User talk:Rua/Archives/2013/February

Proto-Slavic tone
I don't totally understand this myself. Derksen 2008 (look on the net) says "*dômъ m. u (c)" with a long falling (circumflex) accent. Class c means that there's mobile accent, with an original (short) accent on the root syllable in this case. All original short accents were, or became, falling; short rising accent came later as a result of pan-Slavic shortening of original long-rising (acute) accent. The neoacute retraction generated new rising accents. Derksen seems to claim (based on his reconstructed forms) that the neoacute retraction did not lengthen short vowels, but this is denied both by Schenker (who seems to say, AFAIK, that the neoacute neoacute retraction lengthened newly stressed short vowels in "North Slavic") and Kortlandt, who says in "West Slavic Accentuation" that lengthening under the neoacute did happen (unclear whether he means in South Slavic, too). The key to understanding the long falling accent in *dômъ is apparently item 8.8 in Kortlandt's chronology:, which says "Lengthening of short falling vowels in monosyllables", where by "monosyllables" he seems to mean words that would be monosyllabic after loss of the yer.

So the order would be something like this:


 * 1) Proto-Slavic starts with short (falling), long rising (= acute), and long falling (= circumflex).
 * 2) Acute is shortened, producing short falling (= old short), short rising (= old acute) and long (falling = circumflex).
 * 3) Neoacute is retracted, leading to new instances of short and long rising (short rising only in South Slavic according to Schenker), producing short falling (= old short), short rising (= old acute or short neoacute), long falling (= circumflex), long rising (= long neoacute).
 * 4) Lengthening of short falling (= old short) syllables in "monosyllables", producing short falling (= old short not in a "monosyllable"), short rising (= old acute or short neoacute when it occurs), long falling (= circumflex or old short in a "monosyllable"), long rising (= long neoacute).
 * 5) Following this, many more changes.

At least, the first four steps seem to produce nearly all the reconstructed forms in Derksen 2008.

I still don't totally understand step 5 here. There's a "traditional" claim in Schenker that there was "compensatory lengthening" before yers in "many languages"; this may cover step 4 and other cases (Polish, Ukrainian?). Kortlandt has an additional step 10.6 lengthening the acute in Czech in multisyllabic words (not counting yers). He says a whole paragraph in "West Slavic accentuation" denying the claim that the long acute was preserved in Czech. Lengthening in Czech dům and bůh (Czech has length only in the nominative; Slovak has no length in either place) is mentioned a couple of times in Kortlandt; he cannot really explain this in any way but analogically. He denies that such length can be phonological (e.g. lengthening before a yer) because the "counterexamples are prohibitive". Polish has short dom but long bóg, which Kortlandt chalks up to "lengthening before devoiced obstruents", which seems questionable.

So in truth, West Slavic accentuation seems hella complicated and badly understood and full of weird special cases. This doesn't seem to apply to East Slavic and South Slavic so obviously. The Serbo-Croatian (including Chakavian) dôm, dȍma appears the reflect the expected outcome of steps 1-4, Slovenian probably too given the numerous Slovenian-specific cases where the accent was moved later and earlier.

Kortlandt (and Derksen following him) assert that the "glottalic tone" (a special register distinction corresponding to the original Balto-Slavic acute) was continued all the way almost to the end of Common Slavic, but I don't think this is generally accepted. Benwing (talk) 06:10, 1 February 2013 (UTC)

BTW
 * the general relengthening of the acute in East Slavic is given as step 10.4 in Kortlandt.
 * The order in Kortlandt is evidently rather unsettled, given that his "West Slavic accentuation" has a statement placing step 10.6 prior to step 9.4.
 * Shortening of circumflex in Czech and Slovak is Kortlandt step 9.4.
 * I don't know about the lengthening in Czech práh (gen. prahu) vs. Slovak prah, same type of pattern as dům and bůh. Kortlandt would probably claim that this is likewise analogical, with a pattern of long nom. vs. short gen. developing in various cases in Czech and getting extended (??).

