Started sketching out some ideas today and they turned out pretty well:
Teavislian /tə'vaɪ̯.li.ən/ is the language of Teavislia /tə'vaɪ̯.li.ə/, the English-language name of Taewiulie /tɛ'vil.jə/, an archipelagic country in the Atlantic ocean, consisting of three main islands, about 550 miles west-northwest of Galicia. Its official name is the Kingdom of Teavislia, or Reonj Taewilor /ʁeɲ 'tɛ.vi.loʁ/. It is notable as a holdout of a pre-Indo-European language in Europe, though Basque is, of course, the only continental European language to hold such a distinction.
Though a language isolate, Teavislian has had contact with European languages since the beginning of its recorded history. It was first written in a version of the Greek alphabet, mainly found on gravestones and stelae commemorating chiefly/princely accomplishments and offerings to deities. It began to be written in the Latin alphabet after the country's conversion to Christianity in the 400s AD.
As a case study in the phonological history of the language, we will look at a relatively early Latin loanword into Teavislian, saeord ("priest"), from the Latin sacerdos. It was loaned into the language sometime before 200-400 AD, likely before the arrival of Christianity, and before a major sound change known as Roedrig's Law.
The earliest form in which sacerdos appears in Teavislian is saherdu (likely pronounced /saxerdu/). This form replaces the Latin -os with the Teavislian nominative singular suffix -u. The Latin /k/ appears as h because of Roedrig's Law, in which Ancient Teavislian /k/ lenites to /x/.
Eventually, historical /x/ further lenited to /h/, and then disappeared. But before its disappearance, a series of vowel mutations occurred, which eventually had significant consequences for the morphonological structure of Teavislian. The first of the changes was a-mutation, in which any /i/ or /u/ that preceded a syllable with /a/ was lowered to /e/ and /o/, respectively. This did not affect saherdu. This was followed by "high mutation", in which any non-back vowel followed by a syllable with /u/ was diphthongized, and likewise for non-front vowels and /i/. For the purposes of this sound change, /a/ is a central vowel.
The outcomes were:
I-mutation
/a/ > /ae̯/
/o/ > /oe̯/
/u/ > /ui̯/
U-mutation:
/a/ > /ao̯/
/e/ > /eo̯/
/i/ > /iu̯/
The changes to the nominal paradigm were notable:
Pre-mutations:
Sing. nom.: saherdu
Pl. nom: saherdi
Sing. acc.: saherdus
Pl. acc.: saherdis
Sing. gen: saherdula
Pl. gen.: saherdila
Post-mutations:
Sing. nom.: saheordu
Pl. nom: saherdi
Sing. acc.: saheordus
Pl. acc.: saherdis
Sing. gen: saherdola
Pl. gen.: saherdela
Both the effects of vowel mutation and the loss of /h/ were reflected in Teavisilian orthography, with the two changes appearing in two separate "waves" from the scriptorium of the principal monastery in Teavislia, as new generations of monks opted to spell words differently to reflect their own pronunciation.
With the total disappearance of /h/, certainly no later than 1000 AD, many instances of vowel hiatus were created. By around 1200, both these hiatuses and the historical diphthongs were being smoothed into monophthongs. However, this change was not reflected in orthography. The basis for modern Teavislian orthography was established around this time period.
Not all outcomes will be listed here, but the outcomes for saheordu are shown below. Also shown are the modern forms of the case and number suffixes, which were reached no later than 1450. This is also reflected in the orthography. (Also the palatalization of /t/ and /d/ before mid and high front vowels, which I initially forgot about.)
Sing. nom.: saeord /sœrd/
Pl. nom: saerdj /sɛrd͜ʒ/
Sing. acc.: saeorde /sœrdə/
Pl. acc.: saerdje /sɛrd͜ʒə/
Sing. gen: saerdol /sɛrdol/
Pl. gen.: saerdjel /sɛrd͜ʒel/
Around 1550-1600, a change known as the front rhotic merger occurred. Before /r/, /ɛ/ merged into /a/, and /e/ merged into /i/. However, this was counterfed by the unrounding of front rounded vowels between 1650 and 1750. This produced the forms:
Sing. nom.: saeord /sɛrd/
Pl. nom: saerdj /sard͜ʒ/
Sing. acc.: saeorde /sɛrdə/
Pl. acc.: saerdje /sard͜ʒə/
Sing. gen: saerdol /sardol/
Pl. gen.: saerdjel /sard͜ʒel/
During the 1800s, the uvular R spread from continental Europe to Teavislia, becoming the standard pronunciation of /r/ in the prestige dialect of Teavislian, leaving us with the modern word /sɛʁd/ for saeord.
