The Correspondence Library

[alice]

Why is the Italian word for office ufficio? I didnt see anything in http://www.spinnoff.com/zbb/viewtopic.p ... 902#294902 about initial /o/ going to /u/, although there is a mention of a sporadic shift of /au/ to /u/. Is it a loanword from a dialect or is it an irregular?

Before labial consonants, unstressed /e/ and /o/ generally become /o/ and /u/ respectively.

[Whimemsz]

Hmmm...I notice all the romance SCs are from Vulgar Latin. Does anyone have the SCs from Classical Latin to Vulgar Latin?

As Rad said, CL didn't strictly evolve into VL. But there are correspondences between the two that can be expressed as sound change rules.

A few of the more significant include the VL loss of the CL accusative -m, a lot of vocabulary differences, and a reshaping of the vowel system:

a, a: → a
e → ɛ
i, e: → e
i: → i
o → ɔ
u, o: → o
u: → u

Those are the ones I know of off the top of my head.

EDIT: Oh, here (thanks Aszev and Dewrad!)

[Mecislau]

Proto-Semitic > Biblical Hebrew

Regular Consonant Correspondences:

p > p
b > b
θ > ʃ
θˁ > sˁ
ð > z
t > t
tˁ > tˁ
k > k
kˁ > q
g > g
ʔ > ʔ
s > s
sˁ > sˁ
z > z
ʃ > ʃ
ɬ > ʃ
ɬˁ > sˁ
l > l
r > r
x > ħ
ɣ > ʕ
ʕ > ʕ
ħ > ħ
h > h
m > m
n > n
w > w
j > j

/w/ and /j/ underwent some mergers and drops as well, but I'm not sure of the rules. /h/ also was frequently dropped. Some of these distinctions were probably still maintained in early Biblical Hebrew in speech, though not in writing, as is evidenced by how the letter ע was recorded in Greek sometimes as G (from /ɣ/) and sometimes as null (from /ʕ/).

Stressed vowel correspondences:
a: > a: (finally), o: (elsewhere)
i: > i:, i:ə (before h, ʕ, ħ), e: (finally)
u: > u:, u:ə (before h, ʕ, ħ)
o: > o:
a > a: (open syllables), a (less commonly), ɛ (less commonly)
i > e:, e:ə (before h, ʕ, ħ), e (in closed syllables in verbs), a (in closed syllables in verbs before h, ʕ, ħ)
u > o:, o:ə (before h, ʕ, ħ), o (in closed syllables in verbs), a (in closed syllables in verbs before h, ʕ, ħ)
aw > a:w
aj > aj, e: (in an open syllable), ɛ: (word-finally)

Unstressed vowel correspondences:
a: > a:
o: > u:, i: (in an open syllable before o:)
u: > u:, i: (in an open syllable before o:)
a > Ø (word-finally, in open syllables two or more syllables before the stress), ə (before or after h, ʕ, ħ, ʔ if an adjacent syllable has ə), a: (immediately before the stress), i (in closed syllables), Ø elsewhere
i > i, Ø (word-finally, in open syllables two or more syllables before the stress), ə (before or after h, ʕ, ħ, ʔ if an adjacent syllable has a frontal allophone of ə), a (before or after h, ʕ, ħ), e: (immediately before the stressed syllable)
u > ə, Ø (word-finally, in open syllables two or more syllables before the stress), ə (before or after h, ʕ, ħ, ʔ if an adjacent syllable has a backed allophone of ə), ɔ (closed syllables), u (before a geminated consonant), Ø (elsewhere)
aw > o:
ay > e:

ə could be realized as an ultrashort /a/, /e/, or /o/ depending on its surroundings

So basically: Every vowel can become virtually every other vowel.

