Aelf's Scratchpad

[Ælfwine]

NB: Originally called "Pelsodian Scratchpad," I decided to turn this thread into a more general scratchpad for all of my work.

Pelsodian is a romance language spoken by a couple thousand people around the modern city of Kostel (OTL Keszthely in Hungary). Its name derives from the Latin name for Lake Balaton, Lacus Pelsodis. Inspired by Dan Jones's Dravian and Ray Brown's Britainese, Pelsodian follows the tradition of naturalistic romance languages, as opposed to "bogus" romance languages, like Brithenig, and artificial romance languages, like Interlingua.

Much of my effort is focused reconstructing the Vulgar Latin spoken in the Pannonian region through analyzing inscriptions found there. Several papers I've used to help me, such as Attila Gonda's work on Pannonian Latinity. (See here) While there is a (somewhat) clear demarcation line between the "Eastern" Romance of southern Italy and the "Western" Romance of northern Italy, the line is less clear where the Gallo-Romance languages end and Balkan Romance languages begin. The weight of the evidence throws me behind a more Gallo-Romance model, for example, Pannonia shares a possible Celtic substrate with the North Italian and Rhaeto-Romance languages. Furthermore, historically the state of Pannonia was ruled by the Carolingian Empire and would've received Frankish influence. Nonetheless, the language's position on the periphery of the Empire needs to be considered as well. I believe the language most similar to it will be Friulian. Differing it from Friulian will be key to its success as a romance language and a conlang.

Questions to ask myself:

Will Pelsodian adopt a Western Romance vowel paradigm or an Eastern Romance vowel paradigm?
Will Pelsodian lose gemination or keep it?
Will Pelsodian have voiced its intervocalic voiceless consonants or kept them devoiced?
Will Pelsodian adopt plurals ending with a vowel or an /s/?
Will Pelsodian preserve part of the Latin case system (like Romanian) or merge it?
How will Pelsodian treat the consonant groups /ks kt gn/?
How will Pelsodian treat the initial consonant groups of /pl bl fl/?
How will Pelsodian treat nasal consonants and liquids?
How will Pelsodian treat historical Latin /k/ and /g/ before /i/? Before /e/? Before /a/?
Likewise, will the outcome of palatalizing Latin /k/ be /ts/ or /tʃ/? What about before /a/?
How will Pelsodian treat the labiovelar consonants? Before front vowels? Before /a/?
How will Pelsodian treat vowels in stressed environments? Unstressed environments?

Most progress will be covered on my blog and then updated here.

[mèþru]

Eastern paradigm
adopt vowel plurals
preserve case
palatalise k and g before /i e/ at the very least
the outcome of palatalising /k/ will be /tʃ/

[Ælfwine]

Eastern paradigm
adopt vowel plurals
preserve case

Preserving any remnant of Latin's case system is the least likely out of the above, especially because our location is more on the periphery of the Balkan Sprachbund than Romanian is (Hungarian has very few Balkan features iirc). Instead, it'd probably conform to the "central european" sprachbund (see here), for example, fronting /u/ to /i/, devoicing final consonants, possibly having phonemic vowel length and so on.

To give you an impression: VL quando -> quand -> quant -> quiant -> chant [t͡ʃaŋ] or something of the like.

Ultimately however, It would be up to whether Pannonia patterns with the Rhaeto-Romance languages like Friulian or the "Eastern Italic" languages like Dalmatian (and possibly Romanian). I'm leaning towards a more Rhaeto-Romance character myself. However, Atilla Gonda's own findings through the inscriptions seem to suggest that Pannonian Romance, at least near Aquincum, adopted an Eastern vowel paradigm, a Balkan case system and -i instead of -s in the plural, putting it much closer to the latter.

palatalise k and g before /i e/ at the very least
the outcome of palatalising /k/ will be /tʃ/

This is agreeable. According to the inscriptions, post-consonantal /j/ was more widespread in Pannonia than Dalmatia, and so palatalization could've happened had the Pannonians not met their demise. Pannonia likely maintained close connections with both Romania (which had palatalized its velars before both /e/ and /i/) and North Italy, and I'm sure the outcome would be the same as both regions (although I believe parts of North Italy had /ts/ too? Correct me if I am wrong, I'm not quite familiar with North Italian dialectology.)

[mèþru]

In that case you can have the loss of the case system happen later. I think that Common Slavic, the lingua franca of the Avar empire, would be a major influence.


The link you put doesn't work; it sends to you a non-existent part of this website.

[Ælfwine]

In that case you can have the loss of the case system happen later. I think that Common Slavic, the lingua franca of the Avar empire, would be a major influence.


The link you put doesn't work; it sends to you a non-existent part of this website.

