 |
| Piatra Galbenei, Apuseni Mountains, Bihor county, Romania 2005, Razvan Antonescu |
Well, now I can show what a solid year of ripping up and starting again twice has finally yielded.
Also: I underestimated how long it would take to write this all up, so I've missed my traditional conlang date of October 25th. Eleven days isn't that late.
Note: Much of this is now deprecated. Don't calculate any interplanetary flight manoeuvres with this.
Notes on Progress
The Getan Language is not derived from Proto-Albanian, though, it mostly serves as a guide for a dialect of the blanket-termed 'Paleo-Balkan' in the Carpathian area. No Dacian texts survived antiquity, but through comparing Albanian with Daco-Romanian substratum words we can determine some dialectical differences. I've amended some sound changes that were demonstrated last year, in particular *eH2 (ā) → /o/ which is now *eH2 → /a/ (preceded by /*o → a → ə/). This reflects PIE *méH2ǵh-ulos to D.Rom. mazăre rather than Alb. modhull, possibly via Dacian *mādzula. Albanian's shift to /o/ occurred in both Tosk and Gheg dialects after their situation in Albania, so this is marked out as a post-Dacian innovation.
The core sound changes of Getan have not changed much with the above important exception. However, the changes for the last five hundred years have been drastically rewritten, and it is these changes that wreak the most havoc with the nominal case system. The complete sound change documentation will remain unpublished until the language is more complete. That could be some time as I must state this is not a full grammar. It's not even close, not least because this sketch doesn't even touch on verbs.
Proposed History
Alternate history embraces the infinite unfolding possibilities of quantum theory. From some divergence point flows all manner of changes to what we know as history. To my mind, there are two scenarios for the continuation of Dacian as the above presented Getan language. The first is that Rome never crosses the Danube to seize Dacia, which requires the construction of an elaborate new history to explain why Trajan would look the other way. The second is that the Dacians simply survive by moving up into the Carpathian mountains away from Rome and the Avars later. This is a lot less hassle and with precedent. According to one theory the Aromanians are Romanised Thracians who retreated into the hills. History has plenty of irony, so let the imaginary un-romanised Daco-Thracians north of the Danube do what the real romanised Daco-Thracians south of the Danube did.
Like the Aromanians, the Getae would likely practice vertical transhumance. In researching how geography can shape ethnicity I read much on the valleys of Northern Italy and Switzerland - of note is the much overlooked fourth Swiss language Romansch. It's difficult to know how much to transpose from the history of the Alps to the Carpathians - especially whether it's credible that valleys in the vicinity could provide adequate refuge to Getan-speaking populations from successive invaders on the plains. Certainly on the back of a dramatic shift in environment would come pronounced semantic shift. Consider the existence of the Romanised Dacians after Rome pulled back over the Danube - the Latin word for 'city' became Daco-Romanian for 'fortress'.
Albanian exhibits many examples which lend credence to the theory Proto-Albanian migrated to its present position from different climes. With change of scenery comes change of life and it's believed the Proto-Albanians originally reared cattle (Orel, 2000), whereas life in the mountains would be better suited to herding goats and sheep, resulting in a large vocabulary for dairy production. Contact with Latin speakers would also result in the adoption of loanwords for items borrowed. Amongst the languages of Europe only Welsh, Basque, and Albanian avoided total assimilation while preserving sizeable quantities of Latin loanwords (eg. Welsh ffenest ← L. fenestra, Alb. qytet ← L. cīvitātem).
Given their flight into the rugged hinterland from a kingdom of relative size and power, the national psyche would doubtless be shaped by these circumstances. Might the Getae resent the Roman Empire when the rest of Europe was singing the praises of civilisation? We know the world wasn't appreciative of the end of Rome when Vandal ceased to be a proper noun (although the Germans later sought to recast their ancestors as fresh air in a stale house), and Iran is one of the few places on Earth that doesn't love Alexander the Great as he brought about the fall of the Achaemenid Empire.
Within the context of alternate history, the Getans take the place of the Ruthenians on the map of Europe. If they are stateless and an equivalent of our Romania exists then they are spread across Marmatia (Romania), Zakarpattia Oblast (Ukraine), and Subcarpathian Rus (Slovakia). If they do have a state, composed of those three regions, then in terms of size and regional history they would take the place of Moldova. When the proposed history is more concrete the Getans will probably have been in and out of Poland-Lithuania, the Austrian Empire, and the Russian Empire across the centuries.