Lenition:
p > f
b > v
t > θ
d > ð
k > x
g > ɣ

This refers to nongeminated non-initial consonants only

Miscellaneous changes:
ʕ > Ø (in coda position)
j > Ø (after ɛ, e, i)
at > a: (word finally, in feminine noun endings)

And just for interest, here's an example of the evolution of a segolate nouns (a type of broken plural in Hebrew)

*darku "path, way" > *darɛk (loss of case endings) > dɛrɛk > dɛrɛx
*darki:ma "paths, ways" > *dəraki:m (loss of case endings) > drakim > draxim
*darkiya "my path, my way" > darki: > darki


Biblical Hebrew > Modern Hebrew

Mind you, these aren't all true sound changes per se, since Modern Hebrew was artificially revived and is an amalgamation of dialects.

a(:) > a
e(:), ɛ > ɛ
i(:) > i
o(:) > o
u(:) > u
ə > Ø (see note), ɛ (elsewhere); realizations such as [a] become /a/, never dropping
w > v
θ > t
ð > d
x > χ
ɣ > g
ʕ > ʔ
sˁ > ts
ħ > χ
tˁ > t
q > k
h > Ø (word-finally)
ʔ > Ø (everywhere except in onset of stressed syllable - colloquial)
h > Ø (colloquial)

In addition, all geminate consonants simplified (making the former contrast of things such as /p/ [f], /p:/ [p] phonemic as a /p/~/f/ contrast). Also, the historical r (probably a trill [r]) became a uvular approximant.

Schwa note: The schwa was dropped everywhere where it would not be awkward to do so. In other words, virtually anywhere except between clusters such as initial nasal + C (mədab:er "I/you/he speaks" > mɛdabɛr), or when the schwa was already being realized as [a] or [e] (in the presence of gutteral consonants). Also, Hebrew does not allow clusters of three or more consonants, so if there were two schwas in a row, the first dropped and the second vocalized: ti:ləmədi: "you will learn (f)" > tilmedi.


Of course, there were lots of other less regular changes involved, but this hopefully covers most of the major ones.

[Tropylium]

[PROTO-URALIC TO FINNISH]

Undergoing revamp @ FrathWiki.

[Petusek]

Hello Everyone!

I'm looking for a list of Salish regular correspondences (e.g. from Kuipers' Salish Etymological Dictionary (2002), which I don't have ).

Similar lists from the other so-called "Almosan" language families will be welcome, too, especially Wakashan (North Wakashan, at least, another book I don't have is Fortescue's Comparative Wakashan Dictionary) and Algonquian.

Thank you very much for any help or advice!

Best wishes,

Petusek

[LoneWolf]

I request sound changes from Proto-Germanic to Old Norse. I was actually suprised when browsing the thread not find anything on this.

Anyway, I hope someone has them. I'm not getting my hopes up about it tho', but you just never know.

OH, and thanx to anyone that can help !!!

[Kalor]

Does anybody have sound chanes of Middle Dutch > Modern Dutch?

I hope somebody has them... Thanks!

[Izambri]

Has anyone got the Proto-Celtic to Breton sound changes?

[jmcd]

Has anyone got the Proto-Celtic to Breton sound changes?

You can find some stuff on it here and on wikipedia but I'm sure there's some changes missing

[Izambri]

Has anyone got the Proto-Celtic to Breton sound changes?

You can find some stuff on it here and on wikipedia but I'm sure there's some changes missing

Thanks jmcd. These ones are useful.

-------------

BTW, would be possible to restore the Latin > Catalan list? It has lots of question marks.

And although I almost know the answer I must ask if there exist lists for:

PIE > Phrygian
PIE > Thracian
PIE > Ancient Macedonian

A shame that Paleo-Balkan languages are poorly attested...

[Tiamat]

I would like to request Old norse > western old norse dialect and western old norse > Icelandic.

[Thomas Winwood]

If it's English sound changes you want, there's some here.

If anyone wants these in a listed / not roundabout form (some of those links reference the same change, which, at least, confused me at first), I've decided to list them out in regular notation:

[...]

Could anyone restore these sound changes to their original not-brokenness?