That is very odd.

This one should be good: http://real.mtak.hu/31568/1/Gonda_Attila.pdf

While I found a Hungarian to translate the summary for me, I am still quite interested in the O~U timbre fusion. If I am reading it correctly, the unaccented (unstressed) U merged with O. But that is an odd change IMO. Wouldn't it likely be the other way around?

[Ælfwine]

Hello again, faithful readers!

While I have not had much time to directly work on the language, I have had a lot of time to think about it. So far much of the early Pelsodian language has been worked out, though there are a few things I can still change my mind upon. I invite you, reader, to read my justifications for each distinctive sound change and comment on what you think might be the most obvious decision here.

Vowels

One of the first decisions I had to make is whether Pelsodian would have a "Western Romance" vowel paradigm of /a ɛ e i ɔ o u/ or an "Eastern Romance" vowel paradigm of /a ɛ e i o u/. It seems plausible at first that Pelsodian might have had an "Eastern Romance" vowel paradigm, given it's further east than any other native Romance language except for Romanian. However, I began to doubt it when I started looking into Slavic languages such as Slovene and Kajkavian Croatian.

Both Slovene and Kajkavian Croatian have the seven vowel system plus schwa comparable to that of western Romance, while Štokavian to the south only has six vowels and schwa, similar to Romanian. I have reason to suspect that this isogloss goes back all the way to ancient times, when the Celts inhabited north of the Drave while Illyrian tribes populated the south. Because of this, I strongly believe that, had a romance language survived in Pannonia, it too would have adopted a "western" romance vowel system of /a ɛ e i ɔ o u/.

According to Atilla Gonda on his paper of , the vowel system of Aquileia in the later part of Roman times was evolving a "mixed" system where the unstressed velar vowels fell in line with the Balkan system while the stressed velar vowels well in line with the "western" system. While Friulian does not seem to preserve this system, Dalmatian in fact does have something identical to this, and so I believe Pelsodian would also adopt this system, if it had not adopted a "pure" Balkan vowel system. The only upset to this theory is that while we have evidence of this unique vowel system in Aquileia, we have no evidence for O ~ U mixture in Pannonian inscriptions, in fact we only have mergers of the U > O variety.

A > E

One of the characteristics unique to Pannonian Romance is the fronting of /a/ to what I assume to be /æ/, as an unusually high amount of inscriptions show the presence of the E grapheme in place of A. As you may know, French also fronts /a/ to /e/, and this seems to be a common feature of Gallo-Romance varieties. I have a theory that this change initially began in Pannonia, and spread west after the collapse of the province. This would make sense, as Norman French does not share this change, but many southern French dialects had. I have likewise included this A > E change too in Pelsodian, but only to /æ/ for now...

Consonants

One of the questions I have posited in the OP is how Pelsodian might treat the consonant groups /ks kt gn/. In Romanian we see the velar consonant before an alveolar consonant get conflated with a bilabial consonant, hence why we have noapte < Latin noctem. Friulian and Dalmatian on the other hand simply simplify the cluster /kt/ to /t/.

While it is tempting to share Romanian's unique "eastern" consonant shift, I believe the weight of the evidence from both Gonda's findings in the consonant system of the Danube provinces and Pelsodian's neighbors to the west suggests that Pelsodian would have also simplified the cluster /kt/ to /t/. Additionally, I surmise that similar clusters would also reduce: /ks/ to /s/, /ps/ to /s/, /pt/ to /t/, and /gn/ to /n/.

K(e,i)

Initially in Proto-Romance the velar consonants became the postalveolar consonants /t͡ʃ d͡ʒ/. In Western Romance we see the reflexes of a velar consonant before a front vowel merging with the affricate [ts]. In Eastern Romance, including Italian, the difference between [ts] and [tʃ] is maintained.

Posner in The Romance Languages, p. 113 makes the suggestion that the distinction between [ts] and [tʃ] was largely a distinction between whether [tʃ] was apical or laminal, conservative languages would favor a laminal articulation, being naturally more resistant to palatalization. Pelsodian, being isolated early on, would therefore fall into this "conservative camp." While I have made the point on my blog that palatalization is likely to have occurred in Pannonian Romance, I have made no comment on the exact quality of the palatal consonants. Given that Romanian to the east has followed Italian in preserving /tʃ/ instead of shifting it forward to a pure alveolar affricate, I think it is likely that Pelsodian would follow and keep these two phoneme distinct, especially as nearby Slavic languages also do.

G before a front vowel may initially had also developed a laminal articulation, becoming the post-alveolar affricate /dʒ/. Furthermore, given the rarity of the /dʒ/ phoneme in southern Central Europe, I believe it might later fall together with /j/ medially as it did in Spanish, while initially it might become a simple fricative.