The Flag
 |
| Flag of the Getae |
In the centre is a rampant wolf. The use of the wolf is due to the importance of the Draco in Dacian culture. There is an alternate version with a more accurate dragon-wolf chimera but it didn't look right.
Phonology
IPA notation follows.High vowels: /i (ɨ) u/
Middle: /e ə o/
Low vowels: /a/
The presence of central vowels is a hallmark of Balkan languages (Alb. ë, D.Rom ă, Bul. ъ).
Stress movement in the modern language phonemicised /ə/ which descended from short unstressed /a/ (eg. *ḱm̥tóm → θətá → [θˠətˠə ~ θˠɨtˠə]). It may raise to /ɨ/ under stress as unstressed /e a/ move to [ə] (depending on the preceding consonant, the unstressed schwa is written as «e» for palatalised or «a» for velarised). Unstressed /o/ becomes [u] and the high vowels are unaffected. Thus in unstressed syllables only [i ə u] are distinguished.
Stress is always on the penultimate syllable unless marked otherwise by an acute accent.
Nasals: /m nˠ nʲ ɲ/
Stops: /p b t̪ˠ tʲ d̪ˠ dʲ k kʲ g gʲ/
Affricates: /ʦ̪ ʣ̪ ʈʂ ɖʐ ʨ ʥ/
Fricatives: /f v θˠ θʲ ðˠ ðʲ s̪ˠ sʲ z̪ˠ zʲ ɕ ʑ x xʲ h/
Approximants: /w ɫ̪ l j/
Rhotics: /ɾˠ ɾʲ r/
Palatalisation is fully phonemic after syncope and apocope took place, though it doesn't stick to the labials (nor /ʦ ʣ ɫ r h/) which decompose into [Cj] unless word final when the palatal glide is fortified to [ʃ~ʒ] (eg. wolf DIR PL DEF [vu:pʃ], wool id. [vjiɫʒ], cf Hungarian /lopj/ [lopç]).
Velarised coronals are dentally articulated.
The palatal nasal may be more correctly identified as an alveolo-palatal (non-standard notation /ȵ/). It contrasts with the palatalised alveolar nasal (/ɲebjəh/ from *nébhos, /nʲebjəh/ from *snóyghwos).
Word final /h/ is (re-)fortified as [ɣ] (eg. *H3lígos → alléɣə → ɫjehə → [ɫjeɣ]).
The postalveolar space is crowded and primed for dialectical mergers. Though I toyed with some different set-ups, one of the desired outcomes was an opposition of three affricate series as is common in Eastern Europe (Polish and Serbo-Croatian /ʦ-ʈʂ-ʨ/, Hungarian /ʦ-ʧ-cç/) and also a three-way contrast between theta, sigma, and esh along the lines of Old Spanish /s̪-s-ʃ/.
Phonotactics aren't finalised yet, though probably falling between Italian and less extreme Slavic.
The Alphabet
26 letters:A a [a ˠə] B b [b] C c [k ʧ] D d [ð] Đ đ [θ] E e [e ʲə] F f [f] G g [g ʤ] H h [h : ɣ] I i [i j ʲ] J j [ʲ ь] L l [l] M m [m] N n [n] O o [o u] P p [p] Q q [kʲ] R r [ɾ] S s [z s] T t [t] U u [u w ˠ] V v [v] W w [ʷ:] X x [x] Y y [ə~ɨ] Z z [ʣ ʦ]
7 digraphs:
Ch ch [kʲ] Dd dd [d] Gh gh [gʲ] Ll ll [ɫ] Rr rr [r] Ss ss [s] Zz zz [ʦ]
5 trigraphs:
Chi chi [ɕ] Ghi ghi [ʑ] Nhi nhi [ɲ] Thi thi [ʨ] Dhi dhi [ʥ]
7 diacritical variants:
Ą ą [ã] Ę ę [ẽ] Į į [ĩ] Ô ô [õ] Ų ų [ũ] Ŷ ŷ [ə̃]
Among the various things the Free Dacians would have adopted in proximity with the Romans is their alphabet. Unlike Romania, the Getae remain within the Latin rites and therefore never write in Cyrilic - though I flirted with it because of the preponderance of letters for post-alveolar consonants, particularly the Macedonian variety for dze, dzhe, tshe, and dje (Ѕ Џ Ћ Ђ). There were plenty of compromises to be made to get the alphabet nailed down. Firstly, of course, Latin was never designed with anything other than the phonology of Latin in mind - there was much consulting of this guide (unfortunately now offline). A Slavic transcription on the Polish and Czech model was ditched since all the consonantal diacritics clashed with the vowel diacritics that a formerly intended pitch accent would have used. Only ogoneks for nasal vowels were held over from Polish (perhaps via a period of Commonwealth rule). In the end basic Latin and all the digraphs that it would entail gave it a centuries lived-in look that had been elaborately adapted. For a naturalistic language, avoiding the sterility of Esperanto is best practice.
Velarised and palatalised consonants take their articulation from historical succedent vowels and are spelt as such. Velarised consonants before front vowels are indicated «ue ui» (from *ō *ū). In coda they are unmarked. Palatalised consonants before back vowels are indicated «ia io iu» and in coda they are marked «Cj» (eg. *nékʷtey → /ɲatʲ/ «nhiatj»). The velars have their own set of rules.
The velar plosives and the alveolo-palatals are modelled on Eastern Romance. «C G» becoming /ʧ ʤ/ before front vowels is part of Getan's phonology and not simply aping Romance, and like Italian and Romanian they can occur before back vowels too «ci- gi-», though they can also occur word finally before a now-deleted front vowel «cj gj» (eg. Roger /rˠoʤəɾˠ/ «Rroger» → Reg /rˠeʤ/ «Rruegj», not much shorter to spell I acknowledge). Conversely, /kʲ gʲ/ before front vowels is «ch gh» and word finally before a historic front vowel they resemble Corsican «chj ghj». The letter Q is a holdover from earlier times and mostly only occurs in proper nouns (eg. Trad. «Qenhii» → Mod. «Chenhii»).
Old [kj gj] (from *Ké amongst other sources) became /ɕ ʑ/ (unlike Italian and Romanian [c ɟ]) and are spelt the same «chi- ghi-», never occurring word finally because the stressed vowel cannot be deleted. The exception is in borrowings, eg. Rachel /ɾˠeʨəɫ/ «Ruéthiall» → Rach /ɾˠeʨ/ «Ruethj». Before back vowels a diaeresis «chïa ghïa» dissimilates them from the orthographic iota in /kʲa gʲa/ «chia ghia». The use of heta to block palatalisation was extended to the alveo-palatal affricates and the palatal nasal in line with renaissance spellings like Carpathia [karpatja] - /ɲ ʨ ʥ/ «nhi- thi- dhi-» (eg. PIE *súHsos → /ʥux/ «dhiux», cf. ALB gjysh /ɟyʃ/, ROM ghiuj /ɟuʒ/).
In turn this impacted the representation of the interdental fricatives /θ ð/ which are the fairly conventional «th dh» in Albanian and in most other languages they're phonetic. «Đ» was taken on-board from Greek theta (as was «X» from chi). This isn't too drastic a leap since Gaulish borrowed tau gallicum from Eastern Greek for its dental affricate. The voiced interdental fricative takes a single delta, with the geminate representing the plosive in all places. The sibilants /s z/ and /ʦ ʣ/ follow Italian, voiced in between vowels (leaning tower of Pisa) and voiceless when geminate in the same position (pizza). These also stand in coda. Otherwise they're ambiguous unless next to a plosive whereupon the voicing assimilates (eg. *ḱérddhH1 → /θʲaːzdˠ/ «điahsdd»). The labial fricatives once followed the same pattern until «V» was invented.
The velarised lateral and the trill are geminate «ll rr» on account of primarily arising by intervocalic gemination. Their presence initially is due to apherisis (eg. *H3rēǵnós → /rõ:nˠə/ «rrôwna»).
«H» is used in internal codas for lengthened vowels (as in German) from *n *r and those that arose from *m *l use «W». Word initially before vowels and intervocalic «I U» function as the semivowels /j w/ (eg. *ōusa → /wex/ «uex») with exceptions. Word initially or after another vowel, [ji je] is a realisation of /ʲi ʲe/ and does not take a preceding iota (eg. I-stem SG PREP INDEF /-eje/ «ee»). «Y» was chosen for the schwa over «Ë» as it was more distinct and didn't require extending the alphabet.
The nasalised schwa takes a circumflex «Ŷ» rather than an ogonek since that doesn't compose well with the descender. This is also preferred for «Ô» as an ogonek in some fonts may approach Q in form, in addition to the fact that pre-composed O with ogonek isn't available in Latin-2 («Į Ų» are in Latin-4, I'm aware).
Spelling reformers may wish to make use of the plenty of other glyphs in Latin-2 - a few were included in the draft alphabets. I'd like to use them again since I like a neat system of monographs, yet what it produced was really a half-system. They just don't pair up: the velar lateral could be covered by barred-L but that would leave geminate R, C-cedilla covered «cj» but that left «gj», and so on. This is precisely why I dislike Albanian using «Ç» rather than «ch» as it disrupts the pattern of the other postalveolars (xh sh zh).
Nominal Declension
Compared to last year, the nominal declension is deceptively similar. There is one major change, and that is I felt obliged to make Getan a member of the Balkan Sprachbund morphologically in addition to phonologically. One of the most glaring areal features of the Balkans is postposed articles. My implementation of this was in the direct cases a proper effort, and in the oblique cases a quick kludge.
The definiteness of the direct cases is fully inspired by Glen Gordon's theory on Paleoglot to explain PIE's marked nominative. The definite animate nominative and accusative are *-s and *-m, the indefinite animate *-Ø (zero) and *-t (PIE ablative, cf Finnish partitive -ta from old ablative). The inanimates and obliques are unspecified for definiteness. Later on as part of the sprachbund, Getan derives oblique definiteness from a post-clitic demonstrative (*tom, etc). At a later stage of completion I may make the definiteness contrast shift toward a generic versus specific sense; whereby the indefinite a wolf, interchangeable with one wolf, becomes understood as this particular wolf. Expressing definiteness would then require fronting new referents with the verb in the passive voice (cf. English passive allowing the patient to become the subject - I saw THE WOLF → THE WOLF was seen by me).
For the most part all the stress patterns of Proto-Indo-European can be boiled down to two noun types in Getan (as in Albanian for the most part): consonantal and vocalic. The declensions are levelled in favour of weak (oblique) cases and the weak cases dictate whether the declension is consonantal or vocal. The simple rule is any in which stress falls on the last syllable are vocal. All others are consonantal. The exception is proterokinetic eH2/Ā-stems as the dative singular of 'woman' *gʷnéH2ei (trisyllabic ablaut of é-Ø-Ø ↔ Ø-é-Ø) is post-PIE *gʷnā́i (bisyllabic ablaut of é-a ↔ Ø-ā́), essentially becoming amphikinetic and ultimately oxytonic.
While I did have Latin ('classical studies') as a mandatory subject in early secondary school, I am not well versed in the declension formats of Latin or Greek so I will not follow those schemes and instead present them in the order they were finalised. Citation form is the SG DIR DEF case.
Class I (Consonant Stems)
Consonantal reflects PIE rhizotonics *-os *-om → -ə → -Ø (zero). This includes:
-rhizotonic common O-stems (*H1éḱwos 'horse' → /ʑep/ «ghiep»)
-rhizotonic neuter O-stems (*wérǵom 'work' → /vjaːðˠ/ «viahd»)
-rhizotonic common Ā-stems (*písdeH2 'vulva' → /pjedˠ/ «pedd»)
-rhizotonic neuter Ā-stems (*H2wĺ̥H1neH2 'wool' → /vjiɫ/ «vill»)
-acrostatic athematic root nouns (*nókʷts 'night' → /ɲatˠ/ «nhiat»)
-acrostatic lengthened athematic root nouns (*mḗH1ns 'moon' → /monˠ/ «mon»)
-proterokinetic R-stems (*swesōr 'sister' → /vjehər/ «veharr»)
-proterokinetic neuter N-stems (*séH1mn 'seed' → /hjẽvjə̃ɾˠ/ «hęviąr»)
-acrostatic neuter N-stems (*H1nómn 'name' → /ɾʲem/ «rem»)
-acrostatic neuter S-stems (*ḱléwos 'fame' → /sʲavjəh/ «siaveh»)
PIE *wĺ̥kʷos 'wolf' [rhizotonic O-stem] (wúlpas → vúlpə → /vu:p/)
Direct Indefinite: sg. vuwpull, pl. vuwp
Direct Definite: sg. vuwp, pl. vuwpj
Oblique Indefinite: sg. vuwps, pl. vuwpux
Oblique Definite: sg. vuwpsat, pl. vuwpuxt
The rhizotonic direct indefinite singular is extended by a diminutive -ull (from *-ul-, cf Latin suffix) which saves it from collapsing into the plural and definite singular (PIE ablative *-t/d is lost early on as in Proto-Balto-Slavic).
The Oblique case is a merger of the Genitive and Prepositional in this stem class only, which is in line with one of the key Balkan features of genitive-dative syncretism.
The PL PREP metathesised -əxu (< *oisu) to -ux, losing the schwa to apocope.
Class II (A-Stems)
Vocal reflects PIE oxtytonics *-ós *-óm → -á → -ə. This includes:
-oxytonic O-stems (*yugóm 'yoke' → /ʥuhˠə/ «dhiuha»)
-oxytonic Ā-stems (*dikéH2 'goat' → /ʣjekə/ «zeca»)
-hysterokinetic Ā-stems (*dnǵhwéH2s 'tongue' → /dʲeðˠə/ «ddeda»)
-proterokinetic Ā-stems (*gʷēn 'woman' → /gũɾˠə/ «gųra»)
-amphikinetic athematic root nouns (*pṓds 'foot' → /pjedˠə/ «pedda»)
-hysterokinetic athematics (*ḱérd 'heart' → /θʲeːdˠə/ «đehdda»)
-amphikinetic N-stems (*H2éḱmō 'stone' → /kɨmə/ «cyma»)
-hysterokinetic N-stems (*ḱwṓ 'dog' → /sˠũɾˠə/ «sųra»)
-amphikinetic M-stems (*dhéǵhōm 'earth' → /zdˠɨmə/ «sddyma»)
-acrostatic R-stems (*méH2tēr 'mother' → /matɾˠə/ «matra»)
-hysterokinetic R-stems (*pH2tḗr 'father' → /pɨtɾˠə/ «pytra»)
-hysterokinetic TER agents (*H2stḗr 'star' → /sˠɨrə/ «syrra»)
-amphikinetic TOR agents (*ǵénH1tōr 'parent' → /ðnʲetɾˠə/ «dnetra»)
PIE *nisdós 'nest' [oxytonic O-stem] (nixdás → nexdá → /nʲedˠə/)
Direct Indefinite: sg. neddo, pl. nedda
Direct Definite: sg. nedda, pl. neddi
Genitive Indefinite: sg. neddass, pl. neddaue
Genitive Definite: sg. neddast, pl. neddauet
Prepositional Indefinite: sg. nedde, pl. neddax
Prepositional Definite: sg. neddet, pl. neddaxt
Pretty much the majority of PIE nouns end up in the vocalic category.
The vocalic class has an expanded PL GEN -a+ue by analogy with the full grade ending of this paradigm versus the zero grade of the rhizotonics.
Class IIIa (I-Stems) directly continue the common PIE I-stems.
PIE *méntis 'thought' [proterokinetic feminine -tis] (metéys → metéi → /mjetʲe/)
Direct Indefinite: sg. mete, pl. metj
Direct Definite: sg. metj, pl. metęrj
Genitive Indefinite: sg. metess, pl. metiue
Genitive Definite: sg. metest, pl. metiuet
Prepositional Indefinite: sg. metee, pl. metex
Prepositional Definite: sg. meteet, pl. metext
Class IIIb (I-Stems) meanwhile continue the neuter PIE I-stems. These have irregular direct definite forms. The nominative plural exhibits vowel breaking followed by metathesis of liquid plus glide sequences. The plural DIR INDEF unexpectedly descends from INST *mríH1 (probably indifferent to number, cf. English 'by hand').
PIE *móri 'sea' [proterokinetic neuter -i] (NOM PL máriH2 → marjā ~ metath. mairra → /mjer/)
Direct Indefinite: sg. mre, pl. mri
Direct Definite: sg. marr, pl. merr
Genitive Indefinite: sg. mress, pl. mreue
Genitive Definite: sg. mrest, pl. mreuet
Prepositional Indefinite: sg. mree, pl. mrex
Prepositional Definite: sg. mreet, pl. mrext
Class IVa (O-Stems) continue the common PIE U-stems.
PIE *ǵéwstus 'taste' [proterokinetic masculine -tus] (zustjáus → zutô → /zˠutˠu/)
Direct Indefinite: sg. suto, pl. sut
Direct Definite: sg. sut, pl. sutųrj
Genitive Indefinite: sg. sutoss, pl. sutoue
Genitive Definite: sg. sutost, pl. sutouet
Prepositional Indefinite: sg. sutaui, pl. sutux
Prepositional Definite: sg. sutauit, pl. sutuxt
Class IVb (O-Stems) continue the neuter U-stems. As with IIIb, these have irregular direct definite forms. The plural DIR INDEF unexpectedly descends from INST *drúH1.
PIE *dóru 'tree' [proterokinetic neuter -u] (NOM PL dáruH2 → darwā ~ metath. daurra → /dˠor/)
Direct Indefinite: sg. ddro, pl. ddrui
Direct Definite: sg. ddarr, pl. ddorr
Genitive Indefinite: sg. ddross, pl. ddroue
Genitive Definite: sg. ddrost, pl. ddrouet
Prepositional Indefinite: sg. ddraui, pl. ddrux
Prepositional Definite: sg. ddrauit, pl. ddruxt
Class V (Nasalised Stems) are between consonantal and vocal, preserving the liquid stem of the strong cases of the PIE heteroclitic nouns. In the weak cases the nasal stop sometimes pops up (nasal and tap sequences) and other times is lost (nasal vowel).
PIE *wódr 'water' [proterokinetic -R/N] (wáder → vaðer → /vaðʲər/)
Direct Indefinite: sg. vedę, pl. vedęri
Direct Definite: sg. vaderr, pl. vaduerr
Genitive Indefinite: sg. vedęss, pl. vedęue
Genitive Definite: sg. vedęst, pl. vedęuet
Prepositional Indefinite: sg. vedęre, pl. vedęx
Prepositional Definite: sg. vedęret, pl. vedęxt
PIE *sóH2wl 'sun' [sole proterokinetic -L/N] (háwul → hávul → /havuɫ/)
Direct Indefinite: sg. hyvą, pl. hyvąri
Direct Definite: sg. havull, pl. havuell
Genitive Indefinite: sg. hyvąss, pl. hyvąue
Genitive Definite: sg. hyvąst, pl. hyvąuet
Prepositional Indefinite: sg. hyvąre, pl. hyvąx
Prepositional Definite: sg. hyvąret, pl. hyvąxt
Pronouns
Pronouns are in the final stages of completion, though I won't hold up this partial grammar any longer. At the time of writing the basic DIRECT and OBLIQUE forms are:
DIRECT
1st person: sg. Ghied, pl. Ve
2nd person familiar: sg. To, pl. Dhiu
2nd person formal: sg. Dhiu, pl. Dhiu
3rd person masculine/neuter: sg. I, pl. De
3rd person feminine: sg. Ia, pl. De
OBLIQUE
1st person: sg. Me, pl. Nam
2nd person familiar: sg. Tue, pl. Vam
2nd person formal: sg. Vam, pl. Vam
3rd person masculine/neuter: sg. E, pl. In
3rd person feminine: sg. Ia, pl. In
Reflexive: He
The first and second person plurals fell apart and were replaced by the dual number which survived mostly unscathed. This is in stark contrast to their rapid downfall in the nominal declension.
The End?
Not by a long shot. Noun Phrases aren't even complete since I haven't worked on adjectives and I've still to work out how grammatical gender has been effected or whether it even exists anymore in Getan. Beyond that is the immense mountain of the verbs.
Written 26th, 27th, 28th of October, 3rd, 4th, and 5th of November
Updated 10th of November
[3124]
No comments:
Post a Comment