -> πέντε, *gʷʰonos -> φόνος, *kʷukʷlos -> κύκλος. By now we have reached the historical period for Greek, since Mycenean Greek preserved the labiovelars.

p. Vns -> V:s / _#, _[+sonorant] (this remained as a rule, sometimes)
-> Vs /_C
V: /_V

Example: *tons -> τοῦς.

r. Some vowel contraction rules. Some remain in the grammar, morphologically conditioned.
- V[iu] remains as a diphthong.
- V₀V₀ -> V₀:.
Otherwise:

-If either the first or the second vowel is one of /o o: ɔ:/, then the result will be one of /o: ɔ:/.
-If the first vowel is /a(:)/ and the second is not /o o: ɔ:/, the result is /a:/.
-If the first vowel is /e ɛ:/ and the second is not /o o: ɔ:/, the result is /ɛ:/.

This is the cause of the notorious "contracted verbs" in Attic.

s. a: -> ɛ:/ everywhere except [e i r]_
This change happened in Attic and Ionic, but not Doric. Because the exceptions noted did not happen in Ionic, it is possible that this happened in two stages, with a: -> ɛ: unconditionally first, and then ɛ: back to a: after e, i and r in Attic.

t. Obstruent voicing assimilation: if two obstruents are in a cluster, the first will match the voicing of the second. Similar assimilation occured with aspiration: if the second obstruent was aspirated, the first became so, as well.

u. Various cluster changes:
tt ttʰ -> st stʰ
ts -> ss -> s
ksk -> sk
[p b pʰ]N -> mN
[k g kʰ]m -> ŋm
sn -> n / #_
ks -> s / #_ (not always)
mr nr -> br dr/ #_
ml -> bl
[dn]l -> ll

v. V:u -> Vu
V:i -> Vi / except finally
Diphthong shortening.

w. u(:) -> ü(:) /C_ or #_ (written <y>)
Fronting of /u/ except in diphthongs.

x. ei, ou -> e:, o: (later, -> i: u:)

y. Grassman's law:
Cʰ -> C /_(sonorant)Cʰ

z. w -> Ø / _C, C_, V_V
-> h or Ø / #_V
/w/ deletes, except at the beginning of a word, where it sometimes becomes /h/

α. Ø -> h /#_r, #_ü(:)

β. Vowel shortening. These changes both remain as phonologically conditioned rules; neither applied obligatorily.

V: -> V / _+V(:) where + is a word-internal morpheme boundary.

V:+V[+short] --> V[+short]+V:
A long vowel followed by a short vowel over a word-internal morpheme boundary swap lengths.

[hwhatting]

The sound changes here are standard late-stage PIE reflexes:
...
Syllabic [h₁ h₂ h₃] -> e a o

They have that change for Late PIE? It can only be Late stage to Greek, because most other IE subfamilies merge syllabic [h₁ h₂ h₃] into one vowel (/i/ in Indo-Aryan, /a/ or its reflexes in the other dialects), which speaks for them still being syllabic laryngeals in Late PIE.

ih uh -> ī ū

So they don't have laryngeal breaking (e.g. iH2 > ia)?

ph th kh -> pʰ tʰ kʰ

Interesting. I don't know many IEanists who still work with those for PIE.

In addition, the PIE palatovelars *ḱ ǵ ǵʰ became k g gʰ at this stage. The book I have does not include palatovelars (its authors believe they derive from clusters), but if you accept their existence, this is what happened.

I'd like to see how they argue for that, especially for Late PIE.