Talk:Proto-Slavic language
![]() | This article is rated C-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
History of this page before Feb 15, 2013
See Talk:History of the Slavic languages.
Table of all possible consonant clusters?
I think a table of all possible consonant clusters (which implicitly are the only ones allowed at the beginning of a word/syllable) would be useful. Simplification of "disallowed" combinations plays an important role in verb inflection, since it often caused the loss of a stem-final consonant. And that, in turn, is useful when discussing the grammar. There is already a table in the "alternation" section, which could be moved and expanded so that it simply shows every pair of consonants, and the result (which is simply iotation when the latter is *j). CodeCat (talk) 03:52, 16 February 2013 (UTC)
- There's a table in "The Slavonic Languages" p. 128 -- this is actually for OCS but surely the Common Slavic table was essentially the same. It's a bit complex exactly which clusters occur (which may not be the same as which ones are potentially allowed). But basically it goes something like this:
- sp, st, sk, sm, sn, sv, sl, sr; sx and sc only at morpheme boundaries. No sň, sľ or sř.
- zm, zn, zv, zl, zr; zb only at morpheme boundaries; zd and zg either word-medial or at morpheme boundaries.
- gn, gv, gl, gr; [d,t,k,x,sk] + [v,l,r].
- [p,b] + [l,r,ĺ].
- gd, gň, dm in a few words word-medially or at morpheme boundaries.
- pn, bn in a few words at morpheme boundaries.
- žl, žr, žň; žd word-medial or at morpheme boundaries.
- [m,v] + [l,r]; [m,v] + ĺ at morpheme boundaries.
- čl, čr, št, šl, cv, dzv, nr, str, smr, svr.
- stv word-medial or at morpheme boundaries.
- Also at morpheme boundaries: šň, šľ, spl, spr, stl, [sx/zg] + [v/l/r], [zb/zm/zv] + [l/r], zd + [v/r], štv, štř, ždř, tvr.
- Hopefully I didn't miss anything. For Common Slavic this is maybe something like [optional s or z] + (either: [optional n or m]; or: [optional stop or x] + [optional v]) + [l, r, j], but also disallowing certain homorganic combinations like nl, pv, bv; also allowing a few other possibilities, e.g. gn, šl, a few other possibilities only across morpheme boundaries, and some "allowed" combinations occurring only at morpheme boundaries or sometimes also word-medially.
- Note also that the existence of pn, bn in a few words didn't prevent these clusters from getting eliminated early on, suggesting that these are later and rare developments. Benwing (talk) 04:37, 18 February 2013 (UTC)
- Lunt's OCS grammar also contains such a table, which I was considering using to make a table for Proto-Slavic. There are definitely some differences, though. OCS has metathesis of liquid diphthongs, so that adds additional clusters that probably didn't occur in Proto-Slavic. I know that at least vr is one of those, because Slavic had undergone vr > r earlier on. sň and such are probably not allowed because of syllabic synharmony, so what appears in their place? snj > šň? CodeCat (talk) 04:44, 18 February 2013 (UTC)
- Indeed, sň isn't allowed because snj > šň. As for Common Slavic, you're right about some clusters with l and r. Surely this applies to žl, žr, šl, šr. OTOH ml is a real Proto-Slavic cluster (Derksen *mlinŭ "pancake".) Probably *mr too. Don't know about *vl. Can you quote an example where *vr > *r? Also, clusters št, žd are of course < *ty, *dy. Presumably štv < *stv(e) or *tvj?. Also šč < *sk(e) missing from above table. Example of dm is *sèdmŭ "seventh". Note also that bř, bľ, pľ occur at the beginning of Common Slavic words < PIE breu-, bleu-, pleu-. No reason all other combinations with ř and ľ can't appear wherever r and l can appear. Benwing (talk) 11:24, 18 February 2013 (UTC)
- stve would not have given štve (because se doesn't become še either), nor would tvj because that probably would have resulted in tvlj in OCS. Actually, tvj would have just become tъj because of how syllabification works in PIE, so it would have been restricted to more recent environments. So there is only one other source for ťv... I think that it may derive from gtv/ktv. I've created a table here with some of two-consonant pairs and their outcomes. A dash - indicates that there are no examples of the combination. Please add more to the table if you're able. CodeCat (talk) 14:41, 18 February 2013 (UTC)
- Lunt's OCS grammar also contains such a table, which I was considering using to make a table for Proto-Slavic. There are definitely some differences, though. OCS has metathesis of liquid diphthongs, so that adds additional clusters that probably didn't occur in Proto-Slavic. I know that at least vr is one of those, because Slavic had undergone vr > r earlier on. sň and such are probably not allowed because of syllabic synharmony, so what appears in their place? snj > šň? CodeCat (talk) 04:44, 18 February 2013 (UTC)
p | b | t | d | k | g | c | dz | č | s | z | š | ž | ś | x | m | n | l | r | v | j | |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
p | t | d | k | g | c | dz | č | s | z | š | ž | ś | x | m | n | pl | pr | pj | |||
b | t | d | k | g | c | dz | č | s | z | š | ž | ś | x | m | n | bl | br | bj | |||
t | st | tl | tr | tv | ť | ||||||||||||||||
d | st | dl | dr | dv | ď | ||||||||||||||||
k | ť | km | kn | kl | kr | kv | č | ||||||||||||||
g | ť | gm | gn | gl | gr | gv | ž | ||||||||||||||
c | - | cv | č | ||||||||||||||||||
dz | - | dzv | ž | ||||||||||||||||||
č | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | č |
s | sp | - | st | - | sk | - | sc | - | šč | s | - | - | - | - | s | sm | sn | sl | sv | š | |
z | - | zb | st | zd | - | zg | - | zdz | zm | zn | zl | zdr | zv | ž | |||||||
š | - | š | |||||||||||||||||||
ž | - | ž | |||||||||||||||||||
ś | - | - | |||||||||||||||||||
x | xv | š | |||||||||||||||||||
m | nt | ml | mr | mj | |||||||||||||||||
n | nt | ň | |||||||||||||||||||
l | lt | ľ | |||||||||||||||||||
r | rt | ř | |||||||||||||||||||
v | dt | dj | |||||||||||||||||||
j | dt | - |
- It isn't always so simple. Depends on whether you're talking about clusters inherited from Balto-Slavic or created during Slavic times, e.g. *zębnǫti vs. *gynǫti < *gūbn-. Cf. also *sedmŭ < PIE *septm- vs. věmi < *vēdmi, *ěmi < *ēdmi, *ěstĭ < *ēd-ti, ěsnŭ "clear" < Balto-Slavic *aiskn- (Derksen), *borda "beard" (Lith. barzda) but *borzda "furrow", *bręknǫti "swell", *bręždžati "jingle" < *brenzgj-, *brǔzda "bit" < PIE *bhrus-dh- showing that ruki does not apply to -zd- (likewise mĭzda "payment" < PIE *misdh-, *droždža "yeast" supposed < PIE *dhraghy-, *dŭzdjĭ "rain", *mŭldni "lightning" < PIE *mldh-n-. Etc. Look through Derksen. Benwing (talk) 18:56, 18 February 2013 (UTC)
- Ok, I see. I think a synchronic table would be better then. CodeCat (talk) 19:16, 18 February 2013 (UTC)
- Actually, you can't tell if ruki applied because the output *žd would merge back into zd anyway. As for štv-, I don't know if kt always produces ť or only before front vowels, e.g. in *dŭkti gen. *dŭkter- "daughter", *lękti "bend". Note *brutǔ "nail" < *brouk-to- (Lith. braũktas), pętŭ "fifth" < *penkwtos (Lith. peñktas). Benwing (talk) 19:21, 18 February 2013 (UTC)
- Are you saying that tvj might give ťv? That seems unlikely to me... are there any sources that go into more details about the outcomes of the palatalizations and iotation? CodeCat (talk) 19:30, 18 February 2013 (UTC)
- I don't really know. I'm sure there are, somewhere. The real thing is to figure out which word in OCS has -štv- in it and what its etymology is. Benwing (talk) 20:54, 18 February 2013 (UTC)
- Are you saying that tvj might give ťv? That seems unlikely to me... are there any sources that go into more details about the outcomes of the palatalizations and iotation? CodeCat (talk) 19:30, 18 February 2013 (UTC)
- Actually, you can't tell if ruki applied because the output *žd would merge back into zd anyway. As for štv-, I don't know if kt always produces ť or only before front vowels, e.g. in *dŭkti gen. *dŭkter- "daughter", *lękti "bend". Note *brutǔ "nail" < *brouk-to- (Lith. braũktas), pętŭ "fifth" < *penkwtos (Lith. peñktas). Benwing (talk) 19:21, 18 February 2013 (UTC)
- Ok, I see. I think a synchronic table would be better then. CodeCat (talk) 19:16, 18 February 2013 (UTC)
CE or AD?
