List of SCs from PIE to Latin

[Zju]

I've been searching wikipedia and the internet with no success, so I thought I'll ask here. Do you know a list of all known sound changes from PIE to classical Latin? Preferably in standard linguistic notation and not going into a few pages of explanations about every sound change, but still any exhaustive list will do.

[alynnidalar]

It's not exhaustive, but there's always the Index Diachronica.

[Terra]

Leiden's Italic Etymological Dictionary lists the sound changes from PIE to PI, so here they are, copy and pasted, and with some editing, because copying doesn't correctly capture long vowels, syllabic consonants, 1's and i's, and other random things :

The first number of each section indicates the order in the relative chronology of sound changes, whereas the second digit stakes no such claim.

A. Before Proto-Italo-Celtic split up:
1.1 PIE *h1e > *e, *h2e > *a, *h3e > *o
1.2 PIE *eh1 > *e:, PIE *eh2 > *a:, PIE *eh3,*oH > *o: before a consonant
1.3 PIE *iH, *uH > i:, u: before a consonant
1.4 PIE *Tt> *t^ht (and *-dd^h- > *-d^hd-?)

2.1 Pretonic shortening of long vowel before resonant (Lat. vir, ferus)
2.2 *uHijV> *wiHjV(pius-rule)
2.3 *CHC> *CaC
2.4 *CRHC> *CRa:C
2.5 *CRHTC > *CRaTC (Italic), *CRHT/s > *CRaT/s (Celtic)
2.6 *RHT/s- > *RaT/s- (Italo-Celtic)
2.7 *R=DC > *RaDC (Lat. magnus)
2.8 *CCCC> *CaCCC

B. Before Proto-Italic split up:
3.1 *t^ht>*ss
3.2 *CLHV> CaLV- (e.g. cale:re, vale:re, palma)
3.3 *CNHV> [C@NV-] (e.g. similis, sine, tenuis)
3.4 *HLC- > aLC- (argentum)
3.5 *h1/2/3NC- > e/a/oNC- (umbili:cus, ambi)
3.6 *p_(R)k^w- > *k^w_(R)k^w- (coquo:, qui:nque)

4.1 *b^h, *d^h, *g^h, *g^w^h > *b, *d, *g, *g^w after *s and *N
4.2 *g^h > *g before *l, *r (Lat. glaber, gra:men, tra:gula)
4.3 *b^h, *d^h, *g^h, *g^w^h > *f, *þ, *X, *X^w
4.4 PIE *ouV> *awV in pretonic position (Thurneysen-Havet-Vine's law) (Lat. lavo: etc.)

5.1 *þ- > *f- (Fal. filea, Lat. filius)

6.1 * fw- > *f- (Lat. forum, fores)
6.2 *-g^w-, -X^w- > *-w- /V_V (Lat, voveo, nudus)
6.3 *mj > *nj

7.1 *o > a /b, I, m, w, k^w _CV (Lat. badius, canem, lacus, lanius, manus, mare)

8.1 *ew > *ow (Lat. novus, moveo)
8.2 *r=, *l= > *or, *ol (Lat. morior etc.)

The phonological developments between Proto-Italic and Old Latin are too numerous to discuss here. For details, I refer to the main handbooks in the field: Sommer 1914, Leumann 1977, Sommer / Pfister 1977, Schrijver 1991, Sihler 1995, Meiser 1998, Baldi 1999, Weiss fthc.a, to name only the best-known English and German works. Also, the relative chronology of the post-Pit. sound changes has not yet been clarified in all details, and would justify a separate study. Hopefully, the present dictionary may contribute to that aim.

[Salmoneus]

I was going to say what Terra said.

But instead i'll check Matasovic on Proto-Celtic, and mention that he doesn't believe in Italo-Celtic. He comments also that CRHC > CRa:C is a two-step process in Celtic (so it might be in Italic too), and he says that he thinks Dybo's Law (loss of laryngeals before resonants, or as Terra's list puts it shortening of /i/ and /u/ before resonants) occurs only partially in Italic. He also puts TT>ss very early (I guess where Terra's list has Tt>tht, Matasovic goes directly for Tt>ss).

[Tropylium]

*/TT/ > *[TsT] is reconstructible already for PIE.

[Salmoneus]

Interesting - everybody else seems to disagree! Could you perhaps give some reasoning for this certainty, rather than having it just be your word against theirs?

Matasovic, for instance, has the change as *TT>*ss directly, locates it as 'dialectal' PIE, and specifically notes the same change also occuring in Italic and Germanic - the implication being it doesn't occur elsewhere. Similarly, as noted above, de Vaan believes the change occured in two steps, both post-PIE: TT>tht in proto-Italo-Celtic, and then tht>ss in Italic (and presumably independently in Celtic).

[Pabappa]

I think the idea is that */TT/ > *[TsT] is not the same change as */tt/ > /ss/, but might have helped it along. I think there is evidence of the first shift in Sanskrit or something, but Sanskrit deleted fricatives between two stops, so I'm not sure. Wikipedi says "An epenthetic /s/ was inserted already in PIE after dental consonants when followed by a suffix beginning with a dental." as if to imply that it's only for root + suffix situations.