[jmcd]

I would like to find Old Korean to modern Korean sound changes. Are the digraph-like vowel symbols(e.g. a+i=ae, eo+i=e) enough evidence of past changes?

[PROTO-URALIC TO FINNISH]

Undergoing revamp @ FrathWiki.

If you don't mind, may I make a few comments?
You changed the symbol for the labial(dental) approximant from w to v before mentioning anything about any change.

As far as I can tell, there's no need to ever posit /T/ since it comes from /ts/ and goes to /ts/ with no changes in between. Unless you've forgotten to add those changes of course.

[Tropylium]

[PROTO-URALIC TO FINNISH]

Undergoing revamp @ FrathWiki.

If you don't mind, may I make a few comments?
You changed the symbol for the labial(dental) approximant from w to v before mentioning anything about any change.

Very difficult change to date.

As far as I can tell, there's no need to ever posit /T/ since it comes from /ts/ and goes to /ts/ with no changes in between. Unless you've forgotten to add those changes of course.

That's only if you look at Standard Finnish. Old Finnish dialects have a variety of outcomes. Granted, for Karelian and the †s outcome, there's indeed no need for a *T.

Anyway, here's also a bunch of Proto-Uralic to Proto-Mansi/Proto-Khanty changes I just typed for another purpose.

Vowels for Proto-Uralic to Proto-(Ob-)Ugric is not well understood; either there are a bunch of irregular splits, or there were 12 vowels to begin with and they were retained without changes. ("Tense" *i *e *æ *M *u *o; "lax" *I *E *Y *V *U *Q. Take that only as a phonological grouping.)

Proto-Ob-Ugric Proto-Mansi stressed vowel development can be described like this:
1) Mid illabial mergers: *e *i, *M *V
2) Height length, tenseness height:
*i *æ *æ *æ:
*u *o *a *a:
*I *E *i *i:
*Y *V *y *M:
*U *Q *u *u:

As for Khanty, things are much less regular - this is what I understand of it:
1) Front merger-chainshift: *e *Y *I, then *i *æ e
2) Back chainshift: *U O, then *u *U
3) *V *Q *a
4) Lax vowels:
*I *U *e_X *o_X (extra-short)
*E *O *æ *Q
5) Irregular splits:
*a *a *u (yes, not *Q *Q *U)
*æ *æ *i
6) Labialization of *e(_X) *ø(_X), adjacent to velars
7) Various umlaut phenomena

Consonant-wise most stuff goes on around PU POU:
*D *D_j *l *l_j (also in Permic, Hungarian, Samoyedic)
*s *S *K, then *s\ *s (also in Samoyedic)
*ts\ *tS
coronal + *k *k + coronal
*x *k *w *G / V_
*x 0 / _C
*N *Nk
non-homorganic consonant clusters generally broken up with *@
POU PMs includes *K *t, *tS *S
POU PKh includes *l K / V_

[jmcd]

As far as I can tell, there's no need to ever posit /T/ since it comes from /ts/ and goes to /ts/ with no changes in between. Unless you've forgotten to add those changes of course.

That's only if you look at Standard Finnish. Old Finnish dialects have a variety of outcomes. Granted, for Karelian and the †s outcome, there's indeed no need for a *T.

Yes but, unless the squiggle in the dialects shown on the map with yellow and brown is [T], I still don't see any reason to posit *T.

[Tropylium]

That squiggle is /T/ in the UPA, yes (either invariant /T:/ or gradational /T: ~ T/). Old Finnish, based on western dialects, quite regularly used <tz> to write this too. Note that the map is per 1940, I believe [T] is recorded from an wider area in older speech.

Also the /ht/ forms (incl. the variant with /h/ in gradation) aren't really achievable directly from /ts/, at least they also require intermediate /T/.

The smaller isolated instances of /s(:)/ are supposed to be also from /T/ by Swedish influence, but that would seem to work just as well starting from /ts/.

[jmcd]

Couldn't ht have come from t_h and then back to ts? Considering the T: area is so small, maybe it could've been only in that area that ts>T. That's seems more Ockamy to me.

