Talk:Slevan
Unresolved questions with the revision of Slevan
(for the two Jans primarily)
I have been thinking of how best to map the Vulgar Latin and Common Slavic vowel systems to each other. As Jan van Steenbergen pointed out to me it's quite unrealistic that VL ē maps to CS *ě and VL ĕ to CS *e; rather it should be the other way around with VL ĕ eventually becoming je, or to state it abstractly vowel quality should be more important than vowel quantity in the mapping of VL to CS!
OTOH I do think it's realistic to have VL ō become u since in CS *u2 developed from *ō which in turn developed from *au, and VL au in fact merged with ō, so that the product of this merger could have been qualitatively identified with CS *ō < *au before this was raised to [u].
I'm now quite certain that unstressed VL ē and ō merge with ĭ and ŭ (*ь and *ъ) respectively -- except in absolute final position, since I need to "preserve" certain endings, notably the ablative singular of the second declension, the first person singular present indicative of verbs and the nom/acc plural of the third declension.
I've cheated with the liquid metathesis, having no qualitative change. As JvS pointed out Mrác's ancestors should have been MORTIUS rather than MARTIUS! Should I let myself get away with this?
Consonants: I have decided that VL dj and gj become j à la Slovene since Slvanjec badly needs postvocalic js. So dz is out; MEDIUS becomes mjej and Mr. Hrodzán becomes Hroján; also MAGIS becomes maj by way of *MAGIUS.
Last but not least I have cheated with the third palatalization and VL ct, assuming c, z in all contexts. Actually I'm not too keen on changing this because I want more cs. What do you think? Is such an over-generalization permissible in a hybrid language? BPJ 10:37, 16 Jul 2005 (PDT)
Jan I's comments
- I have been thinking of how best to map the Vulgar Latin and Common Slavic vowel systems to each other. As Jan van Steenbergen pointed out to me it's quite unrealistic that VL ē maps to CS *ě and VL ĕ to CS *e; rather it should be the other way around with VL ĕ eventually becoming je, or to state it abstractly vowel quality should be more important than vowel quantity in the mapping of VL to CS!
Indeed, the other way around would be more logical. However, in Wenedyk I finally decided to stick with what I already had:
- VL ē and oe match with CS ě
- VL ĕ and ae match with CS e
I know it is against my own advise, and I seriously contemplated the change. But my ultimate conclusion was that it would change Wenedyk beyond recognisability. And frankly, also beyond my taste; strange enough almost every example I tried with the "new" system just felt plain wrong to my intuition. So I decided to keep the old system. And I'm happy with it. ;)
- OTOH I do think it's realistic to have VL ō become u since in CS *u2 developed from *ō which in turn developed from *au, and VL au in fact merged with ō, so that the product of this merger could have been qualitatively identified with CS *ō < *au before this was raised to [u].
I agree, that's totally realistic. Frankly, I can't see why there should by any tension between this and point 1. If anything, the fronted counterpart of this would be e > i, which frequently happens in Slezan.
- I'm now quite certain that unstressed VL ē and ō merge with ĭ and ŭ (*ь and *ъ) respectively -- except in absolute final position, since I need to "preserve" certain endings, notably the ablative singular of the second declension, the first person singular present indicative of verbs and the nom/acc plural of the third declension.
That's an interesting thought. I haven't thought about that. Can you come up with a few samples? I just might adopt this idea into Wenedyk (provided that I won't have to rewrite 1/3 of my entire lexicon).
- I've cheated with the liquid metathesis, having no qualitative change. As JvS pointed out Mrác's ancestors should have been MORTIUS rather than MARTIUS! Should I let myself get away with this?
Why not? Names tend to behave a little differently from "normal" vocabulary. Another thing is that often names are from much later date than the era of the Roman (even in romlangs). And finally, Mrác could have other sources in Latin as well: Mǐratius or whathaveyou.
- Consonants: I have decided that VL dj and gj become j à la Slovene since Slvanjec badly needs postvocalic js. So dz is out; MEDIUS becomes mjej and Mr. Hrodzán becomes Hroján; also MAGIS becomes maj by way of *MAGIUS.
A very sensible choice! I didn't know it happens in Slovene as well, but I dó know that it is standard in (most?) Romance languages. In fact, I think it is a great improvement; according to my sources, dj and gj were already pronounced /j/ in the Vulgar Latin period (Italian forms like maggio IIRC seem to be later scientific reborrowings).