Benwing (talk) 06:28, 1 February 2013 (UTC)
 * So as I understand it, the original form is indeed *dȍmŭ with short falling accent, which then later was lengthened to *dȏmŭ in step 4. I think steps 3 and 4 of your list could be seen as a larger whole in a sense, since they both involve (compensatory?) lengthening before a weak yer. Step 3 just has an additional clause: if the yer was accented, it retracts and forms a rising accent. I think as far as dating goes, step 4 must have occurred after the resolution of liquid diphthongs, because the "length" they produced in South Slavic and southern West Slavic turned short *o into long *a. Then lengthening of steps 3 and 4 didn't produce *a though, so it must have occurred after *o and *a had become more differentiated in quality. Thank you for your reply, I think I understand a bit better now. I haven't been able to find any Derksen 2008 online, but I would be very interested in an online etymology dictionary of Slavic, if that is what it is. CodeCat (talk) 15:13, 1 February 2013 (UTC)
 * Derksen 2008 is a book "Etymological Dictionary of the Slavic Inherited Lexicon"; see the reference in Proto-Slavic. I agree about 3 and 4 being similar types of compensatory lengthening, although the conditions are rather different -- step 4 evidently occurs only in monosyllables and everywhere in the Slavic area, whereas neoacute lengthening is more general in the words it can occur in, but more restricted in that it doesn't occur in Serbo-Croatian and possibly Slovenian. You are right that steps 2 to 4 all occurred after the metathesis of liquid diphthongs, at a time when the former long and short vowels were distinguished by quality. None of these steps turned e o ь ъ into ě a i u/y or vice-versa. Benwing (talk) 04:09, 2 February 2013 (UTC)
 * Ok thank you. I've found the book online in PDF format and it looks like it could be very useful. I wonder though, whether the reconstruction of Proto-Slavic given in that book is a bit misleading, considering that it ignores the metathesis while applying the accent shifts. It seems clear to me that such a language never actually existed, but that it is basically the unification of all features that are shared equally by all Slavic languages, regardless of their relative dating. So it is a diachronic reconstruction but not a synchronic one. I'm not sure if that is a problem as such, but I do think that if we follow that premise in the Proto-Slavic article, it needs to be made clear.
 * I'm still a bit confused about the outcomes of the various accents as well. Supposedly Slovene hasn't shortened any of the accents (that's what the article says), but it does have short accented vowels in (mostly) final syllables, so are they shortened old acutes or did they develop some other way? I think that the "Accent" section might need rewriting, maybe with a table to show the outcomes more clearly. CodeCat (talk) 11:48, 2 February 2013 (UTC)
 * Slovene is a special case. J.M.S. Priestley in "The Slavonic Languages" (ed. Comrie et al.) gives the following list of Slovenian-specific accent changes:


 * 1) "Long rising vowels became short (rising)." (If we are to believe this, it means that there was a general merger of acute and neoacute. However, possibly this really just means that the old acute was shortened, as happened pan-Slavically.  See #5.)
 * 2) "Short falling vowels became long (falling)." (This means that original short vowels merged with original and secondary circumflex.)
 * 3) "Stress shifted from long falling non-final syllables one syllable to the right, producing new long falling syllables." (This is the so-called "progressive shift" -- original short and circumflex vowels move to the right, remaining as long circumflex.)
 * 4) "Stress shifted from short final syllables one syllable to the left onto preceding long vowels, producing new long rising vowels." (This seems like another neoacute-type retraction. Note that final syllables due to progressive shift will never be retracted but they're long. Note also that all final syllables become short during Late CS, although in Kortlandt's theory certain of them later lengthen and trigger a neoacute retraction. So it's unclear exactly which stressed short final syllables are being referred to -- all those that were not due to the progressive shift, or perhaps in slightly more limited circumstances?)
 * 5) "Old neoacute and all short rising vowels in non-final syllables were lengthened". (See also #1. This is actually listed in the Proto-Slavic section on accent -- the acute is relengthened in initial syllables of multisyllabic words, although this step here is slightly more general. Note that #1 and #5 appear to reverse each other other than in final syllables.)
 * 6) "Short rising vowels in final syllables become short falling".
 * 7) (only in some areas, including standard Slovene) "stress shifted from short final syllables one syllable to the left onto preceding short e o, producing new long rising low-mid vowels"
 * 8) (only in some areas, not usually in standard Slovene but sometimes as optional forms) "stress shifted from short final syllables one syllable to the left onto preceding short ǝ, producing new stressed shwa"