[Qwynegold]

Ooh ooh, another suspectable cognate: Finnish hevonen or hepo vs. the ancient Celtic horse goddess Epona?

[schwhatever]

Standard Tamil to Colloquial Tamil

(All of this is according to Harold F. Schiffman's A Reference Grammar on Spoken Tamil)

First the syllable structure collapses a bit:

ai ee (exception: never finally in monosyllables, never initially in multisyllabic words)
avu, ayi aw ay (aU) and aI) as diphthongs)
iCa, uCa eCa oCa
k v delete /V_V
a: e: i: o: u: a E i o u /_#
am a:m a:n o~ a~: a~: /_#
an {~ /_#
om on o:m o:n O~ O~ o~: o~: /_#
en em e:n e:m E~ E~ e~: e~: /_#
um u~ /_#
m n n`:> m1 n1 n`1 /_#
l` deletes sporadically /_# (noted in plural forms, pronouns, and verbs, elsewhere much less common)
r\`:> deletes extremely sporadically, leading to compensatory lengthening /V_V (very incomplete)
l l` l:} l`:} when word finally, before a short vowel which is not proceeded by a diphthong or a long vowel
l l` l} l`} when word finally, if there is a long vowel or diphthong in the word
r 4 (in most dialects)
r\`:> l` (except in very careful speech, and even then very rarely is r\` produced intervocally)
r l l`:> delete /V_S (S=stop)
i u 1 } /non-initial syllable, unstressed, short
1 } delete unless word final
1 } reinserted in free variation to reduce clusters (ex: saappit`t`ukit`t`irukku sappt`t`kt`t`kk} sapt`}t`t`r}k})
nk [ng] Nk [Ng]
i i: e e: u u: o o: /m_R, v_R, p_R (here R = any retroflex
y (/j/) deletes /V_#, when the vowel isn't front
y y:i /V_#, when the vowel is front

Some more cluster simplification and palatization:
tt nt cc nc [c: Jc] /i_, y_
t`k kk

Then we lose the alveolar/dental distinction and reduce the dental/retroflex too:
n t: n_d t_d:
n` n (sporadic and dialect dependent)
l`(which already merged with r\` mainly) l l (again sporadic)

And the palatals start falling apart:
c s /_a, _o, _u, _e
c: tS: (most dialects)

Vowel Harmony!?:
o e u i /_Cu, _Ci (highly sporadic)

Sandhi:
y inserted between morphemes when the final vowel of the first is front
v inserted between morphemes when the final vowel of the first is back

There's also only a few changes necessary to turn this into the British dialect (which didn't merge retroflexes with alveolars):

i i: e e: u u: o o: /_l`
#eC@, #oC@ #Ce: #Co:

[The Unseen]

Standard Tamil to Colloquial Tamil

(All of this is according to Harold F. Schiffman's A Reference Grammar on Spoken Tamil)

First the syllable structure collapses a bit:

ai ee (exception: never finally in monosyllables, never initially in multisyllabic words)
avu, ayi aw ay (aU) and aI) as diphthongs)
iCa, uCa eCa oCa
k v delete /V_V
a: e: i: o: u: a E i o u /_#
am a:m a:n o~ a~: a~: /_#
an {~ /_#
om on o:m o:n O~ O~ o~: o~: /_#
en em e:n e:m E~ E~ e~: e~: /_#
um u~ /_#
m n n`:> m1 n1 n`1 /_#
l` deletes sporadically /_# (noted in plural forms, pronouns, and verbs, elsewhere much less common)
r\`:> deletes extremely sporadically, leading to compensatory lengthening /V_V (very incomplete)
l l` l:} l`:} when word finally, before a short vowel which is not proceeded by a diphthong or a long vowel
l l` l} l`} when word finally, if there is a long vowel or diphthong in the word
r 4 (in most dialects)
r\`:> l` (except in very careful speech, and even then very rarely is r\` produced intervocally)
r l l`:> delete /V_S (S=stop)
i u 1 } /non-initial syllable, unstressed, short
1 } delete unless word final
1 } reinserted in free variation to reduce clusters (ex: saappit`t`ukit`t`irukku sappt`t`kt`t`kk} sapt`}t`t`r}k})
nk [ng] Nk [Ng]
i i: e e: u u: o o: /m_R, v_R, p_R (here R = any retroflex
y (/j/) deletes /V_#, when the vowel isn't front
y y:i /V_#, when the vowel is front