I'm not sure why it was necessary to switch the BC/AD references to BCE/CE. I don't care all that much but I'm more familiar seeing BC/AD, and it also stands out more: I have to pay close attention to the BCE vs. CE notation because the two are so similar. I've heard some people object to the traditional notation, perhaps because of the embedded religious references (although they're still present behind BCE/CE as well, and even in the word "goodbye"). BTW I'm neither religious nor a Christian. Wikipedia doesn't have a policy prescribing one vs. the other (it just says don't unilaterally change notation, similar to not unilaterally switching spelling or other conventions). Benwing (talk) 03:56, 18 February 2013 (UTC)
- Yes, I'm sorry, I didn't see the part about unilaterally changing it. For me, AD/BC has religious connotations that don't seem very fitting in an article like this, I prefer BCE/CE because it's more neutral. I will change it back if you prefer though. CodeCat (talk) 04:20, 18 February 2013 (UTC)
- Do what you'd rather. I don't really see BCE/CE as more neutral -- after all, it's still the same calendar based off of the same supposed date for Christ's birth viewed as the coming of a new era. As an example, if for historical reasons we ended up using an "AH" calendar like the Muslims do, dated from 622 AD (Mohammed's flight from Mecca), then renaming the AH to something else wouldn't likely satisfy Christian etc. complaints about religious significance in the dating. But in any case it's not really important to me. Benwing (talk)
11:39, 18 February 2013 (UTC)
- IMHO, switching to BCE/CE is more appropriate in historical spheres unconnected to Christianity (such as Ancient China). I don't think that can really be said of Slavic peoples. — Ƶ§œš¹ [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ] 13:39, 18 February 2013 (UTC)
- It is definitely more "neutral" as being a step further away from strict reference to one particular religion. People can use it without having to break their own religious rules, which is why it is now the standard in history and anthropology. I say that if scholars are using it, this article should use it. It's not about what each of us individually think "matters." Using Latin/Catholic church abbreviations no longer makes sense as the international language of science is no longer Latin. In future, surely, some other system will be used (for dates way before 2500BCE. most anthropologists use BP (before present). But the problem with that is that every 25-50 years, one would have to change the date again, which is why most prehistorians and historians are using BCE. We are definitely in a different era than antiquity and using the old religious-based system of reckoning it is the best solution right now. I am a prehistorian. QueenofRods (talk) 18:02, 10 June 2021 (UTC)
- Technically, the Proto-Slavic speakers weren't Christians... yet. CodeCat (talk) 14:03, 18 February 2013 (UTC)
- Well, no one was in the BC era. — Ƶ§œš¹ [ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ] 14:33, 18 February 2013 (UTC)
- Technically, the Proto-Slavic speakers weren't Christians... yet. CodeCat (talk) 14:03, 18 February 2013 (UTC)
Common Slavic vocalism
František Václav Mareš and Georg Holzer have argued, from the evidence of the Nebenüberlieferung (secondary transmission, i. e., loanwords from and into Slavic, names and glosses, that is, only individual words, as opposed to connected texts in Slavic, which we don't have for the period in question), that the Slavic vocalism was still essentially identical to the Balto-Slavic one until the end of the 8th/beginning of the 9th century AD (i. e., about 800 AD), giving not only Proto-Slavic (which Holzer dates to the end of the 6th/beginning of the 7th century AD, i. e., about 600 AD), but even Common Slavic of the 7th and 8th centuries, which was already in the process of developping regional differences, a distinctively "Baltic" look (as can be seen from the reconstructions in Proto-Slavic borrowings; see also Slavic liquid metathesis and pleophony). That is, *u and *ū were still back rounded vowels, *a was still unrounded, long and short vowels were still essentially identical in quality, liquid metathesis and pleophony hadn't yet happened (vowel-initial metathesis under the acute, the so-called first liquid metathesis may partly date somewhat earlier, but still to the period after Proto-Slavic), there were no nasal vowels yet, and the second and third palatalisations hadn't occurred yet, either (at least in the Proto-Slavic stage). Many characteristic Slavic changes (*a > o, delabialisation of long and short u, centralisation of the short high vowels, second liquid metathesis, development of the nasal vowels) happened only after the late 8th century. See, for example, Holzer's EEO article on Proto-Slavic and his Historische Grammatik des Kroatischen (both in German). I'm not aware that any historical linguist disagrees with these conclusions; the main reason that the traditional, OCS-like notation of Proto-Slavic persists and the new, Baltic-looking reconstruction is only rarely used seems to be academic inertia and not much more, apart perhaps from the fact that the "new Proto-Slavic" looks so different from any attested Slavic language (attested in connected texts, that is). Moreover, both notations can be converted to each other relatively easily, which has probably also impeded adoption. (For scholars who resist Balto-Slavic, the "Baltic" look may also be a factor to reject "new Proto-Slavic", but Balto-Slavic seems to be generally accepted now.) But the facts are clear and should be properly explained in this article. --Florian Blaschke (talk) 19:44, 21 August 2013 (UTC)
- Since the discovery of Old Novgorod birch bark letters we know that the old reconstruction of Proto-Slavic is wrong and obsolete.. To account for its evidence, for example Słownik prasłowiański reconstructs both *cěvь and *kěvь for ONv. кевь which lacks second palatalization (in Holzer's notation that would be *kēwi). Croatian linguists generally accept Holzer's theory (there is the Historical grammar of Croatian written by Ranko Matasović which uses it exclusively), and of others Thomas Olander who is in fact writing a book on the reconstruction of inflection of such Proto-Slavic "proper". However, given the inertia prevalent in academia (e.g. Russians still massively use Pokorny-like PIE reconstructions and not the laryngeal framework), I doubt this would be widely accepted anytime soon. --Ivan Štambuk (talk) 01:09, 28 September 2013 (UTC)
- Holzer's theory makes a lot of sense to me and I think User:Florian Blaschke is right about a significant issue being how different the new reconstructions look vs. the old ones. However there's also the issue crystallized in this quote from Lindstedt, quote in Holzer: "due to a simplistic ... model whereby languages, once separated from a common ancestor, can share no further innovations whatsoever". The assumption of definitive splits in a protolanguage is still the norm and there isn't really a good theoretical apparatus for handling the case where innovations continue to propagate across a dialectally differentiated area (a "Common X" period corresponding to Common Slavic). This works well in many cases either because the "Common X" period was fairly stable or because the dialectal differences during this period were eventually erased in all extant dialects prior to anything being recorded. But it's particularly problematic in Slavic where the "Common Slavic" period involved enormous changes and where we have direct documentation from within this period. Cf. similar issues in PIE w.r.t. the laryngeals (leading to vague "Post-PIE" reconstructions). An even more dialect parallel is Proto-Germanic where it's likely that reconstructed Proto-Germanic *ō was actually pronounced *ā at this period (there are many indications of this; one is the Gothic word "Rūmōn-" "Roman" from Latin "Rōmān-", which is inexplicable if *ō was /o:/ but easily explainable if it was /a:/). Benwing (talk) 00:40, 28 October 2013 (UTC)
- Once the Hittite was discovered with actual reflexes of laryngeals old Brugmannian forms immediately became obsolete (now called "post-PIE"). Similarly, in Old Novgorod birch bark letters we have actually attested вьхe and кѣле, this one was discovered in 1956 and I can't find any other reason for sticking to the traditional (OCS-like) forms other than inertia. The traditional Stammbamum model of protolanguages is inapplicable when there is a dialect continuum across large area because isoglosses do not spread uniformly and completely, and that's the reason why Holzer has been advocating such an early stage for Proto-Slavic "proper". Of course, if notation is just convention and what is meant is not what is written, and each model can more or less easily be transferred into its more "correct" equivalent, than it's not really an issue. --Ivan Štambuk (talk) 01:49, 28 October 2013 (UTC)
- It may not be an issue to experts, but for laymen it is terribly confusing. Therefore the problem needs some kind of explanation in the article. The notation of Proto-Slavic is completely different from its reconstructible phonology. You write *o and understand */a/. That's hardly intuitive.