 * These changes are quite confusing, esp. because they seem to reverse each other. I think the basic idea is that, in non-final syllables all stressed vowels end up long, with acute and neoacute merging as long rising. Furthermore, not mentioned here is that all non-stressed syllables are shortened. These two things lead to the actual Slovene prosodic system, where length and stress always co-occur except in words with final stress, which can have either short or long stressed syllables.
 * Note that this also implies that neoacutes are shortened in monosyllables, which I have a hard time believing.
 * The last two steps I've heard of before. Step 7 in particular leads to phonemic stressed long low-mid vowels, whereas the previous long mid vowels usually turn out as high-mid.
 * Step 4 is what's known as the "progressive shift".
 * An additional change not mentioned here at all is the "neo-circumflex", where original acute vowels change into long circumflex in non-final syllables in certain morphological circumstances (e.g. before an internal weak yer, before long vowels in the stem, etc.). See


 * I'll need to do a bit of looking at Derksen's forms at some point to make sense of this. Perhaps you can do the same and let me know what you figure out. Benwing (talk) 12:15, 4 February 2013 (UTC)

Slovenian tone
OK, I looked at this, and it's clearer now:


 * Acute maintained short in monosyllables, lengthened in non-final syllables:


 * *čàsŭ m. o (a) "time": Chak. čȁs (gen. čȁsa), Sln. čàs (gen. čása)
 * *kùpŭ m. o (a) "heap, mound": Chak. kȕp, Sln. kùp (gen. kúpa)
 * *kľùka f. ā (a) "hook": Scr. kljȕka, Sln. kljúka
 * *làpa f. ā (a) "paw": SCr. lȁpa "paw" vs. Sln. lápa "snout, mouth"
 * *kolě̀no n. o (a) "knee": Chak. kolȉno/kolȅno, Sln. kolẹ́nọ
 * *làskati v. "flatter": SCr. lȁskati (1sg. lȁskam) vs. Sln. láskati (1sg. láskam)


 * Neoacute maintained long in monosyllables:


 * *kľũčĭ m. jo (b) "key": Chak. kľũč, Sln. kljúč
 * *kǫ̃tŭ m. o (b) "corner": Chak. kút, Sln. kǫ́t
 * *devętŭ num. o "ninth": Chak. devẽtī, Sln. devę́ti


 * Neoacute maintained short in monosyllables:


 * *kõšĭ m. jo (b) "basket": Chak. kȍš, Sln. kòš


 * Neo-circumflex (acute -> circumflex in certain cases in non-final syllables, e.g. when a long vowel follows):


 * *làjati v. (a) "bark": SCr. lȁjati (1sg. lȁjēm) vs. Sln. lâjati (1sg. lâjam)
 * *làziti v. (a) "crawl, creep": SCr. lȁziti vs. Sln. láziti (1sg. lâzim)


 * Circumflex maintained in monosyllables:


 * *kâlŭ m. o (c) "dirt": SCr. kâl "dirt, mud, puddle", Sln. kâł "mud in a puddle, dregs, puddle"
 * *kôlsŭ m. o (c) "ear, spike": Chak. klâs, Sln. klâs


 * Original short vowel lengthened to circumflex in monosyllables:
 * Progressive shift from circumflex to the right:


 * *bôgŭ m. o (c) "god" (< *bȍgŭ): Chak. bôg (gen. bȍga), Sln. bộg (gen. bogâ)
 * *bôlgo n. o (c): Chak. blâgo "cattle", Sln. blagộ "good, goods, cattle"
 * *dêrvo n. o (c) "tree, wood": Chak. drîvo, Sln. drẹvộ
 * *kȍkošĭ f. i (c) "hen": Chak. kȍkōš, Sln. kokộš (short falling lengthened, then progressive shift to right)
 * *kȍlo n. s "wheel": SCr. kȍlo, Sln. kolộ (same as prev.)
 * *dȅsętĭ num. i (c) "ten": SCr. dȅsēt, Chak. dȅset, Sln. desę̂t


 * Leftward shift from short onto long syllables:


 * *kāzàti v. "show": Chak. kāzȁti (2sg. kãžeš) "say, tell, show", Sln. kázati (1sg. kážem -- neoacute remains lengthened)
 * *klęčàti v. (c) "kneel": Chak. kľečȁti (2sg. kľečĩš), Sln. klę́čati (1sg. klečím -- neoacute remains)
 * *kūrìti v. (b) "smoke": Chak. kūrȉt (1sg. kũrin), Sln. kúriti (1sg. kúrim -- neoacute remains)
 * *lě̄xà f. ā (b) "strip of land, bed": Chak. liehȁ, Sln. lẹ́ha
 * *lě̂nŭ adj. o (c) "lazy, slow": Chak. lîn (fem. līnȁ), Sln. lện (circumflex remains) (fem. lẹ́na)