Some more cluster simplification and palatization:
tt nt cc nc [c: Jc] /i_, y_
t`k kk

Then we lose the alveolar/dental distinction and reduce the dental/retroflex too:
n t: n_d t_d:
n` n (sporadic and dialect dependent)
l`(which already merged with r\` mainly) l l (again sporadic)

And the palatals start falling apart:
c s /_a, _o, _u, _e
c: tS: (most dialects)

Vowel Harmony!?:
o e u i /_Cu, _Ci (highly sporadic)

Sandhi:
y inserted between morphemes when the final vowel of the first is front
v inserted between morphemes when the final vowel of the first is back

There's also only a few changes necessary to turn this into the British dialect (which didn't merge retroflexes with alveolars):

i i: e e: u u: o o: /_l`
#eC@, #oC@ #Ce: #Co:

Over what time period did these changes occur? [random]Tamil has always struck me as a cool lang[/random]

[TzirTzi]

So having read through this thread a couple of times looking for Germanic sound changes (specifically ON>Modern Swedish, Middle Dutch>Modern Dutch, Old Frisian > Modern West Frisian) and failed to find anything I noticed lots of requests for sound changes of various Germanic languages, none of which had ever been responded to. So I thought I'd go and see what I could find in the uni library to address this need. However, after a frustrating 2 1/2 hours reading, I haven't been able to find anything of any real use - almost everything I could find was on syntax.

So now I'm just reiterating the various requests for Germanic sound changes. It's a notable lack in this thread!

And to make this post slightly more interesting, here's a weird sound change I found an article on:

North Frisian Lenition
Voiceless stops are voiced and spirantised following open syllables containing short vowels which are then lengthened.

[p t k] > > [v r ɣ] / [+syllabic-long]_[+syllabic], [+syllabic-long]_#
[+syllabic-long] > [+long] / _.

Why is this interesting? Because lenition didn't take place following etymological long vowels - which a priori would have sounded like a more likely environment. Also because - as far as I could tell from the examples - word final CVC syllables behaved like open syllables. If anyone's interested, here's the reference:

Goblirsch, Kurt Gustav (2002) ‘The North Frisian lenition and Danish linguistic hegemony’ in Raugh, Irmengard; Carr, Gerald F. (2002) New Insights in Germanic Linguistics III, Peter Lang Publishing:46-65

[Khvaragh]

Just a quick little nitpick:
This is not correct:

Proto-Semitic > Biblical Hebrew

Regular Consonant Correspondences:

ɬ > ʃ

This is the Arabic (and possibly Geʕez) sound change. Graphically, it is true through, since the lateral fricative was written with the same letter as that of the palato-alveolar fricative, shin, until the invention of dissimulating niqqud.
i.e. */?\aKrum/ > Ar. /?\aSrun/ "ten," Hebrew /?\esEr/
ɬ > s is the Biblical Hebrew change. There is an exception of "attraction" in a small percentage of words in Hebrew and other languages where this is true:
*/KamSum/ "sun" > /SamSu/, where the lateral fricative has become assimilated to the following palato-alveolar fricative. This explains other strange cognates which otherwise don't make sense, such as getting Ar. /Samsun/ from the same, because Arabic /S/ derives clearly from PS */K/ - so we would expect Caananite /samSu/ instead of /SamSu/. See McCarter Jr. - "Hebrew," (324-25 in Cambridge Encyclopedia of the World's Ancient Languages)

Note I am using the superscript sign for pharyngealization for the emphatic (ħuru:f al-mutˁbaqah) for consistency, however, some linguists believe that the original consonants were glottalized/ejective.

