Wiktionary talk:About Proto-Slavic

Some proposed changes

• removing 80% of the article which deals with history - that's encyclopaedic. WT:AEN doesn't explain Great Vowel Shift. Some of it is quite disputable (e.g. proto-language being spoken, IPA values for reconstructed segments, where exactly was Proto-Slavic spoken)
• document existing templates used for headword line and inflection
• references on which entries should be created (i.e. no OR)
• templates (textual, not MW) for the list of descendants, new entries (noun, adjectives, verbs).
• notation for accents - I prefer the classical one, the one by Derksen based on Serbo-Croatian gives me headaches --Ivan Štambuk (talk) 04:50, 9 August 2013 (UTC)

Stuff to do

• Create an appendix on Common Slavic accent paradigms
• Add an optional parameter to {sla-PoS} to indicate accent paradigm and link to the appendix. Should accept two values (in case of uncertainties).
• Rewrite inflection templates in Lua so that their number, as well as the number of parameters they require, is reduced - i.e. detection of the type of stem, and whether the consonant should undergo palatalization before the desinence. --Ivan Štambuk (talk) 07:39, 9 August 2013 (UTC)

Early Proto-Slavic

Recently mostly Early Proto-Slavic (i.e. the "real" Proto-Slavic, as opposed to Common-Slavic which we term Proto-Slavic) forms is dealt with in the literature, i.e. forms with distinctive lengths, diphthongs, closed syllables etc. We could provide both proto-forms, together with inflections (which differ from sources to source, because the development of word-final clusters is disputed, so there would be multiple inflections per specified sources), and a description of changes (sounds, accents) that occurred in the Common Slavic period. --Ivan Štambuk (talk) 06:36, 11 August 2013 (UTC)

*kt, *gt -> *ť?

Whence is this rule? What for? Especially when only page titles follow it. E. g. there is page Appendix:Proto-Slavic/moťi but all derivatives linking to it list PS form *mogti. So what’s the purpose of this change in PS lexicon? Same with *noktь/*noťь – Silmethule (talk) 12:40, 14 August 2013 (UTC)

That change was Common Slavic. Strictly speaking, both *noktь and *noťь are true, it's just that the former is an earlier form of the latter. (note that before velar consonants, kt and xt change to t instead) *noktь is the traditional form used in the dictionaries. Personally I'm agnostic on the issue.

The reason why etymologies link to *mogti is because the Proto-Slavic page was originally created as *mogti, and the person who moved it hasn't followed the standard procedure of updating the references to the old page as well. --Ivan Štambuk (talk) 12:50, 14 August 2013 (UTC)

The difference between -gt-/-kt- and -tj- is not reconstructable via the comparative method. They all share the same outcome in the Slavic languages which means that they merged in Proto-Slavic at some point. Only by internal reconstruction, or by referring to earlier forms (Balto-Slavic, PIE) can we know which of the two was the original form. We could decide to write it as tj instead, but ť seems more suitable as an abstract symbol as the exact pronunciation of this sound is rather uncertain. w:Proto-Slavic and w:History of the Slavic languages give more information. —CodeCat 13:15, 14 August 2013 (UTC)

There is lots of evidence in onomastics (toponyms, personal names) written in other languages, as well as borrowings from Latin and Greek to Slavic. We can even date these changes very precisely. Georg Holzer in Historische Grammatik des Kroatischen shows that kt, xt > k'ť, x'ť, and only much later do these change to ť. --Ivan Štambuk (talk) 13:59, 14 August 2013 (UTC)

Also, looking at other present forms of descendants we can reconstruct some of the original kt', gt' groups, like *mogti basing on eg. modern Polish mogę, możesz (however it only works, when first element is part of stem and the second – of ending).