Well, this took longer than I thought. I have more stuff about historical phonology and noun morphology I'll post later.
Musings on Teavislian
-
Porphyrogenitos
- Lebom
- Posts: 168
- Joined: Thu Dec 22, 2011 1:13 pm
- Location: Ohio
Musings on Teavislian
Last edited by Porphyrogenitos on Thu Jun 18, 2015 7:25 pm, edited 1 time in total.
-
Porphyrogenitos
- Lebom
- Posts: 168
- Joined: Thu Dec 22, 2011 1:13 pm
- Location: Ohio
Re: Musings on Teavislian
Okay, so a bit more with Latin loanwords.
After Roedrig's Law, there were two sources of new /k/. One was /q/ - Ancient Teavislian had /q/, and it fronted to /k/ after original /k/ was lost. The second source was loanwords.
Let's look at this word, kluku ("chick, young animal") - modern Teavislian kluk ("child, kid"). It is loaned from Latin glocīre, an onomatopoeic term referring to a chicken's clucking. For unclear reasons the Latin /o/ became Teavislian /u/ - though this transformation has been attested in other loans before velars.
The oddest thing here is actually how Latin /g/ came to be Teavislian /k/ - why is that? The answer is that Ancient Teavislian (henceforth AT) didn't have /g/. It gained /g/ from later loanwords, but during this time period foreign /g/ simply became /k/.
If we look at the consonant inventory of AT both before and after Roedrig's Law, it leaves a slightly odd, though not terribly remarkable gap:
/m n/
/p t k (q)/
/b d/
/s (h)/
/w/
However, it is possible to prove through internal reconstruction that Proto-Teavislian (henceforth PT) actually had /g/. AT /w/ - modern /v/, though spelled <w> - is its reflex.
We can take a look at this Modern Teavislian (henceforth MT) word: naowaor /nɔvɔʁ/, said to encapsulate a Teavislian national value. It's kind of one of those "untranslatable words" you see in lists online, and basically means "speaking the truth, but in a simple manner; owning up to the facts without trying to dress them up".
Its AT form, nahuwaru starts to reveal its etymology. In Modern Teavislian, naou /no/ (pl.: nai /ne/) means "correct" or "true". (Or "veracity" or "truth" - like in Spanish, adjectives in Teavislian are not truly distinct from nouns as a word class.) In AT, this word was nahu, and meant "truth" more in the sense of "truth that is told, especially as in a prophecy".
The second component is what's revealing. In MT, we have aor /ɔʁ/ ("language") - in AT, haru ("speech, rhetoric"). So nahuwaru, to put it in modern terms, essentially meant "real talk". But why do we have nahuwaru and not nahuharu? Well, some very old compound words in Teavislian - dating back past AT even to PT - exhibit a form of sandhi very similar to Japanese rendaku. It's not productive past PT - new compound words in AT don't exhibit it - but we know it's there. We know that nahu and haru must have been *naku and *karu in Proto-Teavislian and early AT. So apparently when voiced, PT /*k/ became /*w/ - though it would certainly make more sense if it became /g/. In fact, that is likely how it was initially voiced, but by the time of AT it had lenited to /w/.
This phenomenon is even seen in the very name of Teavislia:
*ta ("three") + *gilu ("land, territory, allotment")
*tagilu ("Teavislia")
Tawilu
Taewiulu
Taewiulu + Latin -ia
Taewiulie
After Roedrig's Law, there were two sources of new /k/. One was /q/ - Ancient Teavislian had /q/, and it fronted to /k/ after original /k/ was lost. The second source was loanwords.
Let's look at this word, kluku ("chick, young animal") - modern Teavislian kluk ("child, kid"). It is loaned from Latin glocīre, an onomatopoeic term referring to a chicken's clucking. For unclear reasons the Latin /o/ became Teavislian /u/ - though this transformation has been attested in other loans before velars.
The oddest thing here is actually how Latin /g/ came to be Teavislian /k/ - why is that? The answer is that Ancient Teavislian (henceforth AT) didn't have /g/. It gained /g/ from later loanwords, but during this time period foreign /g/ simply became /k/.
If we look at the consonant inventory of AT both before and after Roedrig's Law, it leaves a slightly odd, though not terribly remarkable gap:
/m n/
/p t k (q)/
/b d/
/s (h)/
/w/
However, it is possible to prove through internal reconstruction that Proto-Teavislian (henceforth PT) actually had /g/. AT /w/ - modern /v/, though spelled <w> - is its reflex.
We can take a look at this Modern Teavislian (henceforth MT) word: naowaor /nɔvɔʁ/, said to encapsulate a Teavislian national value. It's kind of one of those "untranslatable words" you see in lists online, and basically means "speaking the truth, but in a simple manner; owning up to the facts without trying to dress them up".
Its AT form, nahuwaru starts to reveal its etymology. In Modern Teavislian, naou /no/ (pl.: nai /ne/) means "correct" or "true". (Or "veracity" or "truth" - like in Spanish, adjectives in Teavislian are not truly distinct from nouns as a word class.) In AT, this word was nahu, and meant "truth" more in the sense of "truth that is told, especially as in a prophecy".
The second component is what's revealing. In MT, we have aor /ɔʁ/ ("language") - in AT, haru ("speech, rhetoric"). So nahuwaru, to put it in modern terms, essentially meant "real talk". But why do we have nahuwaru and not nahuharu? Well, some very old compound words in Teavislian - dating back past AT even to PT - exhibit a form of sandhi very similar to Japanese rendaku. It's not productive past PT - new compound words in AT don't exhibit it - but we know it's there. We know that nahu and haru must have been *naku and *karu in Proto-Teavislian and early AT. So apparently when voiced, PT /*k/ became /*w/ - though it would certainly make more sense if it became /g/. In fact, that is likely how it was initially voiced, but by the time of AT it had lenited to /w/.
This phenomenon is even seen in the very name of Teavislia:
*ta ("three") + *gilu ("land, territory, allotment")
*tagilu ("Teavislia")
Tawilu
Taewiulu
Taewiulu + Latin -ia
Taewiulie
-
Porphyrogenitos
- Lebom
- Posts: 168
- Joined: Thu Dec 22, 2011 1:13 pm
- Location: Ohio
Re: Musings on Teavislian
More on Diphthongs
So, I realized that having the diphthongs (and resultant triphthongs) wait until after /h/ disappeared to monophthongize would result in catastrophic damage to the sound system, so to speak. So I organized the progression of sound changes into stages a bit more.
First Diphthongization: c.500 AD
E.g. High Mutation, U- and I-Mutation
- After A-Mutation
- Before loss of /h/
I-mutation
/a/ > /ae̯/
/o/ > /oe̯/
/u/ > /ui̯/
U-mutation:
/a/ > /ao̯/
/e/ > /eo̯/
/i/ > /iu̯/
First Smoothing: c. 700 AD
- After High Mutation
- Before loss of /h/
/ae̯/ > /ɛ/
/oe̯/ > /ø/
/ui̯/ > /y/
/ao̯/ > /ɔ/
/eo̯/ > /ø/
/iu̯/ > /y/
This resulted in a relatively simple nominal paradigm. The penultimate vowel in each noun recieves a different inflection depending on case and number. Nominative and accusative singular receive the u-mutation, nominative and accusative plural receive the I-mutation, and the genitive singular and plural receive the unmutated form (due to A-mutation triggered by -la).
Nouns (with a few exceptions) fall into one of these five inflectional patterns:
A-stems: /ɔ/ /ɛ/ /a/
E-stems: /ø/ /e/ /e/
I-stems: /y/ /i/ /i/
O-stems: /o/ /ø/ /o/
U-stems: /u/ /y/ /u/
At this stage, for example, saheordu was an E-stem:
Sing. nom.:saheordu /sahørdu/
Pl. nom: saherdi
Sing. acc.: saheordus /sahørdu/
Pl. acc.: saherdis
Sing. gen: saherdola
Pl. gen.: saherdela
Second Diphthongization: c. 1100
- After loss of /h/
In this stage, the loss of /h/ produces a large number of new diphthongs as vowels enter into hiatus and then diphthongize. All resulting diphthongs are falling.
This has two main effects on nouns. First, it creates new diphthongs word-internally, such as nahowaoru becoming naowaoru /nao̯vɔru/.
Secondly, it affects the inflection of nouns when it puts the penultimate vowel into a new diphthong. This is the case with saheordu > saeordu /saø̯rdu/ and naohu > naou /nɔu̯/.
This results in the creation of a large and diverse class of nouns collectively known as H-stems. These are broken down into initial H-stems (words like saeordu, in which the falling component of the diphthong inflects) and final H-stems (words like naou, in which both the syllabic and falling components of the diphthong inflect).
Luckily (for me, and perhaps for the speakers of Teavislian), no triphthongs or any larger vowel sequences were created as a result of the loss of /h/. This is because of a sound law - a surface filter - active from PT through AT all the way up to Roedrig's Law, in which it was impermissible to have two syllables in a row that began with unvoiced consonants of the same place of articulation. Thus, *rukiku was impermissible, but *rukigu (eventual rukiwu) was. This surface filter ceased to be productive after the lenition of /k/ to /h/. Note that PT was strictly CV, and so was AT, until the entrance of loanwords.
Second Smoothing: c. 1450-1600 (I'm moving this up in time a bit)
Even with my attempt to mete out the processes of vowel changes gradually, this was still somewhat of a disaster - at least for the H-stems. I haven't decided all the outcomes yet, but it mainly assigned the new monophthongs to preexisting vowels, though it did result in the innovation of /œ/.
The unrounding of the front rounded vowels definitely also wreaked more havoc. And I haven't decided at all about the verbal morphology yet, but I imagine it was pretty significantly impacted by all this.
So, I realized that having the diphthongs (and resultant triphthongs) wait until after /h/ disappeared to monophthongize would result in catastrophic damage to the sound system, so to speak. So I organized the progression of sound changes into stages a bit more.
First Diphthongization: c.500 AD
E.g. High Mutation, U- and I-Mutation
- After A-Mutation
- Before loss of /h/
I-mutation
/a/ > /ae̯/
/o/ > /oe̯/
/u/ > /ui̯/
U-mutation:
/a/ > /ao̯/
/e/ > /eo̯/
/i/ > /iu̯/
First Smoothing: c. 700 AD
- After High Mutation
- Before loss of /h/
/ae̯/ > /ɛ/
/oe̯/ > /ø/
/ui̯/ > /y/
/ao̯/ > /ɔ/
/eo̯/ > /ø/
/iu̯/ > /y/
This resulted in a relatively simple nominal paradigm. The penultimate vowel in each noun recieves a different inflection depending on case and number. Nominative and accusative singular receive the u-mutation, nominative and accusative plural receive the I-mutation, and the genitive singular and plural receive the unmutated form (due to A-mutation triggered by -la).
Nouns (with a few exceptions) fall into one of these five inflectional patterns:
A-stems: /ɔ/ /ɛ/ /a/
E-stems: /ø/ /e/ /e/
I-stems: /y/ /i/ /i/
O-stems: /o/ /ø/ /o/
U-stems: /u/ /y/ /u/
At this stage, for example, saheordu was an E-stem:
Sing. nom.:saheordu /sahørdu/
Pl. nom: saherdi
Sing. acc.: saheordus /sahørdu/
Pl. acc.: saherdis
Sing. gen: saherdola
Pl. gen.: saherdela
Second Diphthongization: c. 1100
- After loss of /h/
In this stage, the loss of /h/ produces a large number of new diphthongs as vowels enter into hiatus and then diphthongize. All resulting diphthongs are falling.
This has two main effects on nouns. First, it creates new diphthongs word-internally, such as nahowaoru becoming naowaoru /nao̯vɔru/.
Secondly, it affects the inflection of nouns when it puts the penultimate vowel into a new diphthong. This is the case with saheordu > saeordu /saø̯rdu/ and naohu > naou /nɔu̯/.
This results in the creation of a large and diverse class of nouns collectively known as H-stems. These are broken down into initial H-stems (words like saeordu, in which the falling component of the diphthong inflects) and final H-stems (words like naou, in which both the syllabic and falling components of the diphthong inflect).
Luckily (for me, and perhaps for the speakers of Teavislian), no triphthongs or any larger vowel sequences were created as a result of the loss of /h/. This is because of a sound law - a surface filter - active from PT through AT all the way up to Roedrig's Law, in which it was impermissible to have two syllables in a row that began with unvoiced consonants of the same place of articulation. Thus, *rukiku was impermissible, but *rukigu (eventual rukiwu) was. This surface filter ceased to be productive after the lenition of /k/ to /h/. Note that PT was strictly CV, and so was AT, until the entrance of loanwords.
Second Smoothing: c. 1450-1600 (I'm moving this up in time a bit)
Even with my attempt to mete out the processes of vowel changes gradually, this was still somewhat of a disaster - at least for the H-stems. I haven't decided all the outcomes yet, but it mainly assigned the new monophthongs to preexisting vowels, though it did result in the innovation of /œ/.
The unrounding of the front rounded vowels definitely also wreaked more havoc. And I haven't decided at all about the verbal morphology yet, but I imagine it was pretty significantly impacted by all this.