Arabic Sound Changes from Proto-Semitic:

p > f
b > b
θ > θ
θˁ > ðˁ
ð > ð
t > t
tˁ > tˁ
k > k
kˁ > q
g > gʲ > dʒ
ʔ > ʔ
s > s *1
sˁ > sˁ
z > z
ʃ > s
ɬ > ʃ
ɬˁ >dɬˁ>dˁ
l > l
r > r
x > x
ɣ > ɣ
ʕ > ʕ
ħ > ħ
h > h
m > m *2
n > n
w > w
j > j

*1 /s/ in some initial contexts change to /h/ and /?/, but not within the root itself, but only anaphora and certain derivational prefixes. For example, */su?a/ > Ar. /huwa/, */safta?\ala/ > /hafta?\ala/>/?ifta?\ala/. This is common to many other Semitic languages as well.
*2 /m/ changes to /n/ in certain contexts, notably in the nunation.
The vowels generally remain unchanged, but there is assimilation in some of the longer vowels. In particular, some of the examples of /a:/ come from contracted VjV/VwV sequences, for example: /kˁawama/>/qa:ma/ "he stood."; this explains the so-called irregular formation of this verb, in the imperfective, it is /jaqu:mu/ "he stands," which was probably originally /jakˁwamu~jakˁwumu/.

[hwhatting]

*1 /s/ in some initial contexts change to /h/ and /?/, but not within the root itself, but only anaphora and certain derivational prefixes. For example, */su?a/ > Ar. /huwa/, */safta?\ala/ > /hafta?\ala/>/?ifta?\ala/. This is common to many other Semitic languages as well.

Just a question from an ignorant bystander - is it clear that this is a sound change? Or could that be a case of different sets of anaphora, some with *h- and some with *s-, and then these paradigms become mixed up and the different anlaut consonants spreading to parts of the paradigm / anaphora set where they're not etymological? (And, due to hypercorrection, from there to prefixes as well?)

[dhok]

What about Proto-Uralic to Hungarian?

[Tropylium]

What about Proto-Uralic to Hungarian?

Vx Vː / _C
x ɣ / _V
ð l
l ∅ / CʲV_m (odd, but examples are numerous)
s, ʃ θ (probably phonetically [ɬ])
(t)sʲ s
(It's debated if *ʃ *tsʲ actually existed in PU at all)
ŋ ŋg
i y u ɪ ʏ ʊ
iː uː i u (there was no yː)
Vː V
ɤ ɯ / _i (sometimes elsewhere too)
ɤ a / _a
ɯ i
θ ∅ / #_, _C
kθ ʃ
w ∅ / C_
w v
m, ŋ n / _t
p t k b d g β ð ɣ v z v / V_V
mp nt ns ŋk mb nd nɟ ŋg b d ɟ g
k ∅ / {r l}_
k h / _[back vowel, ʃ]
p f / _V
ðʲ, lj ɟ (*ðʲ was probably phonetically [lʲ])
j ɟ (sporadic)
m v / V_V, l_V (sporadic)
tʃ ∫ (sporadic)
ɪ ʏ / w_[coronal]
Cː C
nl nʲː
(C)VCV (C)VːC (with some CVC)
e(ː) æ(ː) ɛ(ː)
a o ɒ (short only)
ɪ ʏ ʊ e o ø (with some exceptions)
ev(V) ɛv(V) av(V) øː øː oː
in most dialects, e ɛ
tj c

Just a quick & dirty basic version. I'm probably missing quite a few of the more specific rules, and the chronology may be out of order too. (Those that are common Ugric are in blue; degemination must be later than the Hungarian-specific medial voicing, however.)

/ʒ/ and initial /p b d c g z/ only occur in loanwords (from Turkic, Slavic or whatever).

[edit]Minor updates 2010-01-24