And, regarding to writing ť instead of tj – it makes no sense, when nj, rj and lj are preferred – one consistent way of iotation would look better. // Silmeth ^@talk 17:29, 27 August 2013 (UTC)

Every other book seems to have its own way of marking palatalized/palatal consonants. However, I agree that it's inconsistent to have digraphs in case of */lj/ or */nj/, while simultaneously having */ť/ and */ď/. There is also the problem of not having the distinction of iotated/palatalized consonants and normal sequences such as *lj and *nj* in internal history of Proto-Slavic in extended etymologies (see *korljь; and also *t' and d* that originated from the third and second palatalizations which occurred before iotation, and which later became */c/, */dz/). Apostrophe <'> or <’> seems to be the most widespread one, perhaps we should settle on it? --Ivan Štambuk (talk) 08:42, 3 September 2013 (UTC)

Now that I think about it, we should definitely provide all alternative forms for reconstructions, as they appear in the literature. I was thinking of using some box to list these instead of cluttering the headword line, which would then contain only the "Wiktionary standard form". The box would have two columns, one for reconstructions and another for the list of works that use them. Works would be listed by shortcuts which would on click link to an appendix page that lists their full name, author, etc.. --Ivan Štambuk (talk) 23:23, 14 August 2013 (UTC)

Some proposals

• For nouns, adjectives and verbs an optional parameter ap= is added which could take values a, b, c and d (the d paradigm is reconstructed by Russians), including their combinations when there could be multiple reconstructed (e.g. a-b, similar like we have for genders). These parameters would also categorize.
• For adjectives we make principal parts feminine and neuter, in addition to the masculine. These would be optional, i.e. generated from the masculine. This would be particularly useful in cases when accents are different in different genders. For example, the headword line for *bělъ could be: bě̃lъ m (feminine *běla̍, neuter *bělo̍), which more easily explains e.g. Serbo-Croatian b(ij)ȇl, b(ij)éla, b(ij)élo. If not provided via parameters, these feminine and neuter forms could easily be generated in Lua.
• For the verb we make principal part, beside the infinitive, the first-person singular present. For example, the headword line for *běgati would be: bě̋gati (first-person present tense bě̋gajǫ). In other words, we have both present stem and the infinitive stem listed. --Ivan Štambuk (talk) 01:13, 30 September 2013 (UTC)

Soft sign

If ь = ĭ then actually soft sign doesn't exists in proto-slavic? If exists then which letter will be '(apostrophe) in phonetic transliteration? I mean for exmaple russian голубь = gólubʹ, where ь = '. Useigor (talk) 04:18, 31 December 2013 (UTC)

ь/ĭ was a vowel in Proto-Slavic, and so was ъ/ŭ. ь was pronounced like i, ъ like y or u, but less clear and shorter. Eventually, these two vowels often disappeared in the later Slavic languages, but sometimes ь/ĭ left softening as a "residue" after it disappeared. Russian writers then started to use ь (which was not a vowel with its own sound anymore) to show softening of the preceding consonant, and started to write it in places where the vowel ь had not originally been. In older Russian spelling the ъ was still written in some places too, but it was not pronounced as anything and it was dropped in 1917. —CodeCat 04:26, 31 December 2013 (UTC)

I mean i want to know which letter makes soft consonant. I don't understand why some stems called "soft". Useigor (talk) 22:55, 1 January 2014 (UTC)

Some letters are soft and others are hard. Soft are č, š, ž, c, ś, j, ť, ď (and also nj, lj, rj). A stem that ends with one of those sounds is a soft stem. —CodeCat 23:39, 1 January 2014 (UTC)

Please note that consonants š, ž, c (ш, ж, ц) are almost always hard in Russian (but šč - щ is soft), (adding ь doesn't make them soft), can be both ways in Ukrainian, c can be both ways in Belarusian but š and ž are always hard. There are more soft (palatalised) consonants. --Anatoli ^{(обсудить}/^вклад) 23:58, 1 January 2014 (UTC)

But what about dz? Is it soft like c(ts)? Useigor (talk) 05:53, 4 January 2014 (UTC)

Church Slavonic

Wikitiki89 (talk • contribs) keeps removing it from the descendants list for unknown reason. Church Slavonic recensions are generally treated as separate languages, with their own literatures, spelling rules, dictionaries and grammars. They have attestations for many words not found in the limited corpus of OCS canon. Etymological dictionaries always mark words found in them as such (CS, not OCS). They should thus be marked as a separate language in the descendants list as well. --Ivan Štambuk (talk) 13:57, 16 August 2014 (UTC)

I didn't "keep removing" it; I removed it once. Anyway, I'm not so much against having it, but against its location. It's not its own subfamily of Slavic, but still a part of South Slavic and should probably be as a sub-element of OCS. --WikiTiki89 14:06, 16 August 2014 (UTC)

Russian Church Slavonic is South Slavic? --Ivan Štambuk (talk) 15:58, 16 August 2014 (UTC)

Well it's a direct continuation of OCS in written-only form, isn't it? Use of Medieval Latin in England doesn't make it not an Italic language. --WikiTiki89 16:18, 16 August 2014 (UTC)

It's spoken not written-only. OCS is merely a carefully selected oldest layer of what was already at the time separate literary languages. --Ivan Štambuk (talk) 14:27, 17 August 2014 (UTC)

It's not "spoken", just read aloud. --WikiTiki89 19:37, 17 August 2014 (UTC)

In other words, it was spoken. --Ivan Štambuk (talk) 20:10, 17 August 2014 (UTC)

This is just a matter of semantics and has nothing to do with whether it is a South Slavic language. --WikiTiki89 14:43, 18 August 2014 (UTC)

Church Slavonic recensions were adaptations according to the local vernacular. Yes the basis was "South Slavic" (I'm putting question marks because at that time of OCS MSS there was no geographical separation of the Slavic speech yet, and such notion is anachronic) and these adaptations far outnumber the few isoglosses (phonological, grammatical, lexical) that characterize OCS as South Slavic from the modern perspective. --Ivan Štambuk (talk) 15:19, 18 August 2014 (UTC)

So then it might make sense to place it in a different place depending on where the particular word is found. As the default location, it makes sense to place it as a sub-element of OCS. --WikiTiki89 16:08, 18 August 2014 (UTC)

Why you're using nj, rj, lj and dz?

In my opinion ń, ŕ, ĺ are better, they save space and make text more readable. Look at the list: ť *ď, *č, *ž, *š, *c, *lj *nj *rj, *j — 3 last are like black sheeps. Also in fact first two are using caron (just does not fit with lowercase d,t), so the list in capital would look so: Ť, Ď, Č, Ž, Š, C, Ĺ, Ń, Ŕ (Ś, Ź). As you can see caron is for palatalized hushing consonants (at least not just -j) and acute for soft consonants (-j).

I absolutely don't get why we're using dz while only polish it does. Don't you think that common slavic change is g > dz (or dź) > z (or ź) otherwise we must not write epenthetic l since it's barely appear in west slavic. We can't randomly choose when to listen to west slavs when don't.

• kъnjiga / kъńiga
• prijateljь / prijateĺь (never have seen proto-slavic words/roots with not iotated vowel after i)
• brjuxo / bŕuxo
• cěsarjь / cěsaŕь
• morje / moŕe

Useigor (talk) 08:39, 17 August 2014 (UTC)

I agree, digraphs should be eliminated. I don't understand your second paragraph? --Ivan Štambuk (talk) 14:30, 17 August 2014 (UTC)

I agree too that we should use single characters for these. But it makes more sense to me if we use hačeks for all of them, so ň, ř, ľ. And I don't understand your point in the second paragraph either. The phoneme *dz is reconstructed for Proto-Slavic based on the comparative method. Not just Polish has this sound, but also Macedonian and some forms of Old Church Slavonic. The fact that the *dz ~ *z distinction reflects an older distinction inherited from Proto-Balto-Slavic shows that it is not an innovation within each of these languages. And, being inherited, this implies that it must be reconstructed for Proto-Slavic proper. —CodeCat 14:39, 17 August 2014 (UTC)

I can compare too. In third palatalization g > dz/dź in all Slavic languages, but later it was shortened to z/ź in all Slavic languages except Polish, Polabian and Old Church Slavonic. Therefore g > dz > z - is common slavic by clear majority. But you still use dz for some reason. Also i see that pj and similar changed to plj only in South and East Slavic. In West Slavic languages letter l almost doesn't appear (polish czapla looks suspicious). In this case South and East Slavs are in majority and that's right that we write epenthetic l. My conclusion is that there can be 2 variants: dź & pj or ź & pĺ.