 * Leftward shift from short onto short low-mid syllables, lengthening:


 * *čelò n. o (b) "forehead": Chak. čelȍ, Sln. čélọ

These all correspond to one of the 8 steps listed above in my previous message, except that step 1 is indeed the pan-Slavic shortening of the acute and step 8 apparently isn't standard. The Proto-Slavic forms are from Derksen, except that I've added the neoacute where I know it should go. (All cases of long rising accents in Derksen are neoacute, as are some short neoacute, e.g. masc. accent class b.) Note also that Derksen indicates the lengthening of the original short into circumflex in monosyllables, which he considers pan-Slavic.

The basic conclusion is that Slovenian agrees with Chakavian in monosyllables but lengthens all non-final syllables, and in the process might convert original short rising to falling (the neo-circumflex) and/or move the accent leftwards or rightwards. Note also that some Slovenian dialects preserve nasality (and occasionally add nasality where it wasn't originally), and this is indicated in Derksen's forms whenever available.

Benwing (talk) 01:09, 5 February 2013 (UTC)

Comments on dĭnĭ "day"

 * The proto-form dьnja (presumably meaning dьňa) is from Sussex/Cubberley, assuming that it was a jo-stem. Movement from n-stem to jo-stem was probably already happening in Late CS, although it seems fine to list it as an n-stem.
 * The forms with nasal vowels in Slovene are as given in Derksen, probably based on NW Slovenian dialects.

Benwing (talk) 04:38, 5 February 2013 (UTC)
 * Oh, I thought the Slovene words were just a mistake for underdots. As for "day", the final -e is still attested in OCS, so it seems strange to have a later form as Proto-Slavic. CodeCat (talk) 14:22, 5 February 2013 (UTC)

better ref
Try this:

http://www.scribd.com/doc/97679547/Verweij-Quantity-Patterns-of-Substantives-in-Czech-and-Slovak

You should be able to see it even without logging in. Benwing (talk) 05:05, 7 February 2013 (UTC)
 * I can see it, but the letters are very rough and hard to read, zooming in doesn't help... CodeCat (talk) 22:41, 7 February 2013 (UTC)


 * Parts of it are readable through Google Books, I thought ... but can't find the link now.


 * Here's more tables, these by Kortlandt (around p. 410), who postulates an extra acute (glottal) accent that you should take with a grain of salt ...

http://books.google.com/books?id=gYpB3wAYo-AC&pg=PA397&dq=proto-slavic+accent&hl=en&sa=X&ei=dsQUUYawLIL4yQGP9oDwDA&ved=0CFIQ6AEwBzge#v=onepage&q=proto-slavic%20accent&f=false


 * BTW you *can* make out the letters with a little guesswork, and the accents seem fairly clear ... there are only a few.

Benwing (talk) 09:34, 8 February 2013 (UTC)

E.g. masculine o stems, accent class c:

I'm filling in the neoacutes here based on the following:
 * short rising on a short syllable must be neoacute, because an original acute can only go on a long syllable
 * long rising on a long syllable must be neoacute, because acutes were shortened. The case of vozý is weird; no neoacute retraction possible and final syllables sometimes get lengthened or (usually) shortened in unexpected ways. Possibly this is meant to be a short rising??

BTW, the indication ẹ apparently indicates e2, i.e. former PSl *ai. Also, what looks like a dot above a letter can be assumed to be short rising. Two dots are short falling, even when they seem to merge, as in vȍzŭmĭ (this can't be #vôzŭmĭ because (a) compensatory lengthening like this occurs only in monosyllables; (b) the circumflex is usually clearly drawn as such).

Benwing (talk) 09:57, 8 February 2013 (UTC)

moving proto-slavic
The plan is to move the page and then copy the sections relevant to proto-slavic back into a new article. I thought you were OK with this. Because more of the page is about the language history overall than specific to Proto-Slavic, it makes sense to keep the change history of the stuff going into "History of the Slavic languages". Benwing (talk) 21:26, 7 February 2013 (UTC)
 * I didn't think it was that important to keep the history. If you think it's important to keep it then it's ok. CodeCat (talk) 22:25, 7 February 2013 (UTC)
 * I do think so. Benwing (talk) 10:00, 8 February 2013 (UTC)