 |
| Green Eagle, Quinn Dombrowski, 2010 |
Before getting into the meat, let me mention some of the phonological differences. I used to be fascinated by maps and atlases in school, so I am in love with this map of Europe circa 125AD. Looking at Dacia we can see the wider area was populated by a mix of Iranic (Iazyges, Roxolani, Sarmatians), Germanic (Bastarnae), Celtic (Cotini), and Thraco-Dacian tribes (Daci, Carpi?). Some of these fit in nicely with sound changes in Albanian, such as Iranic /s → h/ between vowels. Other features of Albanian phonology prompted me to consider a Celtic influence at the time they were pushing into the Balkans - such as the gemmination of intervocalic liquids and the reduction of intervocalic /n/ to a tap. That almost resembled the fortis and lenis resonants of Old Irish and so I extended it by assuming /n/ became a nasalised tap /ɾ̃/ and having /m/ parallel it by shifting to a nasalised labial glide /w̃/. These weren't features of contemporary Continental Celtic, but I've gone with it under the rule of cool. Similarly, I've followed Proto-Celtic in debuccalising /p → h/ (also seen in Armenian) and then created a new /p/ from /kʷ/ which is seen in some Celtic languages but also justifiable as occuring in Greek, Osco-Umbrian and also Romanian (Lat. equa → Rom. iapă / PIE *H1éḱuos → Get. jépa, cf. Greek hippos, Gaul. epos).
Curiously enough, only a few months ago I came across a link to an article claiming evidence that Albanian is related to the Celtic languages. Some of the more orthodox research into Albanian claims early influence/origin from/in Balto-Slavic - eg. /o → a/, RUKI, Winter's Law, syllabic resonants gaining prothetic high vowels, satemisation. In the case of Albanian's interesting satemisation of the PIE palatovelars to dental fricatives in some circumstances, I've posited they became alveo-palatals affricates which deaffricated before liquids and high vowels. These then became alveolar affricates and fricatives as velars palatalised, and then shifted into dental fricatives as new alveolar affricates arose from palatalised alveolars (/ḱ → ʨ → ʦ → θ/). Dental fricatives are usually the hallmark of an English speaker's first attempt at a conlang or someone aping flowery Tolkein-esque elf languages (although Quenya doesn't actually have them). More modern changes from the middle ages onwards are influenced by Slavic just as Romanian, whereby Getic acquires palatalised (and velarised) consonants as well as vowel reduction and another wave of palatalisation (roughly along the lines of the Polish fourth). The palatal and velar articulations of consonants again recalls Old Irish. I'm happy with the sound changes as they stand and have let them be in the few months I've turned my attention to the grammar, yet there might come a time I think it looks too Celtic to be plausible and tone it down a bit. As a consequence, the Cyrillic script has been ditched for Latin which admittedly does create a lot of the similarities to Irish (<mh> for /w̃/ in particular) - it's just that having digraphs in Cyrillic isn't aesthetically pleasing to my senses.
And now to the moment you've been waiting for. If you ever had Latin at school (probably very rare these days in state schools), get ready for some riveting declension tables. In the two months I've been looking at Albanian nominal morphology and cribbing notes from the ZBB members who authored Wenetic and Lusitanic, I have thus far produced two declensions to my satisfaction. It's no coincidence the O and EH2-stems are the best documented (accessible for free, I might add) and easiest by virtue of having only two stress patterns (rhizotonic and oxytonic), hence those are the two I have to present. The semi-vocalic stems (I and U) are the next most regular declension with only the proterokinetic stress patterns, yet I'm still not satisfied with those. Especially since I'm not too sure of the amount of analogical levelling that should occur by influence of the two thematic stems. I haven't proceeded into the consonant stems because the R-stems have been a major sticking point. There's only five nouns (all family terms: mother, father, brother, sister, daughter) but there's acrostatic, proterokinetic, and hysterokinetic stress patterns which is hampering the regularisation of the endings. Pronouns are also a work in progress. Nonetheless, there are plenty of thematic stem nouns which I can now decline.
PIE Acrostatic O-stem
Wolf, male, rhizotonic
PIE *wĺ̥kʷos (wulkʷas → wulpah → metath. lūpā > [ɫupˠ])
Direct (Nominative, Accusative, Vocative): singular lupa, plural lupe
Genitive: sing. lups, pl. lupaŭ
Ablative (also Instrumental): sing. lup, pl. lupaŭ
Dative (also Locative): sing. lupi, pl. lupach
Nest, neuter, oxytonic
PIE *nisdós (nixdas > nexdah > niedā > [nʲedˠa])
Direct: sing. nieda, pl. niede
Genitive: sing. niedas, pl. niedaŭ
Ablative: sing. niedo, pl. niedaŭ
Dative: sing. niedi, pl. niedach
PIE Acrostatic Ā-stem
Wool, neuter, rhizotonic
PIE *H2wĺ̥H1neH2 (awūlnā > awillo > βillo > [ɥiɫ])
Direct: sing. yill, pl. yillu
Genitive: sing. yills, pl. yilloŭ
Ablative: sing. yilla, pl. yilloŭ
Dative: sing. yille, pl. yilloch
Grain, female, oxytonic
PIE *dhoHnéH2 (dōnā > dweɾ̃o > dˠøɾ̃ˠo > [dˠẽɾˠo])
Direct: sing. dęro, pl. dęre
Genitive: sing. dęros, pl. dęroŭ
Ablative: sing. dęra, pl. dęroŭ
Dative: sing. dęri, pl. dęroch
Observations. As in Albanian, the nominative and accusative cases collapse into a single direct case which necessitates SVO word order. The majority of the oblique cases have been lost and largely amalgamated - ablative and instrumental; and dative and locative were already in the process of merging early on and the endings were wrecked with later sound changes. The dual number didn't survive, not helped by how poorly reconstructed it is, though I may let it survive in some intrinsically dual words. The -Vŭ and -Vch endings resemble Albanian -ave and -ash. The latter are both from PIE *-oisu which changed to -oy(x/š)u through RUKI. The former are a resemblance only as Getic derives it from PIE Instrumental plural *-ābhi(s) whereas Albanian is thought to have extended the genitive plural *ōm. In the dative singular there is an alternation between front and back high vowel, as in Albanian. In Getic -u/-e follows a consonant that is always velarised, in this case double L [ɫ]. The genitive singular preserves /s/ by analogising *-osyo > -assa, whereas the nominative /s/ is lost to debuccalisation and lends compensatory lengthening.
I shouldn't have to point out that this isn't even a complete nominal morphology so I haven't provided a translation. Verbal morphology is another mountain to climb entirely. Hopefully this is proof of progress, and even if not it continues to work as a tool for understanding language. The first conlang I tried to create in 2003 for my nation in NationStates was dire and lost in the great hard-drive crash of 2004, yet I walked away with a greater understanding than before. And the same again the following two times, until today I can competently throw terms like 'rhizotonic' around and know what it means. We've come a long way, baby.
[1285]
No comments:
Post a Comment