And of course, you always have the option of turning dj into dz anyway in certain cases, which then can very simply be explained as an early scientific reborrowing!
- Last but not least I have cheated with the third palatalization and VL ct, assuming c, z in all contexts.
Could you please tell me what the exceptions are? I just might to change that in Wenedyk, because frankly, I think I ended up with a little too much of c!
- Actually I'm not too keen on changing this because I want more cs. What do you think? Is such an over-generalization permissible in a hybrid language?
IMO it is. It's your language after all. Especially because Slvanjec is based on Slovak far more loosely than Wenedyk is based on Polish, I think you have quite some freedom in setting some of your own rules when that helps the final result.
Cheers, IJzeren Jan 14:43, 16 Jul 2005 (PDT)
BPJ's comments on Jan I's comments
On ē and ĕ
- Indeed, the other way around would be more logical. However, in Wenedyk I finally decided to stick with what I already had:
- * VL ē and oe match with CS ě
- * VL ĕ and ae match with CS e
I think leaving things as they are would also decrease the number of medial jes, which from my æsthetic POV would be a Good Thing, since there are dj lj nj tj all over the place already -- it's becoming a bit too parodic!
On e and o becoming jers
I'm not too sure about Ē and Ō becoming ĭ and ŭ if I don't change the reflexes of Ĕ and Ē. I don't want to get jers all over the place and certainly not Ī and Ū coming up "strong" in the wrong places (where they wouldn't even be strong in Southern Romance) all over the place. BPJ 08:35, 17 Jul 2005 (PDT)
Slovene *dj > j
It's actually only *dj which becomes j in Slovene, but letting DJ and GJ behave the same makes sense from a Romance POV.
Still I'm not totally convinced by your argument on Italian DJ: why would mezzo be a borrowing?, and at least NDJ becomes nz everywhere! Cf. Old French where even MJ and NJ often become /ndʒ/ e.g. SIMIA > singe.
Exceptions to third palatalization
Basically when an u or o vowel follows. Strangely with following a the outcome is variable. One book I read suggested that it may have to do with wether a preceding *i was from PIE *ī or *ei and that following u is the only sure exception, but Jan II and his wife (who is the family linguist) wouldn't agree! BPJ 15:01, 16 Jul 2005 (PDT)
Jan I's comments on BPJ's comments on Jan I's comments
- I think leaving things as they are would also decrease the number of medial jes, which from my æsthetic POV would be a Good Thing, since there are dj lj nj tj all over the place already -- it's becoming a bit too parodic!
Fully agreed. So that's settled then! :)
- I'm not too sure about Ē and Ō becoming ĭ and ŭ if I don't change the reflexes of Ĕ and Ē. I don't want to get jers all over the place and certainly not Ī and Ū coming up "strong" in the wrong places (where they wouldn't even be strong in Southern Romance) all over the place.
Again, I agree. I've given it a little thought and finally decided against it. FWIW, in the development of Polish, stress didn't play a role at all.
- Still I'm not totally convinced by your argument on Italian DJ: why would mezzo be a borrowing?, and at least NDJ becomes nz everywhere! Cf. Old French where even MJ and NJ often become /ndʒ/ e.g. SIMIA > singe.
Yes, that's true. I must check my resources.
Exceptions to third palatalization
- Basically when an u or o vowel follows. Strangely with following a the outcome is variable. One book I read suggested that it may have to do with wether a preceding *i was from PIE *ī or *ei and that following u is the only sure exception, but Jan II and his wife (who is the family linguist) wouldn't agree!
I can't give any conclusive answer. From what I understood from Grzegorz Jagodziński's pages (if I recall correctly), third palatalisation happens very irregularly and most often doesn't take place at all.
But the way you formulated it, I got the impression that there are also exceptions to the KT > TS rule. I that true?
Cheers, IJzeren Jan 12:30, 18 Jul 2005 (PDT)
- Basically that same book as had strange ideas on the contexts of palatalization also had the idea that originally only *kti/*ktī becomes ci.
- The idea with the preceding monphthong/diphthong was of course meant to explain why 3d palatalization is irregular -- basically claiming that it is regular, but that it actually precedes monophthongization, and this influences its occurrence, to which the Havlišes objected. BPJ 13:41, 18 Jul 2005 (PDT)
Jan II.'s comments
I am far from being competent to introduce your Slavo-romance discussion, but I can add some hints on history of Slavic lang:
On III. palatalisation:
It happens after i, soft jer and e-nasal if the palatalised consonant is followed by vowel. If there is cluster V-C-hard jer (grěšnica X grěšnikU) or V-C-y (kUne.ndzI X kUne.gyni), the terminal hard jer or y prevented palatalisation of C, when V fulfils conditions for it. Quite often morphological “equilibration” has happened (nom. grěšnikU, gen. grěšnika). Quite often appeared doubletons (je.ga/je.dza). It should happen along with II. palatalisation, although there are not unified opinions on when in happened. It is also unclear when it begun. What is sure, it happened almost 100 years after monophthongisation.
And here is the discussed topic, the Slavic glottochronology. The source 2 is newer and I found it more competent because of other issues. And for sure, source 2 knew source 1 and used it.
source 1 – G. Shevelov, The historical phonology of Common Slavic, 1964
source 2 – A. Lamprecht, The Common Slavic, 1987
| type of change | source 1 | source 2 |
|---|---|---|
| s > ŝ (š, ch) | before 400 EC | before 400 EC |
| k’, g’ > s, z | before 400 EC | before 400 EC |
| a, o > å | before 400 EC | before 400 EC |
| I. palatalisation | 400-500 EC | 400-475 (±25) EC |
| monophtongisation | 500-600 EC | 510-585 (±15) EC |
| II. palatalisation | 500-600 EC | 575-650 (±25) EC |
| III. palatalisation | 600-850 EC | 575-675 (±25) EC |
| palatalisation of alveolars | N/A | 675-750 (±25) EC |
| epenthetic l’ | N/A | 710-785 EC |
| metathesis of liquids | 700-850 EC | 750-825 (±25) EC |
| appearance of nasals | 600 EC | 750-825 (±25) EC |
| appearance of jers | 800 EC | 800 EC |
| dispalat. ‘ě > ‘a in Pol. | N/A | 850-925 EC |
| extinct. of nasals | N/A | 950-1025 EC |
| contraction VjV | 900 EC | 875-950 EC |
| vocalisation of jers | 900-1150 EC | 950-1025 EC |
Benct's comments to Jan II's comments
- happens after i, soft jer and e-nasal if the palatalised consonant is followed by vowel.
Check. So far I got things right in Slvanjec.
- terminal hard jer or y prevented palatalisation
Alas I wasn't aware of this restriction, hence the name of the lang is not Slvanjek < SILVÁNICUM, and because the name of the lang is quite well established I'm very reluctant to "correct" this. Perhaps it will be permissible to over-apply 3. palatalization in a hybrid language, since having -ICUM become -jec rather than -jek gives a more distinct "flavor" to the language compared to other Romance languages!
One possible reason might be that the Roman contingent in the original mix of peoples didn't grasp the restriction and over-generalized the palatalization, but alas that doesn't tally very well with the dating you give -- even though I'm suspicious of any absolute dating of pre-historic changes!
Another possibility is that Vulgar Latin short U (hard jer) became a schwa-like sound earlier than what happened in true Slavic languages and so failed to block the 3. palatalization.
In a way I would now have liked Slevan to agree better both with Wenedyk and the Slavic languages in this respect, but OTOH I feel a reversal of this feature to be too drastic at this point. Also I don't want there to be too few instances of c and z in Slevan, and input to the 2. palatalization is already scarce as it is. What do you (both) think? BPJ -- written at the library and hence no proper log-in or signing! :-(
- Why not to implement some "morphological equilibrations"? E.g. the genitive forms underwent the palatalisation and had affected the nominatives... thus it would be _slvanjec_ and not _slvanjek_ (I deduce, that NOM.SING had ending -ekU, but the genitive had vowel other than U or y). Or you may simply ignore this restriction (see, only one vowel _u_ in both forms u /ECSL/ > U /CCSL/ or u: > y stopped the III. palatalisation). Jan II.
- There was analogical leveling out between the second and fourth declension, so that at some point the gen.sg. of all words from the 2. declension became -y, e.g. slváncy; otherwise it would be **slvánczi. Also the -ICU ending does not become -ekU, but the je is from a strong soft jer. I looks like I will simply ignore this restriction, but I still would like an explanation for doing so. BPJ 09:26, 19 Aug 2005 (PDT)
Jan I's comments on Jan II's comments and on Benct's comments to Jan II's comments
First of all, thanks for the interesting scheme you provided, Jan. If I may ask: which of the two columns has your own preference? And where do prof. Večerka's opinions fit?
Regarding third palatalisation: like I said before, I don't use it very often in Wenedyk. Grzegorz Jagodziński's pages were not clear about the circumstances under which it occurs, and therefore I mostly adopted the notion that it happens "irregularly". In other words, I use it when it fits my taste. Jan II's input might cause some changes, though: I didn't know all that either, but it makes sense.
All in all, I think the principle of "analogical levelling" gives us plenty of space to follow our own tastes. Indeed, both Wenedyk and Slezan use something based on Slьvanьkъ, fem. Slьvanьka. That might have spawned something like *slvanjek, *slvanjca, which then became Ślewanik, Ślewańka in Wenedyk and Slvanjec, Slvanjca in Slevan.
Of course, there is a remote possibility of having the whole pair. Say, slvanjek for the adjective, and slvanjec for the language (from LINGUA SILVANICA > *lęgva sьlvanьka > ljadzva slvanjca).
Since I recently got into contact with Grzegorz himself, I might as well ask him personally what he thinks about 3rd palatalisation.
--IJzeren Jan 06:25, 20 Aug 2005 (PDT)
Benct's comments on Jan I's comments on...
Of course it is not the 3. palatalization that applies irregularly, but analogical levelling that happens irregularly!
It is a very good idea that analogy in Slevan may have started from the feminine. I didn't think of that! Alas the name of the language can hardly be derived from LINGUA SILVANICA: that would be adjectival use while language names in Latin are normally neuter; that's why you have Latinum, Italiano, Español, Français, not Latina, Italiana, Española, Française. BTW the current form of the feminine is Slvánca -- one of the changes between version 3 and 3.5. BPJ 13:07, 20 Aug 2005 (PDT)
Jan II's comments on Jan I's comments on Jan II's comments and on Benct's comments to Jan II's comments
First of all, thanks for the interesting scheme you provided, Jan. If I may ask: which of the two columns has your own preference? And where do prof. Večerka's opinions fit?
As I said, the second column has my definite preferences. Our old Slavistic/Indoeuropeistic gang in Brno (Večerka-Lamprecht-Blažek) seems to be quite reliable. They included lots of new findings, which made their work quite rigorous. The is already a new generation (Rejzek-Palasová), with them I am in contact (although it looks they are not happy with any slaviconlanging). I have also ordered within amazon.co.uk a book about Slavic glottochronology by Alexander M. Schenker, but amazon is VERY slow in obtaining the copy for me.
Problems with III. palatalisation might be that different Slavic langs treated differently the new forms, as mentioned above. Jan II.
Input from Grzegorz Jagodziński
I asked Grzegorz Jagodziński (whose pages about Polish sound changes were a major inspiration to me in the creation of Wenedyk) about his opinion. He tried to edit this page himself, but was blocked for some reason. After that, he sent me a long mail instead. For now, I'll reproduce the Polish text below (sorry Bengan, but I though that's better than nothing); tomorrow, I hope to translate it into English. --IJzeren Jan 13:11, 22 Aug 2005 (PDT)
It happens after i, soft jer and e-nasal if the palatalised consonant is followed by vowel.
W podręcznikach podają również pozycję po sonantycznym ŕ̥ (< ьr < ir). Znam tylko 2 przykłady z scs. (starocerkiewnosłowiański): mŕ̥cati 'zmierzchać się, to grow dusk' oraz rastŕ̥zati (rastŕ̥ʒati) 'rozerwać, roztargać, to tear up'. Odpowiedniki "innosłowiańskie" (z innych języków, w tym z polskiego) nie mają palatalizacji (staropolskie mierzkać, dziś zmierzchać się z nieregularnym ch na miejscu k). Być może jest to związane z faktem, że sonanty zjawiły się w językach zachodniosłowiańskich później, a we wschodniosłowiańskich nie było ich nigdy. A przecież po grupie -ьr- palatalizacja raczej nie powinna zachodzić... i nie zachodziła.
Nic dziwnego w tym, że istniały różnice dialektyczne: przecież już II palatalizacja zachodziła niejednakowo w różnych językach słowiańskich: *kwaita- > *květъ || *cvětъ (polskie kwiat, ale rosyjskie cvet). Co więcej, formy w rodzaju kěna zamiast ogólnosłowiańskiego cěna są znane z okolic Nowogrodu Wielkiego. Być może były więc dialekty, w których II palatalizacja nie zaszła w ogóle.
If there is cluster V-C-hard jer (grěšnica X grěšnikU) or V-C-y (kUne.ndzI X kUne.gyni), the terminal hard jer or y prevented palatalisation of C, when V fulfils conditions for it. Quite often morphological “equilibration” has happened (nom. grěšnikU, gen. grěšnika). Quite often appeared doubletons (je.ga/je.dza).
Gdyby y lub ъ rzeczywiście wstrzymywały III palatalizację, oczekiwalibyśmy jej w formach z następującą inną samogłoską. A przecież -ika i -ica są niemal równie częste! Tak samo spotykamy równie często -ьca jak -ьka, -ьce jak -ьko itd. Np. *polьʒa 'korzyść' (scs), ale *ulьga 'ulga' (pol., z innym przedrostkiem), staropolskie oboczne jajce i jajko, miejsce i miastko (< *městьce || *městьko), itd.
Moim zdaniem raczej należy szukać: 1) uwarunkowań morfologicznych (tak przypuszczał już Jan Otrębski, który zmarł w 1971 roku), 2) różnic dialektalnych. Oto kilka przykładów:
- *děvica, ale *děvьka (scs. děvica 'virgin, girl', pol. dziewica 'virgin', dziewka 'girl' (trochę przestarzałe)) - przecież nie ma słów *děvikъ ani *děvьkъ,
- *žitnica 'stodoła' (scs, żytnica znana chyba nawet w staropolskim, nie chce mi się szukać), ale (scs) bližika bliska, krewna', ǫžika 't.s (krewna)' - widać, że -ica jest w rzeczowniku nieosobowym, ale -ika w osobowym, jednak obok tego także *děvica (osobowe!),
- scs zna naricati 'nazywać się', dviʒati 'dźwigać', polski zna tylko tylko narzekać (z zupełnie innym znaczeniem), dźwigać,
- twardy jer wcale nie hamował palatalizacji: *otьcь 'ojciec', *zajęcь 'zając', *měsęcь 'miesiąc, księżyc', *pěnęʒь 'pieniądz' (zapożyczenie z germańskiego), *kъnęʒь 'książę, ksiądz', *vьśь 'wszystek, cały' - a przecież wszystkie te wyrazy wywodzą się od form twardotematowych: *otьkъ, *zajękъ, *měsękъ, *pěnęgъ, *kъnęgъ, *vьxъ - i żadna z tych form nie zachowała się w żadnym z języków słowiańskich.
And here is the discussed topic, the Slavic glottochronology. The source 2 is newer and I found it more competent because of other issues. And for sure, source 2 knew source 1 and used it.
source 1 – G. Shevelov, The historical phonology of
Common Slavic, 1964
source 2 – A. Lamprecht, The Common Slavic, 1987
| type of change | source 1 | source 2 |
|---|---|---|
| s > ŝ (š, ch) | before 400 EC | before 400 EC |
| k’, g’ > s, z | before 400 EC | before 400 EC |
| a, o > å | before 400 EC | before 400 EC |
| I. palatalisation | 400-500 EC | 400-475 (±25) EC |
| monophtongisation | 500-600 EC | 510-585 (±15) EC |
| II. palatalisation | 500-600 EC | 575-650 (±25) EC |
| III. palatalisation | 600-850 EC | 575-675 (±25) EC |
| palatalisation of alveolars | N/A | 675-750 (±25) EC |
| epenthetic l’ | N/A | 710-785 EC |
| metathesis of liquids | 700-850 EC | 750-825 (±25) EC |
| appearance of nasals | 600 EC | 750-825 (±25) EC |
| appearance of jers | 800 EC | 800 EC |
| dispalat. ‘ě > ‘a in Pol. | N/A | 850-925 EC |
| extinct. of nasals | N/A | 950-1025 EC |
| contraction VjV | 900 EC | 875-950 EC |
| vocalisation of jers | 900-1150 EC | 950-1025 EC |
- I palatalizacja: początek naszej ery (1 A.D.) - przełom II / III wieku (200 A.D.)
- II palatalizacja: IV - VI w. n.e.
- III palatalizacja: koniec VII w n.e.
- I palatalizacja: wiek III i IV n.e.
- II palatalizacja: "po wieku VI", tj. wiek VII lub VIII n.e.
- III palatalizacja: początek VIII wieku.