[Valdeut]

Interesting - everybody else seems to disagree! Could you perhaps give some reasoning for this certainty, rather than having it just be your word against theirs?

Matasovic, for instance, has the change as *TT>*ss directly, locates it as 'dialectal' PIE, and specifically notes the same change also occuring in Italic and Germanic - the implication being it doesn't occur elsewhere. Similarly, as noted above, de Vaan believes the change occured in two steps, both post-PIE: TT>tht in proto-Italo-Celtic, and then tht>ss in Italic (and presumably independently in Celtic).

Actually, my impression is that most writers agree that "The Double Dental Rule" goes back to PIE. I'm not sure that at least what Matasovic says contradicts this if you take *TT to have phonetically been [TsT]. Only Italic, Celtic and Germanic changed it to *ss.

See this paper by Byrd:
https://www.academia.edu/4165302/The_Ph ... nal_draft_
(You may have to download the PDF)

"(Ι) */VTTV/ → *VTsTV (The Double Dental Rule)
While the previous geminate sequences were reduced to singletons, a geminate dental sequence was fixed by *-s- epenthesis. Simplified in most of IE (*/u̯i̯d-tó/ → *u̯itstó- ‘known’ > Germ. ge-wiss, Lat. vīsus, Gk. ἄ-ϊστος, Ved. vittá-), *-TsT- was maintained in Anatolian (*/h1ḗd-ti/ → h1ḗtsti ‘eats’ > Hitt. ēzzazzi [ēt͡st͡si]). However, if a geminate dental sequence was followed by a sonorant + vowel, a dental was deleted with no CL."

And to pick another example, here is what Beekes has to say on the subject:
"When two dental stops occurred one after the other (which could happen at a morpheme boundary), a sibilant was inserted between them in PIE. Thus, *-tt-, *-dt- yielded *-tst- and *-d^(h)d^(h)- yielded *-d^(h)zd^(h)-. The cluster is retained as such in Hittite; in other languages it is simplified to -tt- (Indo-Aryan), -st- (Iranian, Greek, Balto-Slavic) or -ss- (Italo-Celtic, Germanic)."

The different outcomes in different groups suggests that there was something going on already in PIE (also, Anatolian actually retains *TsT). Also, PIE in general didn't allow geminates, but in other cases they were just simplified.

[Pole, the]

The different outcomes in different groups suggests that there was something going on already in PIE (also, Anatolian actually retains *TsT). Also, PIE in general didn't allow geminates, but in other cases they were just simplified.

What about *TK → *TsK → *KTs? Is this shift widely accepted as having occured in PIE?

[Valdeut]

What about *TK → *TsK → *KTs? Is this shift widely accepted as having occured in PIE?

I think the last step is now widely accepted to not have occured already in PIE, since both Anatolian and Tocharian preserve the (presumably original) dental–velar order.

As for *TK → *TsK, I don't think it's widely accepted but as you can see in the paper I linked to, Byrd does thinks PIE had the rule (though only for tautosyllabic *TK). I believe it may have been Melchert who first suggested this, based on evidence from Luwian. But Kloekhorst and others argue against this rule both for PIE and for Luwian itself (and I'm not sure if any other language gives any evidence for this development):
http://www.kloekhorst.nl/KloekhorstPIEThornClusters.pdf (page 16)

[Tropylium]

What Valdeut said, and there's that /TT/ > /ss/ is implausible as a single-step sound change anyway. I could buy /TT/ > /st/ (simple coda spirantization), but since regular *st does not become /ss/ in Italic/Celtic/Germanic, this was clearly not the pathway. I doubt Matasovich actually claims it to have been a single-step change either.

I'm a bit mystified by what de Vaan is going on about "tʰt". I've heard of no other place with a sound change from an aspirate /tʰ/ or a cluster /tH/ to /s/ in Italic or Celtic, so this presumed intermediate stage seems to be a pointless complication. Is this, perhaps, something to do with an obsolete assumption that fricativization of voiceless stops always progresses through aspirated stops?

[Sleinad Flar]

That is just an erratum on Terra's part; it should be:

1.4 PIE *Tt > *t^st (and *-dd^h- > *-d^zd-?)

(At least in my copy of De Vaan).

And yes, it is common consensus that this sound change already took place in PIE itself.

As for the OP, Michael Weiss has written a book called "Outline of the Historical and Comparative Grammar of Latin". A while ago the chapters of this book were available online, but Mr. Weiss seems to have taken them offline when he released the book. There might still be some copies floating around though.

And if everything else fails, there's always Andrew Sihler - Comparative Grammar of Greek and Latin, which has copies floating around, if you're up to it.

[Sleinad Flar]

.... or you could just go to the Correspondence Library in the L&L Museum on this very site

{Ctrl+F "PIE > Latin".}