- I don't want to get down into the Stammbaum model discussion, only point out that, as usual, Stammbaum-like (divergence induced by the loss of regular contact between speech communities) and wave-like (contact-induced convergence) dynamics and patterns interact. There is no contradiction between the two models, they just describe different aspects of reality. Where physicists have particles and waves, we have trees and waves instead :) And of course, biologists have divergence and convergence phenomena between taxons (especially subspecies) too, so there's nothing unusual about it all.
- As Holzer conceives of Proto-Slavic, and I submit that this is the methodically clearest and most sound approach, it is defined as the MRCA of all idioms recognised as Slavic. Anything earlier or later is best labelled differently, for example Pre-Proto-Slavic or Common/Early Slavic. If Proto-Slavic is understood as a local dialect belonging to a much larger Balto-Slavic continuum, it is also natural that it should exhibit no internal divisions.
- By the way, I wonder the "Great Slavic Vowel Shift" and also some other phenomena like the Slavic umlaut could have been plausibly caused by the influence of Finno-Ugric speakers in Central Russia, who took over Slavic in the 7th century and distorted it with their accent, by mapping an unfamiliar vowel system on their native system. Check Proto-Uralic language#Vowels – there are some striking points of similarity between the Proto-Uralic and Common Slavic system (of course this implies the assumption that the Finno-Ugric dialects spoken in the relevant area in the relevant period retained the relevant properties or had developped even more of them). --Florian Blaschke (talk) 18:06, 28 October 2013 (UTC)
- There are some things I don't really understand here. There's a lot of possible sounds in between *a and *o, so arguing that Proto-Slavic had *a because it could not have had *o is a false dilemma. It's unlikely that the pronunciation flipped quite suddenly from /ɑ/ to /ɔ/. More likely (to me at least) is that the Proto-Slavic (pre-metathesis/polnoglasie) pronunciation should be reconstructed as /ɒ/. That is, a back open vowel like /ɑ/, but rounded nonetheless. There isn't really any requirement that *a (> *o) and *ā must have exactly the same pronunciation, there are many languages where short and long vowels have quite notable differences in quality without upsetting the balance between them. Hungarian (in which short a is also /ɒ/ in fact) is a very notable example. So I really do not understand why Proto-Slavic must be reconstructed without *o, or more specifically, why *o can't be understood to represent /ɒ/.
- Similar arguments can be applied to the second palatalisation in Old Novgorodian as well. It's likely that Proto-Slavic, and certainly Old East Slavic, had allophonic palatalisation of most consonants before the loss of the yers; Russian still preserves this more or less intact. There's no reason to say, therefore, that this didn't also apply to any consonant that preceded *ě from earlier *ai. What appeared in Old Novgorod must have been phonetically [kʲeː] or perhaps [ceː], but because the distinction between [c] and [k] was nonphonemic, the Cyrillic letter к sufficed for both sounds. So Old Novgorod did have syllabic synharmony, but all other Slavic languages then shifted this [kʲ] (which must have been the inherited sound) further to [ts] in a later post-Proto-Slavic stage. CodeCat (talk) 19:08, 28 October 2013 (UTC)
- However, if all evidence points to an unrounded [a]- or [ɑ]-like vowel prior to 800 AD and no evidence points to a rounded [ɔ]- or even [ɒ]-like vowel, it becomes pretty difficult to maintain that the vowel was ever rounded in the period in question, even at the very end of the period. Equally, if there is no indication that Old Novgorodian /k/ was palatalised before *ě from earlier *ai, your reconstruction is simply arbitrary (actually, even you accept that it was [kʲ] and not [ts] originally, and even if we grant that Proto-Slavic had such an allophonic palatalisation, which is by no means certain and in fact rather questionable, the notation c for a nonphonemically palatalised velar is simply misleading; note that Proto-Slavic is assumed even by Holzer to have had such an alternation, but it is the one traditionally called "First Palatalisation", with /k/ vs. /tʃ/ probably already phonemes at that stage, the alternation thus being morphophonological). Occam's Razor strikes here. You could be right – your solutions may even be more realistic (but I see no reason why there cannot be a sudden jump) – but methodically it's easier to assume the simpler solution where it's compatible with and actually suggested by the evidence. --Florian Blaschke (talk) 14:23, 29 October 2013 (UTC)
- What evidence is there for an unrounded vowel at that point? If the comparative method points to a rounded vowel, which it clearly does (all Slavic languages have one), then the evidence for an unrounded vowel must come from either outside Slavic, or be derived through internal chronology of sound changes. In other words, either a word borrowed from or into Slavic excludes a rounded vowel, or there are sound changes that have an earliest possible date that rely on the vowel being unrounded. It is important that borrowings explicitly rule out a rounded vowel, being "compatible" with an unrounded vowel is not enough; it's not unreasonable that both [ɒ] and [ɑ] could have been borrowed into another language as that language's /a/, so this does not rule out [ɒ]. Similar for the reverse, there's no reason that a foreign /a/ could not have been borrowed into Slavic as [ɒ]. Analogy with existing, earlier loanwords also plays a role: if Slavic already had some words in which a foreign /a/ matched their own [ɒ], then it's likely that any new words with /a/ would have also been borrowed with [ɒ] despite the match being less than perfect. You could call it a kind of "phonetic calquing", and English still does this regularly: Latin long /aː/ is still borrowed into English as /eɪ/ even though it's been 500 years since these two vowels actually matched. There's also evidence, I believe, that the Slavic languages themselves substituted matching sounds when borrowing from each other. CodeCat (talk) 15:16, 29 October 2013 (UTC)
- Those are mostly "learned borrowings" made and adapted by literate folks from a dead language and subsequently entering the vernacular, greatly influenced by spelling and schooling, whereas Slavic loanwords/glosses in Latin, Greek, Baltic, Finnic, as well as Romance or Greek words in Slavic are all pure vernacular words. All of them point to the difference of non-rounded *a/*ā until very late Common Slavic (early 9th century) when *a>*o. In fact, Holzer in the first 20 pages of Historische Grammatik des Kroatischen describes changes from classical Latin to spoken Balkans Romance, and in the end derives a correspondence table of Romance sounds to Slavic (with a, ā > a, ā). For your theory to be true, Slavic would have to replace a with *o, and all of the borrowed Slavic word in several languages families would have to do the opposite... Occam's razor again. We're dealing with thousands of words BTW. A fun exercise is to read the list of former toponyms in Greece or Vasmer's Die Slaven in Griechenland and guess Proto-Slavic word in toponyms. E.g. Βάριανη = barjāne (not *borjane), Γαρίτσα= *garīkā (not *gorica), Μαγούλα = *magūlā (not *mogyla), Καρούτια = *karūta (not *koryto) etc. but note the change of *a > *o already occuring in Βοδίκιον = *vodьcь (not *wadīkā; using -κ- instead of -τσ- for the output of third palatalization is due to the orthographic practice in Byzantine Greek) and Κορίτο = *koryto (not *karūta). Pure comparative method will only get you so far, and is inapplicable when the area of change is a very large dialect continuum and you need such clues to get to the earlier stage... You can force it methodologically and end up with something like PIE - Frankestein's monster combining different chronological layers and last-reconstructable-ancestors into a patchwork which works just fine in most of the cases (the most common conjugation *bʰéroh₂-*bʰéresi-*bʰéreti is non-existant in Anatolian and Tocharian? Who cares, it's "PIE proper"!). Same thing with Old Novgorod/Pskov attestations - regardless of whether к in кѣле represents non-assibilated [tʲ] or your [kʲ] or because in those dialects monophthongization occurred after the second platalization - the traditional reconstruction with *c, *z and *s is defective, as is any other representing any stage after it (which is about dozen more Common Slavic changes). -Ivan Štambuk (talk) 18:50, 29 October 2013 (UTC)