The Labiovelars

Initially /kʷ/ and /gʷ/ in Latin, they came to become the clusters /kw/ and /gw/ in Proto Romance. In Romanian and Sardinian, they became bilabial consonants where they did not become velar ones. The conflation between bilabials and velar consonants again seems to be a uniquely Romanian trait, one not shared by any other language save Sardinian. I see no reason to not follow the Empire's trend here. Labiovelars would simplify to pure velar consonants, even before /a/, as opposed to becoming bilabials.

[Ælfwine]

I've decided to turn this thread into more of a general scratchpad for all of my work, instead of merely my romlang, which is stuck in development hell.

Right now, I have a couple other conlangs, all a posteriori's from Proto-Indo-European derived languages, as dealing with a posteriori's outside of IE is well outside my capability at the moment (although, I do have an itch for a Semitic language influenced by Georgian.) One (or two, since I split the project in half) of these languages is a Germanic language heavily influenced by the Celtic languages. For this thought experiment, I had considered the idea that the Vikings had began their raiding and pillaging several hundred years earlier than in our timeline, and what would happen had they conquered Ireland, as they did in our timeline, but earlier and left a more permanent mark.

The second variant of this language is more properly descended from Old Norse, spoken on the Isle of Man (the Calf of Man, to be specific). This project too has entered development hell until I can find some better resources on Manx. So for now, I focus on the first project, which is the conlang in "Ireland."

As per popular demand, this conlang will heavily feature Celtic traits like grammatically induced initial mutations and a broad/slender split in the consonant system.

Examples on how these grammatically induced initial mutations might work:

*bōk -> bóch [ˈboːx] "(a) book"
*hinƦ bōk -> an mbóch [ən ˈmˠoːx] "the book"
*in bōki [ĩn ˈboːki] -> í mbóich [iː ˈmˠoːç] “in a book”

etc...

For now, I use a Celtic orthography, although I am not married to this orthography and will probably adjust it so that it resembles North Germanic languages more. Also, I do want initial "ð" somewhere.

[Ælfwine]

Been thinking about the above conlang again...

In Proto-Norse, back vowels had fronted allophones before /i/. After final vowels lost out, these allophones were phonemized and Old Norse gained the phonemes /æ/, /ø/ and /y/ from /ɑ/, /o/ and /u/ respectively.

However, if I introduce an earlier broad-slender contrast from Old Irish, wouldn't front rounded vowels continue to still be allophonic? To me it seems like they would be reanalyzed as allophones of back vowels before slender consonants.

What do ya'll think?

[Nortaneous]

Front rounded vowels in Karaim (Turkic) were mostly shifted to back vowels with palatalization on the preceding consonant due to influence from Slavic. Turkic harmony was retained, however, as consonant palatalization harmony.

Front rounded vowels were retained in absolute initial position, and /e/ was retained in initial syllables (elsewhere > ʲa). /j/ doesn't trigger front harmony.

[Kuroda]

You're definitely a conlanger after my own heart, and I'm really pleased to see all your work & plans here! Unfortunately... I've become a big IE-dodger, took my Sanskrit degree and lit out for greener pastures. So I seriously doubt I'll understand, let alone fully appreciate, and far less likely have anything to say, about this.

But DON'T STOP! I love this kind of conlanging, a lot.

[Ælfwine]

Front rounded vowels in Karaim (Turkic) were mostly shifted to back vowels with palatalization on the preceding consonant due to influence from Slavic. Turkic harmony was retained, however, as consonant palatalization harmony.

Front rounded vowels were retained in absolute initial position, and /e/ was retained in initial syllables (elsewhere > ʲa). /j/ doesn't trigger front harmony.

That's quite interesting. I am actually tempted to include the reverse of this in my Slavlang, where palatal consonants cause fronting of /o/ to /ö/ before being lost, and a sort of Hungarian influenced type of vowel harmony spreading this change to all instances of /o/ in a word.

I'm quite tempted to redo the language from 900s onward, as I am having difficulty finding good resources on Proto-Norse in English in order to complete this project. I'm afraid it'll end up too similar to Trevor B.'s Írsc, though I had some of the same ideas before coming across his language on the old forum. The nouns greatly simplify due to the merger of the short vowels to /ə/ (this was also a sound change in Middle Irish iirc), but the slender/broad contrast keeps some distinctions alive. Notably, the verbs lose the distinction of person (like in Old Swedish), but not of number. My goal is to create a Nordic language halfway between insular Norse and Scandinavian.