About carons, if we're gonna use soft ź and ś and with acuts therefore soft ń, ŕ, ĺ must use acuts as well to be same with ź and ś, since ž and š are in use already and they represent another palatalized sounds. I mean that acute for soft consonants (e.g. lj / ĺ), caron for palato-alveolar (they are in most descendants).Useigor (talk) 20:33, 17 August 2014 (UTC)

I think you misunderstand what Proto-Slavic is. It's the common ancestor of all Slavic languages, not just of some of them. If we say that Polish dz descends from *z, that would not make sense. We have to use the form that is ancestral to every Slavic language. Majority has nothing at all to do with it. Compare for example what we do with Proto-Germanic *z. Almost all Germanic languages reflect this phoneme as r, the only ones that don't are Gothic and Proto-Norse, both very old languages. So if we went by a majority here, we'd have to use *r for Proto-Germanic, but that would not make sense because *z is the older, ancestral form of the sound. Likewise, Proto-Germanic *ē is reflected as ē in Gothic but as ā in all other Germanic languages, but we reconstruct *ē because that is the original form of the sound. —CodeCat 20:39, 17 August 2014 (UTC)

That said, recent examination of the Old Novgorod dialect suggests that we should reconstruct neither *dz nor *z, but actually *g (or *ǵ?). The Old Novgorod letters were actually written in an obscure form of Slavic that did not take place in the second or third palatalisations, which affected every other Slavic language we know of. The fact that it did not take part in the change in the same way shows that the palatalisation actually took place after Proto-Slavic had split into dialects to some degree. —CodeCat 23:55, 17 August 2014 (UTC)

Seems your position is clear for me. Also i like ǵ, kъnęǵь with it looks good, and i think you know that in Russian feminine of this word is княгиня (kńagińa). But i still don't understand epenthetic l (i mentioned in my both messages) - it's not common for Western Slavs therefore we must write zemja (zem-ja), stьbjь (stьb-jь), čapja (čap-ja) and etc. — Useigor (talk) 22:57, 18 August 2014 (UTC)

I like the idea of using *ǵ and *ḱ and was actually going to suggest it myself. If we do this, I would also move *gvězda and *květъ to *ǵvězda and *ḱvětъ. This is inconsistent, though, with the use of *ś for palatalized *x, but *x́ would look pretty weird. --WikiTiki89 23:45, 18 August 2014 (UTC)

There is one point we need to examine first though. In Common Slavic, vowel fronting after palatal consonants followed the application of the progressive palatalisation. This is why we have *kъnęgъ > (progressive palatalisation) *kъnęǵъ > (fronting) *kъnęǵь > (assibilation in non-Novgorod Slavic) *kъnędzь. The progressive palatalisation (one of the two which did not fully affect Old Novgorod) would have first palatalised the final consonant, and following that its declension would have switched over from "hard" to "soft" due to vowel fronting. This means, then, that we must look for signs of soft declensions in Old Novgorod nouns that underwent the progressive palatalisation in other Slavic dialects. If there is evidence of this, then we can conclude that *kъnęǵь is the latest form common to all Slavic languages, and it would also mean that the progressive palatalisation did apply to ON, but did not progress as far. If there is no soft declension for such nouns in ON, then we must necessarily reconstruct *kъnęgъ for all of Slavic. —CodeCat 00:35, 19 August 2014 (UTC)

Based on w:ru:Древненовгородский диалект#Собственно дрененовгородские явления, some of the attested ON reflexes of the forms of *vьśь are в[ъ]хоу (v[ŭ]xu), въхо (vŭxo), отъ въхоѣ (otŭ vŭxojě), въ въхъ (vŭ vŭxŭ), овхо (ovxo, “fully”), вхого (vxogo), на вхыхъ (na vxyxŭ), вьхѣ (vĭxě), вьхѣмъ (vĭxěmŭ). It also contradictorily states that all the forms have the stem вьх- (vĭx-), but I don't think that makes that big of a difference. It is pretty clear that it uses the hard declension. --WikiTiki89 01:19, 19 August 2014 (UTC)

• The jury is still out on how to explain aberrant features of North Russian speeches. According to some they are not exemption from e.g. second palatalization, but rather later innovations or special developments such as a postponed monophothongization. The are alternative reconstructions of Proto-Slavic (e.g. by Georg Holzer) that go chronologically earlier to account for such anomalies, and for other reasons. According to Holzer's chronology the "middle steps" from k, g > c, dz (> z) are not ḱ, ǵ but rather t', d'. There is also the issue of the little fact that 2nd palatalization is a very old isogloss and has been completed before some other Common Slavic changes took place (such as iotation, the rise of jers and nasal vowels..) so mixing such chronologically different stages is a big no-no. Those alternative reconstructions are slowly being accepted so we might add them in 5-10 years when theories get sufficiently fleshed out.
• Regarding dz - it is attested in Glagolitic (the older part of the corpus) OCS where it is overwhelmingly marked with a special letter: ⰷ (dz), and occasionally with z (this at a point when the change dz > z was completed, the letter ⰷ lost its original function and was used strictly as a numeral). Glagolitic script is generally assumed to be purely phonemic, reflecting Constantine-Cyrillus' vernacular, and that dz is given a special symbol points to its authenticity as a "separate sound". Furthermore, the Cyrillic (younger) part of the OCS corpus regularly writes z, which indicates that dz is indeed a genuine output of the 2nd palatalization, and z a later change.
• Regarding the divergent developments of l-epenthesis there are two explanations in the literature: 1) Some Slavic areas didn't exhibit the epenthesis on morpheme boundaries 2) the change was regular everywhere both within a morpheme and at morpheme boundaries, but l was lost at morpheme boundaries in some areas. Regarding the evidence - again the older part of the OCS canon has more consistent usage of epenthesis in all positions, almost completely regular in Glagolitic corpus. E.g. in Codex Assemanius (early 11th c., Glagolitic) 342 attestations have the epenthesis while 85 lack it; in Sava's book (11th c., Cyrillic) out of 205 attestations 90 are without epenthetic l, while in Codex Suprasliensis (also 11th c., Cyrillic) 695 out of 791 forms lack the epenthesis. West Slavic evidence is a bit less clear since some early 10th century attestations lack the epenthesis, but seconary evidence (toponyms such as Czech Vidovle and Davle, 14th-century Hungarian attestation of Slovak Ribelyn vs. today's Rybany or a 1276 attestation of Hleulan vs. today's Chlievany, isolated words such as Polish kropla, grobla : Kashubian kropla, grable, konople : Lower Sorbian grobla : Slovak hrobľa, hrobeľ) indicates that it was originally a Common Slavic change and developments in Bulgarian-Macedonian and West Slavic developments are secondary. That's what most scholars agree with, though few hold that the few exceptions in West Slavic are not archaisms but secondary developments as well (e.g. change of syllable boundary: grob|ja > gro|bja where the epenethesis was regular within a morpheme). At best the evidence is inconclusive and is another argument in favor of pushing the reconstruction o "true" Proto-Slavic to an earlier stage.

There is also the possibility to use superscript ʲ everywhere which looks much nicer IMHO. ť and ď look like ejectives. --Ivan Štambuk (talk) 00:12, 18 August 2014 (UTC)

*bata, *baťa, *batja (ЭССЯ-1-163)

What's the difference between *baťa & *batja and how to type second word if we change third to baťa? —Игорь Телкачь 17:23, 23 December 2014 (UTC)

ť is the notation we use for the iotated t, which originated from a former -tj- sequence. So ť and tj are the same thing. —CodeCat 19:22, 23 December 2014 (UTC)

Wordless descendants

For me most problematic part in creating pages are descendants since they're using different scripts, special chars, diacritics. So no problem if i'll be creating pages where descendants have no words, except the one that will have word? —Игорь Телкачь 01:03, 11 January 2015 (UTC)

I don't agree with leaving out the descendants. They're crucial for any reconstructed term. —CodeCat 01:05, 11 January 2015 (UTC)

Is it forbidden to create pages with one descendant? People who're familiar with language may type words. —Игорь Телкачь 01:21, 11 January 2015 (UTC)

• It's OK to create appendices with a single reflex (e.g. Russian). You can also type reflexes without checking their spelling, placing an accent, or only transliterations, so someone else can check them later. We have folks like Anatoli that do just that (it gives them great pleasure :) --Ivan Štambuk (talk) 01:25, 11 January 2015 (UTC)

Descendants with mid3

What do you think about dividing 3 Slavic groups into 3 blocks? Church Slavonic can be written above them.

East Slavic Template:mid3 South Slavic Template:mid3 West Slavic

—Игорь Тълкачь 18:09, 15 April 2015 (UTC)

Excellent idea Boris. Other than cell phones and similar technologically primitive replacements for a real computer, modern computer screens have a large width-to-height ratio and horizontal placement of lists results in a more compact entry, that can be read and skimmed more easily, reducing unnecessary scrolling. --Ivan Štambuk (talk) 13:49, 7 August 2015 (UTC)