- Fully agreed, only a detail: Byzantine Greek κ before front vowels was probably pronounced as (IPA) [c], and the digraph κι before vowels probably represented [c], while τσ stands for true affricates such as [ts] or [tʃ], all just like in Modern Greek. Therefore, Βοδίκιον is probably to be read as [voˈðicon] and reflects a Common Slavic pronunciation [βɔdʲɘcɘ] or so. The conclusion that while Common Slavic č was pronounced [tʃ], c was apparently still pronounced [c] and not [ts] or [tɕ] yet doesn't affect your point, though, quite the opposite. --Florian Blaschke (talk) 19:30, 29 October 2013 (UTC)
- I don't find the examples terribly convincing. You still apply a false dichotomy between *a and *o and this skews the conclusion. There is no reason at all that a Slavic *gorica /gɒriːt(s?)ɑː/ could not have been borrowed into Byzantine as Γαρίτσα. Similar for Μαγούλα and Καρούτια. After all, what forced the Greeks to adapt Slavic *a-in-the-process-of-becoming-*o /ɒ/ into their native ο /o/, when α /a/ was a closer match? Compare it to the Hungarian or English /ɒ/: would other languages borrow this as /o/ or /a/? I'd say either one is possible. So all the Greek evidence really suggests is that the vowel was low and not low-mid at the time of the earliest borrowings, but it is not conclusive on the rounding. Given that we know that rounding eventually occurred in all dialects, /ɒ/ seems more likely than /ɑ/. In fact, if you ask me, the raising from /ɒ/ to /ɔ/ was a direct result of the accentual changes that happened in late Slavic that upset the short-long distinction of many of the vowels: it aimed to increase the articulatory distance between originally-short *o and newly-shortened *a to avoid them becoming indistinct. CodeCat (talk) 20:55, 29 October 2013 (UTC)
- Well this sounds more closer to *o than *a to me. Finnish, Greek and vulgar Latin all had short rounded vowel *o (or something very similar), and borrowings in both directions always point to unrounded *a in Slavic. Germanic *ō always gets borrowed as *ō > CS *u, never *a > CS *o, Vulgar Latin *o as *u>*ъ and *ō as *ū>*y. There is simply no evidence for short rounded *o in that stage of Proto-Slavic. "Rounding" (i.e. distinctive lowness/openness) is post-9th century but as has already been said, it's doubtful that you can speak of Proto-Slavic at that stage.. Re other: originally long vowels remained non-distinctively long even after the change and were not immediately shortened (and later rephonemicized, though the theories on the development of Slavic length vary from author to author). --Ivan Štambuk (talk) 04:29, 30 October 2013 (UTC)
- I don't disagree that if a Greek speaker had heard the vowel in that sound file, it would have been heard as /o/. But that sound file isn't actually really accurate. When I think of /ɒ/ I think of the vowel in English dog, and I pronounce it more open than in that sound file (but still with rounding, it's different from the /ɑ/ I use in Dutch) and I think many other English speakers do. I also notice though that the rounding is very indistinct when I pronounce the English /ɒ/. It's like the mouth itself just can't really articulate the rounding properly when the vowel is so open, there's barely any difference in the shape. So it's likely that speakers of other languages can't hear this distinction clearly either, until the vowel becomes more closed. (I know at least one Dutch speaker who substituted the American English /ɒ/, which can be somewhat lengthened, with /aː/ rather than use /ɑ/ or /ɔ/) I suppose this is just something that the evidence can't recover. We know that *a was rounded eventually in all dialects, so the most likely explanation to me is that it was already rounded nondistinctively in Proto-Slavic, but it was still very much an open vowel, and this rounding only became distinctive (to native speakers and foreigners alike) when it was closed to /ɔ/.
- As for Germanic *ō... that was a more closed vowel by that time, and the change to /uo/ is attested in early Old High German, so it's not surprising that CS substituted *u there once the older *au was gone. I do wonder though, if Germanic *ō was borrowed as *u, why not Latin *ō too? The case of Germanic *ō is actually parallel to the Slavic one, because it was also rounded and raised from an earlier *ā, and this also happened fairly late but still completely in all Germanic languages. The example of Gothic Rūmōneis was given above which supports this, it was probably borrowed into Proto-Germanic as /ˈruːmɒːniːz/. CodeCat (talk) 14:36, 30 October 2013 (UTC)
- Sure, but short non-rounded vowel is far more accurately represented by the symbol <a> than <o> surely you must agree. I don't find the discussion on phonetics much productive - all that comparative evidence can give is distinctive features, and roundness wasn't one of them. When you see reconstructions such as *golva you think of two vowels, one rounded and one not, which is misleading.
- Latin *ō was apparently changed to spoken Romance *ū (L Scardōna > Rom. Skardūna > Slavic skordynъ > SC Skradin; Latin Alvōna > Rom. Alβūna > Slavic olbynъ > SC Labin etc.) Holzer lists this change as number XII in his grammar. Much later Slavic also changed the inherited *ō (<*aw, *ew, *āw, *ēw) to redundantly long *u, as well as the one that came from borrowings (beside Germanic also e.g. L Salōna > Slavic Solunъ > SC Sòlūn; Latin Gregōrius > Slavic Gьrgurjь > SC Gȑgūr). So there is no difference to Germanic borrowings, it's only a matter of dating and origin.. --Ivan Štambuk (talk) 20:28, 30 October 2013 (UTC)
- If the matter is purely notation then things are different, yes. We should be clear on what is actually denoted by these letters. The reconstructive process goes more or less like this. First, comparison of phonemes leads to a set of nameless archiphonemes, which are more or less abstract representations of the sound correspondences. Then, attempts are made, given the existing evidence, to give each of those a probably phonetic value. And based on that, finally, a system of notation is developed to make it easier to denote the archiphonemes. So in the end, it doesn't matter for the meaning whether *a or *o is chosen, because both denote the same reconstructed phoneme, except that each one suggests a different phonetic value. But this still purely suggestion, and tradition can and often does lead to representations that match earlier ideas about phonetics and not newer ones. A good example is the PIE opposition between "velars" and "palatovelars", which is now more often thought of as an opposition between uvulars and velars. But the traditional symbols *ḱ, *ǵ are still used for the latter, even though the acute accent is often understood to denote a palatal articulation of a consonant. And then there's the "laryngeals" which are no longer understood to be laryngeal in articulation at all. Tradition can be hard to break sometimes. For Slavic, the traditional notation is still the most widely used, so that's what should be used here too. We should make a note on the possible/likely phonetic values associated with each letter, which the article did until it was removed because of lack of sourcing. I don't disagree with requiring a source for this, but at the same time if we don't discuss the phonetic values at all, then there is an implication, much like you said, that *o is /ɔ/ or /o/ or similar, which would be an error by omission. So there's a bit of a dilemma: either the article has to lie by implication due to lack of sources, or present at least some unsourced information about the phonetic values to avoid giving readers the wrong impression. CodeCat (talk) 20:52, 30 October 2013 (UTC)
- The traditional comparative method is faulty if the comparanda belong to different chronological layers. What appears as an outlier i.e. post-Proto-X development could in fact be an earlier stage, retention with others innovating. Occam's razor/principle of parsimony is valid only if we're dealing with daughters that conveniently split Stammbaumlike. Since we can make an educated guess on the phonology of spoken Greek, Romance, Baltic and Finnic during the period (the first two are much more documented, the latter two more archaic), we can similarly make assumptions on the phonology of Slavic words attested in them, or borrowings in Slavic, as well as internal development, on the basis of hard data. If have evidence that e.g. acuted *ar-, *al- changed to *ra-, *la- before a dozen or so changes that we normally reconstruct for "Proto-Slavic proper", then reconstructions obtained by comparative method alone such as *ordlo "plough" should be thrown out of the window because they do not fit the data. The same argument goes for the second palatalization of Old Novgorodian, using *o instead of more proper *a, or the mentioned simple thematic present of PIE (neo-PIE, not PIE, its absence is archaicism, not its presence). Qualities of what is designated as velars and palatovelars probably changed in pre- and post-Antolian-Tocharian PIE so it's pointless to assign them any kind of specific phonetic value valid for "real" PIE...
- All of this can be sourced and is indeed already used in some articles as Florian mentioned. When I get enough motivation I'll add it unless someone does it before. But there are some other important topics missing articles like the third palatalization, grammar.. perhaps the discussion phonology could be extracted to a separate article if it grows to large - I would also like to cover the reconstructions of vocalism by other major scholars such as Shevelov, Mareš, Lamprecht, Carlton... all of whom have different theories as well. --Ivan Štambuk (talk) 22:35, 30 October 2013 (UTC)
- You're still presenting an entirely different issue here. *o versus *a is a matter of notation, not of phonetics. We could denote this vowel as *V1 if we wanted to (kind of like the PIE laryngeals). It doesn't really matter which one you choose, both could represent the same sound for all I know. The mapping between symbols and sounds is arbitrary and does not necessarily have to be phonemic or even systematic (English...), so in establishing any notation for Proto-Slavic at all, we need to establish what those symbols represent. It really doesn't matter which we choose as long as it's clear what they mean. However, for Proto-Slavic, the majority of sources uses *o to denote the short vowel that developed from Proto-Balto-Slavic *a, and *a to denote the long vowel that developed from PBS *ā. And unless the majority of sources nowadays uses anything else, we should continue to use this notation, if not for representing the majority usage, then just to ease understanding (because it's what users expect). CodeCat (talk) 22:53, 30 October 2013 (UTC)
- But that's just valid for the last stage of Proto-Slavic, and Lunt, Schenker and other sources for article use *a as Early Proto Slavic, what is in fact Holzer's "real" Proto-Slavic (but with some modifications). There is really no issue notation-wise, we're just dealing with different time frames. The point is to illustrate that the traditional reconstructions which are largely based on OCS are inadequate. Usage of <o> is a 19th century remnant and was chosen not because it represents some abstract archiphoneme as you suggest, but because it was believed to indeed represent [o]-like sound. At that time historical linguists haven't even figured out conditions for palatalizations, relationship with Baltic group, let alone inspected loanwords etc. --Ivan Štambuk (talk) 00:20, 31 October 2013 (UTC)
- You're still presenting an entirely different issue here. *o versus *a is a matter of notation, not of phonetics. We could denote this vowel as *V1 if we wanted to (kind of like the PIE laryngeals). It doesn't really matter which one you choose, both could represent the same sound for all I know. The mapping between symbols and sounds is arbitrary and does not necessarily have to be phonemic or even systematic (English...), so in establishing any notation for Proto-Slavic at all, we need to establish what those symbols represent. It really doesn't matter which we choose as long as it's clear what they mean. However, for Proto-Slavic, the majority of sources uses *o to denote the short vowel that developed from Proto-Balto-Slavic *a, and *a to denote the long vowel that developed from PBS *ā. And unless the majority of sources nowadays uses anything else, we should continue to use this notation, if not for representing the majority usage, then just to ease understanding (because it's what users expect). CodeCat (talk) 22:53, 30 October 2013 (UTC)
- If the matter is purely notation then things are different, yes. We should be clear on what is actually denoted by these letters. The reconstructive process goes more or less like this. First, comparison of phonemes leads to a set of nameless archiphonemes, which are more or less abstract representations of the sound correspondences. Then, attempts are made, given the existing evidence, to give each of those a probably phonetic value. And based on that, finally, a system of notation is developed to make it easier to denote the archiphonemes. So in the end, it doesn't matter for the meaning whether *a or *o is chosen, because both denote the same reconstructed phoneme, except that each one suggests a different phonetic value. But this still purely suggestion, and tradition can and often does lead to representations that match earlier ideas about phonetics and not newer ones. A good example is the PIE opposition between "velars" and "palatovelars", which is now more often thought of as an opposition between uvulars and velars. But the traditional symbols *ḱ, *ǵ are still used for the latter, even though the acute accent is often understood to denote a palatal articulation of a consonant. And then there's the "laryngeals" which are no longer understood to be laryngeal in articulation at all. Tradition can be hard to break sometimes. For Slavic, the traditional notation is still the most widely used, so that's what should be used here too. We should make a note on the possible/likely phonetic values associated with each letter, which the article did until it was removed because of lack of sourcing. I don't disagree with requiring a source for this, but at the same time if we don't discuss the phonetic values at all, then there is an implication, much like you said, that *o is /ɔ/ or /o/ or similar, which would be an error by omission. So there's a bit of a dilemma: either the article has to lie by implication due to lack of sources, or present at least some unsourced information about the phonetic values to avoid giving readers the wrong impression. CodeCat (talk) 20:52, 30 October 2013 (UTC)
- Well this sounds more closer to *o than *a to me. Finnish, Greek and vulgar Latin all had short rounded vowel *o (or something very similar), and borrowings in both directions always point to unrounded *a in Slavic. Germanic *ō always gets borrowed as *ō > CS *u, never *a > CS *o, Vulgar Latin *o as *u>*ъ and *ō as *ū>*y. There is simply no evidence for short rounded *o in that stage of Proto-Slavic. "Rounding" (i.e. distinctive lowness/openness) is post-9th century but as has already been said, it's doubtful that you can speak of Proto-Slavic at that stage.. Re other: originally long vowels remained non-distinctively long even after the change and were not immediately shortened (and later rephonemicized, though the theories on the development of Slavic length vary from author to author). --Ivan Štambuk (talk) 04:29, 30 October 2013 (UTC)
- Those are mostly "learned borrowings" made and adapted by literate folks from a dead language and subsequently entering the vernacular, greatly influenced by spelling and schooling, whereas Slavic loanwords/glosses in Latin, Greek, Baltic, Finnic, as well as Romance or Greek words in Slavic are all pure vernacular words. All of them point to the difference of non-rounded *a/*ā until very late Common Slavic (early 9th century) when *a>*o. In fact, Holzer in the first 20 pages of Historische Grammatik des Kroatischen describes changes from classical Latin to spoken Balkans Romance, and in the end derives a correspondence table of Romance sounds to Slavic (with a, ā > a, ā). For your theory to be true, Slavic would have to replace a with *o, and all of the borrowed Slavic word in several languages families would have to do the opposite... Occam's razor again. We're dealing with thousands of words BTW. A fun exercise is to read the list of former toponyms in Greece or Vasmer's Die Slaven in Griechenland and guess Proto-Slavic word in toponyms. E.g. Βάριανη = barjāne (not *borjane), Γαρίτσα= *garīkā (not *gorica), Μαγούλα = *magūlā (not *mogyla), Καρούτια = *karūta (not *koryto) etc. but note the change of *a > *o already occuring in Βοδίκιον = *vodьcь (not *wadīkā; using -κ- instead of -τσ- for the output of third palatalization is due to the orthographic practice in Byzantine Greek) and Κορίτο = *koryto (not *karūta). Pure comparative method will only get you so far, and is inapplicable when the area of change is a very large dialect continuum and you need such clues to get to the earlier stage... You can force it methodologically and end up with something like PIE - Frankestein's monster combining different chronological layers and last-reconstructable-ancestors into a patchwork which works just fine in most of the cases (the most common conjugation *bʰéroh₂-*bʰéresi-*bʰéreti is non-existant in Anatolian and Tocharian? Who cares, it's "PIE proper"!). Same thing with Old Novgorod/Pskov attestations - regardless of whether к in кѣле represents non-assibilated [tʲ] or your [kʲ] or because in those dialects monophthongization occurred after the second platalization - the traditional reconstruction with *c, *z and *s is defective, as is any other representing any stage after it (which is about dozen more Common Slavic changes). -Ivan Štambuk (talk) 18:50, 29 October 2013 (UTC)
- What evidence is there for an unrounded vowel at that point? If the comparative method points to a rounded vowel, which it clearly does (all Slavic languages have one), then the evidence for an unrounded vowel must come from either outside Slavic, or be derived through internal chronology of sound changes. In other words, either a word borrowed from or into Slavic excludes a rounded vowel, or there are sound changes that have an earliest possible date that rely on the vowel being unrounded. It is important that borrowings explicitly rule out a rounded vowel, being "compatible" with an unrounded vowel is not enough; it's not unreasonable that both [ɒ] and [ɑ] could have been borrowed into another language as that language's /a/, so this does not rule out [ɒ]. Similar for the reverse, there's no reason that a foreign /a/ could not have been borrowed into Slavic as [ɒ]. Analogy with existing, earlier loanwords also plays a role: if Slavic already had some words in which a foreign /a/ matched their own [ɒ], then it's likely that any new words with /a/ would have also been borrowed with [ɒ] despite the match being less than perfect. You could call it a kind of "phonetic calquing", and English still does this regularly: Latin long /aː/ is still borrowed into English as /eɪ/ even though it's been 500 years since these two vowels actually matched. There's also evidence, I believe, that the Slavic languages themselves substituted matching sounds when borrowing from each other. CodeCat (talk) 15:16, 29 October 2013 (UTC)
- However, if all evidence points to an unrounded [a]- or [ɑ]-like vowel prior to 800 AD and no evidence points to a rounded [ɔ]- or even [ɒ]-like vowel, it becomes pretty difficult to maintain that the vowel was ever rounded in the period in question, even at the very end of the period. Equally, if there is no indication that Old Novgorodian /k/ was palatalised before *ě from earlier *ai, your reconstruction is simply arbitrary (actually, even you accept that it was [kʲ] and not [ts] originally, and even if we grant that Proto-Slavic had such an allophonic palatalisation, which is by no means certain and in fact rather questionable, the notation c for a nonphonemically palatalised velar is simply misleading; note that Proto-Slavic is assumed even by Holzer to have had such an alternation, but it is the one traditionally called "First Palatalisation", with /k/ vs. /tʃ/ probably already phonemes at that stage, the alternation thus being morphophonological). Occam's Razor strikes here. You could be right – your solutions may even be more realistic (but I see no reason why there cannot be a sudden jump) – but methodically it's easier to assume the simpler solution where it's compatible with and actually suggested by the evidence. --Florian Blaschke (talk) 14:23, 29 October 2013 (UTC)
- Once the Hittite was discovered with actual reflexes of laryngeals old Brugmannian forms immediately became obsolete (now called "post-PIE"). Similarly, in Old Novgorod birch bark letters we have actually attested вьхe and кѣле, this one was discovered in 1956 and I can't find any other reason for sticking to the traditional (OCS-like) forms other than inertia. The traditional Stammbamum model of protolanguages is inapplicable when there is a dialect continuum across large area because isoglosses do not spread uniformly and completely, and that's the reason why Holzer has been advocating such an early stage for Proto-Slavic "proper". Of course, if notation is just convention and what is meant is not what is written, and each model can more or less easily be transferred into its more "correct" equivalent, than it's not really an issue. --Ivan Štambuk (talk) 01:49, 28 October 2013 (UTC)
- Holzer's theory makes a lot of sense to me and I think User:Florian Blaschke is right about a significant issue being how different the new reconstructions look vs. the old ones. However there's also the issue crystallized in this quote from Lindstedt, quote in Holzer: "due to a simplistic ... model whereby languages, once separated from a common ancestor, can share no further innovations whatsoever". The assumption of definitive splits in a protolanguage is still the norm and there isn't really a good theoretical apparatus for handling the case where innovations continue to propagate across a dialectally differentiated area (a "Common X" period corresponding to Common Slavic). This works well in many cases either because the "Common X" period was fairly stable or because the dialectal differences during this period were eventually erased in all extant dialects prior to anything being recorded. But it's particularly problematic in Slavic where the "Common Slavic" period involved enormous changes and where we have direct documentation from within this period. Cf. similar issues in PIE w.r.t. the laryngeals (leading to vague "Post-PIE" reconstructions). An even more dialect parallel is Proto-Germanic where it's likely that reconstructed Proto-Germanic *ō was actually pronounced *ā at this period (there are many indications of this; one is the Gothic word "Rūmōn-" "Roman" from Latin "Rōmān-", which is inexplicable if *ō was /o:/ but easily explainable if it was /a:/). Benwing (talk) 00:40, 28 October 2013 (UTC)
- (unindenting) I suppose we should ask if what we intend to cover is really the last common ancestor, or something else. Do we really need two articles on this, Proto-Slavic language and Common Slavic language? I would say that for the study of the Slavic languages themselves, the latter is more useful and is generally cited in etymologies (including those on Wiktionary). Derksen's dictionary also covers Common Slavic. I have the impression that Proto-Slavic is really a linguistic ideal, and is unrealistic as a concept. No language really has no variety at all. Common Slavic on the other hand strikes me as much more natural, and it would certainly be a single "language" by modern standards. So if we ask what language the Slavic languages descended from, it would be Common Slavic. But the point of divergence is somewhat earlier, if it ever existed as a real language (which I doubt and can probably never be confirmed anyway). CodeCat (talk) 01:42, 31 October 2013 (UTC)
- I have the impression that Proto-Slavic is really a linguistic ideal, and is unrealistic as a concept. No language really has no variety at all.
- Your assumption is that Proto-Slavic was a dialect continuum itself. But what if it was simply a dialect forming part of the Balto-Slavic dialect continuum, a dialect which expanded suddenly in the 6th century and began to fragment no earlier than in the 7th? In fact, Holzer addresses your objection regarding the homogeneity of Proto-Slavic himself by suggesting it was used as a lingua franca in the Avar Empire (and beyond). Every proto-language logically arises as part of a larger dialect continuum. If a speech form is spoken in a very limited region or even a single settlement, where the speech community is very coherent (close-knit) due to regular and frequent contact between the speakers, its homogeneity should not be surprising. Just think of Old Latin, which was originally the dialect of Rome. The surrounding countryside spoke dialects that were slightly different and later absorbed by the Roman, urban form of Latin.
- Anyway, I'm not even sure how all of this is relevant to the article itself. My point was that the area contact/loan/secondary transmission provides evidence for an unrounded realisation of what is traditionally reconstructed as Common Slavic *o (and none for a rounded realisation, which if not Finnic, then at least Romance should provide with its distinction between /a/, /ɔ/ and /o/ timbres in short vowels), a rounded realisation of what is reconstructed as Common Slavic *y and *ъ, etc., as late as the beginning of the 9th century, and no timbre distinction between long and short vowels, which means that even the Middle Common Slavic stage retained a Baltic-like system and did not exhibit the Great Slavic Vowel Shift yet. In any case, the article should note the complications and problems with the traditional OCS-oriented reconstruction and controversy about or different approaches to the reconstructed phonology of Early and Middle Common Slavic. After all, the whole question whether long and short vowels had the same timbre is not only about ultimately perhaps unrecoverable phonetic details, but affects the whole phonological system.
- By the way, is it the fact that we need to rely on external evidence (but even internal evidence, namely Novgorod/Pskov Slavic, indicates that something is amiss) to reconstruct an unrounded realisation of *o, etc., and that "internal" (intra-family) comparative reconstruction alone does not clearly lead to this Baltic-looking system and that it is not obvious from internal evidence alone that Proto-Slavic preceded the Great Slavic Vowel Shift, is it this fact which bothers you so much? Personally, I understand that, but it's just an empirical fact that "internal" (intra-family) comparative reconstruction tends to lead to systems that are not as archaic in many respects as the "true" historical stage probably was (because that's what we usually find when we have actual evidence for such a stage), due to innovations propagating throughout an already differentiated dialect continuum after the breakup of the homogeneous proto-language stage (for example, case and gender syncretism in Romance), so Slavic isn't an exceptional case here. You seem yourself well aware of the limitations of proto-language reconstruction. --Florian Blaschke (talk) 01:31, 6 November 2013 (UTC)
- The turning point in the Slavic vowel system was most likely the shortening of acuted syllables, and other pretonic/posttonic shortenings (I think?), before Dybo's law and the j-retraction. Everything before that point reflects the Balto-Slavic system more or less, although we know for a fact that immediately before that shortening, short and long vowels already had distinct qualities (because the shortening caused no mergers). So any internal evidence of the original vowel system with unrounded *o must come from this time, but no later. The only actual reconstructive evidence for this from within Slavic, that I know of, is the outcome of the liquid diphthongs. In many Slavic languages, *or metathesizes to *ra, and *er to *rě. This is internal evidence that these vowels must have still been paired phonologically at the time of the change, and thus *a was essentially still a "long *o". CodeCat (talk) 18:06, 25 April 2014 (UTC)
Vowel Phonetics
I have to agree with Ivan that Schenker doesn't give phonetic values for the phonemes. I have asked the OP to provide a different source on his talk page. μηδείς (talk) 19:16, 12 October 2013 (UTC)
- You can't list phonetic values because some segments have uncertain features (even and especially jat). --Ivan Štambuk (talk) 21:40, 12 October 2013 (UTC)
- The OP seems rather convinced such values do exist. I am aware of the difficulties, and I pulled Comrie off my shelf the verify the Schenker. I am hppy to wait to see if there's another source for reconstructed phonetic values or not. μηδείς (talk) 21:45, 12 October 2013 (UTC)
- I can add a cited table of distinctive features for Late Proto Slavic (which is apparently described, though dubbed "Middle Common Slavic", whatever that is). There are also some other issues with that section, but those IPA symbols were simply the most conspicuous and annoying part. --Ivan Štambuk (talk) 16:57, 14 October 2013 (UTC)
- I think sources generally disagree on how to label the different stages of Slavic, especially because there are so many changes towards the end of the period. I think User:Benwing was the one who wrote most of the information about the periods... in fact I think most of the content of the current article was written or rewritten by him in some form. A lot of the original content was split off to History of the Slavic languages. CodeCat (talk) 20:45, 14 October 2013 (UTC)
- What is described as Middle Common Slavic is in pretty much every book I saw labeled Late Proto Slavic or Common Slavic. MCS is supposed to me some intermediate stage before LCS and after Late Balto-Slavic/Early Proto-Slavic. --Ivan Štambuk (talk) 23:56, 18 October 2013 (UTC)
- The term "Middle Common Slavic" is from Horace Lunt "On the Relationship of Old Church Slavonic to the Written Language of Early Rus'" (1987). This is p. 135 where he says "the hypothetical linguistic system of Common Slavic posited for the 700s, which is familiar from etymological dictionaries, and which I dub Middle Common Slavic (MCoS)". I adopted this term because I wanted to distinguish Proto-Slavic (a single reconstructed lect) from Common Slavic (a dialectally differentiated entity, but in a period where developments still occur pan-dialectally), and emphasize the fact that there is a still later stage of Common Slavic, and the "Early Proto-Slavic" vs. "Late Proto-Slavic" (or "Common Slavic") terminology doesn't have a term for this period. This last Common Slavic stage is termed "Late Common Slavic" in Lunt and is more-or-less directly attested in both OCS and Old East Slavic (aka "Early Rus"), and is generally thought to have ended upon the breakdown of yers (although granted this occurred at different times in different places). The problem is that there aren't any universally agreed-upon terms or even periodizations. Kortlandt uses "Late Middle Slavic" as approximately the same as "Middle Common Slavic" here, and divides my "Late Common Slavic" into two periods ("Young Proto-Slavic" and "Late Proto-Slavic"); see History of the Slavic languages. The thing is that (at least according to Kortlandt) there are so many important Common Slavic changes, esp. in accentation, that occurred during the LCS period that it seems important to distinguish this stage. Benwing (talk) 00:14, 28 October 2013 (UTC)
- But why not simply write Late Proto-Slavic, with its unambiguous meaning "the form that can be reconstructed by comparative method" ? It's fine to separate the internal history of protolanguage into stages, and show how it evolved, but when you read the article all that it gives is some MCS and it remains unclear whether it's the real protolanguage that the article is supposed to describe, or some antecedent form. --Ivan Štambuk (talk) 02:08, 28 October 2013 (UTC)
- The term "Middle Common Slavic" is from Horace Lunt "On the Relationship of Old Church Slavonic to the Written Language of Early Rus'" (1987). This is p. 135 where he says "the hypothetical linguistic system of Common Slavic posited for the 700s, which is familiar from etymological dictionaries, and which I dub Middle Common Slavic (MCoS)". I adopted this term because I wanted to distinguish Proto-Slavic (a single reconstructed lect) from Common Slavic (a dialectally differentiated entity, but in a period where developments still occur pan-dialectally), and emphasize the fact that there is a still later stage of Common Slavic, and the "Early Proto-Slavic" vs. "Late Proto-Slavic" (or "Common Slavic") terminology doesn't have a term for this period. This last Common Slavic stage is termed "Late Common Slavic" in Lunt and is more-or-less directly attested in both OCS and Old East Slavic (aka "Early Rus"), and is generally thought to have ended upon the breakdown of yers (although granted this occurred at different times in different places). The problem is that there aren't any universally agreed-upon terms or even periodizations. Kortlandt uses "Late Middle Slavic" as approximately the same as "Middle Common Slavic" here, and divides my "Late Common Slavic" into two periods ("Young Proto-Slavic" and "Late Proto-Slavic"); see History of the Slavic languages. The thing is that (at least according to Kortlandt) there are so many important Common Slavic changes, esp. in accentation, that occurred during the LCS period that it seems important to distinguish this stage. Benwing (talk) 00:14, 28 October 2013 (UTC)
- What is described as Middle Common Slavic is in pretty much every book I saw labeled Late Proto Slavic or Common Slavic. MCS is supposed to me some intermediate stage before LCS and after Late Balto-Slavic/Early Proto-Slavic. --Ivan Štambuk (talk) 23:56, 18 October 2013 (UTC)
- I think sources generally disagree on how to label the different stages of Slavic, especially because there are so many changes towards the end of the period. I think User:Benwing was the one who wrote most of the information about the periods... in fact I think most of the content of the current article was written or rewritten by him in some form. A lot of the original content was split off to History of the Slavic languages. CodeCat (talk) 20:45, 14 October 2013 (UTC)
- I can add a cited table of distinctive features for Late Proto Slavic (which is apparently described, though dubbed "Middle Common Slavic", whatever that is). There are also some other issues with that section, but those IPA symbols were simply the most conspicuous and annoying part. --Ivan Štambuk (talk) 16:57, 14 October 2013 (UTC)
- The OP seems rather convinced such values do exist. I am aware of the difficulties, and I pulled Comrie off my shelf the verify the Schenker. I am hppy to wait to see if there's another source for reconstructed phonetic values or not. μηδείς (talk) 21:45, 12 October 2013 (UTC)
User:Ivan Štambuk's deletion of references
I don't think it's a good idea to delete all the references that happen not to be cited in the article, and I don't think there's a WP rule requiring such deletion. At the very least the remaining useful references should go into a "See Also" section or similar. Many of the deleted references are highly important to understanding Proto-Slavic and their presence in the article is useful as a summary of where the reader should look for more info. I agree there are many references that can be pared but this should be done selectively rather than with an axe. Benwing (talk) 00:56, 28 October 2013 (UTC)
- They are wasting 8K of space of an already big article. Additionally, they give the article a false impression of being a synthetic work of all of them, which it is not. I'll put the unreferenced works as ==Further reading==. --Ivan Štambuk (talk) 01:18, 28 October 2013 (UTC)
A missing verb sub-conjugation?
Could a verb conjugation that forms the present stem with -ne- and has a zero-suffix in the infinitive stem be reconstructed? There are ten or more such root verbs in Serbo-Croatian, and while other languages don't have this conjugation (to my knowledge), cognates seem to be found in both the 1st and the 2nd conjugation. Is this just an analogical development following the nasal-infix verbs? — Preceding unsigned comment added by Crom daba (talk • contribs) 14:55, 10 March 2014 (UTC)
- @Crom daba: Can you give an example? --Ivan Štambuk (talk) 06:35, 13 October 2014 (UTC)
rj
- The pronunciation of *ř is not precisely known, but it was approximately a palatalized trill [rʲ]. It survives as a separate phoneme only in Czech
Uhm, Russian has pairs like /rʲad/ vs. /rad/. Are we correct to say that only Czech preserves this as a separate phoneme? — Preceding unsigned comment added by 213.87.134.5 (talk) 02:27, 6 December 2014 (UTC)
- Yes, at least as far as Proto-Slavic goes. The Russian word ряд was *rędъ, so it didn't have *ř. The palatalisation in this word happened only in the separate history of Russian, after Proto-Slavic times. CodeCat (talk) 02:32, 6 December 2014 (UTC)
- Czech's development is no different from Russian in this respect. *r preceding front vowels merged with *ř in both languages, producing palatalized *rʲ. This remained as /rʲ/ in Russian, but turned into a fricative trill in Czech (and Old Polish). No modern Slavic language distinguishes *ř and *r before front vowels. Benwing (talk) 03:47, 7 January 2017 (UTC)
table manners
Now, I might be completely wrong here, but it really FEELS like there has been a slight mixup in one of the conjugation tables at the end, specifically, the final one (nouns in class C), wherein the first three nouns or so in the plural subtable seem to have their Dative and Instrumentative conjugation entries juxtaposed. Being one of the unwashed masses, I haven't the credentials to correct this, but can SOMEONE of such an illustrious and learned background PLEASE look into this and correct if necessary? Thanks much, Mwa! 23.242.248.211 (talk) 09:25, 19 May 2015 (UTC)
- Ugh. Hate "responding" to myself (but hate screwing up attribution entries even more), but slight amendment to the above: make that ALL the tables seem to suffer from this malady, i.e., the first three rows of nouns in each of the three tables seem to have their plural Dative and Instrumentative cases (possibly!) juxtaposed. Can someone just run a searing, omniscient eye over yon and confirm or, uh, disconfirm same? Toots. 23.242.248.211 (talk) 09:34, 19 May 2015 (UTC)
Derksen's and Jasanoff's prosodic notations
In this article, there seem to be four different types of accented syllable for Late Common Slavic: short rising à, long falling ȃ, short falling ȁ and neoacute ã. For MCS, there's also a symbol for the unshortened acute, long rising á. However, I'm finding that some other prominent sources show them differently.
Derksen's Etymological Dictionary of the Slavic Inherited Lexicon distinguishes instead short rising à, long rising á, short falling ȁ and long falling ȃ. Thus, he includes an extra long rising accent, while having no special symbol for the neoacute. The long rising accent he places on *bě́lъ "white" and pǫ́tь "way", where it appears to indicate a neoacute retraction onto a previously long syllable. For neoacute onto originally short syllables , he indicates short rising, e.g. *nòžь "knife".
Jasanoff's The Prehistory of the Balto-Slavic accent instead distinguishes long rising a̋, long falling ȃ, short falling ȁ, short rising à and long neoacute ã. Potentially confusing is that Jasanoff uses a̋ for the acute accent where Derksen uses à, while Jasanoff's à indicates short neoacute as well as short accent in non-initial syllables. Jasanoff's ã appears to correspond exactly to Derksen's á.
These discrepancies make it difficult to add sourced reconstructions to the article, because they may differ in notation and may also carry assumptions that each of the authors make about the nature of the Slavic accentual system. Determining the historical accent type from their notations is not straightforward, in particular with Derksen's. Adding each form in its original notation will be faithful to the source, and avoid original research, but it will confuse the hell out of the readers of this article in a topic that is already difficult to follow. It is in the interest of the readers to use the same prosodic notation throughout the article. But if we do so, how do convert the notation of various sources like the two above, to the one that the article uses? A table showing the differences may be useful, to both readers and editors. Rua (mew) 10:55, 26 April 2019 (UTC)
- I agree, a table would be in order. — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 15:28, 26 April 2019 (UTC)
@Aeusoes1: Here is my attempt:
Jasanoff Kapović |
Olander | Derksen Kortlandt | |
---|---|---|---|
Old short, initial | ȍ ъ̏ | ȍ ъ̏ (ȏ ъ̑) | |
Old short, medial and final | ò | ||
Short neoacute, initial and medial | ò ъ̀ | ò ъ̀ (ó ъ́) | |
Old acute, initial and medial | a̋ | à | |
Old acute, final | à | ||
Circumflex, initial | ȃ | ||
Long neoacute, initial and medial | ã | á |
The vowel symbols I used stand for different vowel classes that the accent can appear on: o stands for short mid vowel, ъ stands for short high vowel (yer), a for any long vowel or liquid diphthong. Short and circumflex in initial syllables are always in AP c, where Dybo's law failed to shift the accent rightward. Medial old short always results from Dybo's law. Medial and final circumflex/yers don't exist because they always undergo the neoacute retraction. For this same reason, neoacute doesn't exist finally.
Overall, the two schemes are the same. They use different symbols for the long neoacute, but both use a symbol that unambiguously indicates it. The key difference between the two schemes is how they treat the old acute nonfinally: Derksen and Kortlandt consider it shortened, while Jasanoff and Kapović don't. They both consider short neoacute to be identical to shortened acute and to medial old short, but whether they actually have the same reflexes in all languages I don't know. Derksen and Kortlandt reconstruct a circumflex in yer-final AP c "monosyllables" originally having a short vowel, like *bȏgъ and *dь̑nь, and Kortland appears to reconstruct long neoacutes for short ones in this position as well (*dь́nъ, gen pl). Jasanoff puts the etymologically original short vowel here, *bȍgъ.
It would be helpful if other authors could be added to the table, perhaps with another column if they use yet another notation. The remaining question is which notation we should use. Both notational schemes can be automatically converted into one another, so that we don't actually lose information by using only one. I'm inclined to use Jasanoff's notation, which is more traditional and thus has less WP:DUE issues than the more radical Leiden school to which Derksen belongs. A compromise/mix between the two notations is also an option, e.g. Jasanoff but with short acute everywhere. Such a compromise would eliminate the need for the symbol a̋, which eases the mental load on the reader a little because it's easily confused with ȍ. Rua (mew) 10:45, 7 May 2019 (UTC)
Further edit: I've also had a look at Olander's notation. It's essentially the same as Jasanoff's, but uses ´ for the long neoacute like Derksen and Kortlandt. Rua (mew) 11:57, 7 May 2019 (UTC)
The English Wiktionary has now adopted the compromise notation I described. See wikt:Wiktionary:About Proto-Slavic#Accents for a description, including a modified version of my table above. I propose using this same notation in Wikipedia as well. Thoughts? Rua (mew) 17:01, 10 May 2019 (UTC)
- I'm for it. — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 17:50, 10 May 2019 (UTC)
Requested move 18 October 2020
- The following is a closed discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. Editors desiring to contest the closing decision should consider a move review after discussing it on the closer's talk page. No further edits should be made to this discussion.
The result of the move request was: weak consensus to move to Proto-Slavic language for consistency and better recognizability. No such user (talk) 08:23, 20 November 2020 (UTC)
Proto-Slavic → Proto-Slavic language – Move requested for consistency, since all proto-languages currently have "language" at the end. This also disambiguates proto-languages from proto-peoples and other "proto-"'s that are not languages. — Sagotreespirit (talk) 18:07, 18 October 2020 (UTC)
- Support as there is more consistency with a title like this when comparing to other Wikipedia articles of proto languages. Quahog (talk • contribs) 20:56, 18 October 2020 (UTC)
- Comment The consistency spoken of was created by the move requestor moving 37 of these articles today. Kanguole 22:33, 18 October 2020 (UTC)
- Weak Oppose. Adding "language" to the article name is a disambiguation tool, as explained at WP:NCL. If there is no article about proto-Slavic people, then there is no need for disambiguation. It seems redundant, but it's also not hurting anything if we add it. — Ƶ§œš¹ [lɛts b̥iː pʰəˈlaɪˀt] 02:01, 19 October 2020 (UTC)
- Support I don't know what "Proto-Slavic" name of the article should mean. As for "Proto-Slavic language" we know what this means or represent. Mikola22 (talk) 19:50, 27 October 2020 (UTC)
- The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Usage of the letter 'e' with ogonek and dot above (ę̇)
I've found an interesting letter used in this article and other Proto-Slavic reconstructions. I've been confuddled by dots both above and below letters, especially the letter 'e' for some reason, in reconstructions of other proto-languages such as Proto-Turkic and I believe Proto-Samic. The letter is an 'e' with an ogonek and a dot above, ę̇, and I found it in the conjugation of the word dušà (soul), both on Wiktionary and in this article (here), in the singular genitive, plural nominative and plural accusative forms of the word. Is this a mistake(I've seen it used in other places so probably not...?)? If not, (I know this isn't a Q&A help site but) could someone explain why and how it is used in Proto-Slavic reconstructions? I can't find anything about the usage of this letter in Proto-Slavic, and it would be super helpful with my studies:)) Logan Sherwin (talk) 01:34, 4 November 2021 (UTC) — Preceding unsigned comment added by Logan Sherwin (talk • contribs) 22:30, 3 November 2021 (UTC)
- If the dot is actually a vertical line, it's a prosodic marker and it's mentioned in Proto-Slavic accent. – filelakeshoe (t / c) 🐱 23:03, 3 November 2021 (UTC)
- No, I don't think so. I copied the character into the heading of this thread (it looks like a dot) and added links to the place where I found it. I've also found it here in the vowel alternations chart, it's in the 8th and 9th boxes of the "After soft consonants" row. It could be a typo where a macron (vertical line) is needed instead, I'm just not sure and I'd need someone to give more insight. Logan Sherwin (talk) 01:34, 4 November 2021 (UTC)
- It's explained in more detail at History of Proto-Slavic#Nasalization. In short, *ę̇ represents the phoneme that must be reconstructed as the outcome of pre-Slavic *uN, *ūN after a palatal consonant. This vowel has a different outcome from "regular" *ę in many languages: it denasalises to *ě in West and East Slavic, but merges with *ę in South Slavic. Rua (mew) 19:42, 4 November 2021 (UTC)
- No, I don't think so. I copied the character into the heading of this thread (it looks like a dot) and added links to the place where I found it. I've also found it here in the vowel alternations chart, it's in the 8th and 9th boxes of the "After soft consonants" row. It could be a typo where a macron (vertical line) is needed instead, I'm just not sure and I'd need someone to give more insight. Logan Sherwin (talk) 01:34, 4 November 2021 (UTC)
jьměti
This verb is mentioned to be class V (athematic). Seems so, but is there a source to prove? I have found only ones proving there are 4 verbs in the class, with the 5th "jьměti" missing. VerbaGleb (talk) 10:42, 8 February 2023 (UTC)
Error in table
Table AP a nouns, entry for jìlъ, has jìlъmъ as instrumental, while judging where it patterns with AP b and c, the form should be jìlъmь (cf. AP b võlъ/volъ̀mь, AP c sy̑nъ/sy̑nъmь). Jalwikip (talk) 19:02, 9 April 2024 (UTC)
Did Proto-Slavic ever exist?
Since there is no evidence at all that Proto-Slavic ever existed it is doubtful whether there were any Slavic languages before 500 AD. One theory by Omeljan Pritsak claims that the Slavic languages derived from Lithuanian mercenaries that were employed by the Awars after the collapse of the Roman Empire. This Balto-Slavic connection has been mentioned in the article. Pritsak´s theory - wildly rejected by Russian nationalists - should be paid attention to. --Ontologix (talk) 02:05, 12 August 2024 (UTC)