You're definitely a conlanger after my own heart, and I'm really pleased to see all your work & plans here! Unfortunately... I've become a big IE-dodger, took my Sanskrit degree and lit out for greener pastures. So I seriously doubt I'll understand, let alone fully appreciate, and far less likely have anything to say, about this.

But DON'T STOP! I love this kind of conlanging, a lot.

Well thank you! I'm interested to know what part of my conlanging interests you?

I am personally influenced by conlangers like Martin Posthumus and Ray Brown. I try to treat my conlangs as if they were extensions of natural languages in the real world, influenced by real events

<hr></hr>

Back to Pelsodian for now...

p t c > b d g / V_V

A common occurance in Proto-Romance, and indeed Romance as a whole, was the lenition of original unvoiced stops in a intervocalic or postvocalic position. Whether I will include this change in Pelsodian depends on several factors, including its distance from the origin of this change (presumably Gaul), and the type of substrate that might influence the development of this sound law.

Intervocalic lenition in Spanish, French and North Italian languages may have been influenced by a common Celtic substrate. Evidence that further supports this theory includes evidence of intervocalic lenition in Gaulish and similar developments in the Insular Celtic languages (Schuchardt 1886). It is possible when city life collapsed in the Roman Empire the mass exodus of Latin speaking civilians to the countryside hastened this process, this aligns up quite nicely with the extinction of Gaulish in the 7th century, the same century when lenition became common. By the time the Frankish Empire gained possession of the Avar Khaganate, it may have spread as far east as Pannonia.

However, there was no collapse of city life in Pannonia as there was in Gallia, the Romans who lived their continued to eek out an existence regardless (or probably because) of the many barbarian invasions into the Pannonian countryside. While the Celts were present in Pannonia (and likewise influenced the A > E change), later evidence of them becomes scant.Finally, only the East Francien Empire controlled Pannonia after its split, and there may not have been any contact between Gaul and North Italy and Pannonia. I’ll make a final decision after reading the entirety of the Upper Danubian Provinces.

A possible compromise is partial lenition of intervocalic /p/ > /v/ alone, with /t/ and /k/ merging into their geminated equivalents.

ɪ ʊ > ∅ / _(C)#

In both Common Slavic and Hungarian, reduced vowels were often lost in unstressed positions. Unlike Common Slavic however, Hungarian only lost reduced vowels in word final position. This development is shared with Rhaeto-Romance, Dalmatian and Romanian, and likely would have also occurred in my Pannonian Romance language, likely with the Common Slavic and Hungarian language acting as an impetus. In both languages, this change is generally dated to the 9th and 10th century. The same process in Pelsodian would have likely occurred around the same time due to areal influence.

[Ælfwine]

Here's my attempt at deriving a vowel system from Proto-Germanic for the Irish Nord-lang:

https://imgur.com/a/YDXn2IA

This isn't a very consist graph, I was trying to detail phonemic outcomes but sometimes mixed them up with phonetic ones. What is supposed to happen is the former allophones of back vowels before front vowels are reanalyzed as allophones of back vowels in between two palatal consonants (with the front vowel allophone palatalizing the consonant in front of it).

To give you an idea of what I am trying to achieve, imagine this:

[CøCi] /CoCi/ --> [CʲøCʲ] /CʲoCʲ/

General palatalization occurs before final short vowels drop out in Proto-Norse. This happens after the period where short vowels are broken before these vowels. Later on, the outcomes of these broken vowels merge with the reflexes of back vowels in palatalizing environments.

Happening alongside this is lenition and the development of initial mutations.

Edit: Somehow this board doesn't support images. That is lame. Oh well, you can copypaste the link.

[Ælfwine]

I have to apologize for my inactivity: finals are coming up and I have to study for them, although I'd rather much conlang! Once I am done, I am going to probably work on my romlang. I am a bit worried its too close to Raetoromance like Friulian at the moment, instead of anything resembling an "Eastern" Romance language, though I suspect that's a given.

My reading goals for this winter is to finish Pannonia and Upper Moesia: A History of the Middle Danubian Provinces and The Rhaeto-Romance Languages by Haiman. I also have on my reading list The Dialects of Italy by Maiden and Perry. Hopefully, once I am done reading all three I'll have a fairly good idea on where to take this language.

[Nortaneous]

What is supposed to happen is the former allophones of back vowels before front vowels are reanalyzed as allophones of back vowels in between two palatal consonants (with the front vowel allophone palatalizing the consonant in front of it).

Something similar happened in Karaim, but only the preceding consonant was affected, and front rounded vowels (and the back unrounded vowel) were preserved in word-initial syllables.

Edit: Somehow this board doesn't support images. That is lame. Oh well, you can copypaste the link.

Imgur links aren't images. You have to link directly to the image URL (https://i.imgur.com/C4BRJmu.png), like this: