Battlax

I have been compelled to create yet another phonology sketch. Well, more like a family of four. They're like snacks, only the come out of you rather than going into you.

Proto-form

P	tʰ	ʧʰ	kʰ
	t	ʧ	k
	d	ʤ	ɡ
	s	ʃ	x
m	n	ɲ
w	l	j

Not quite decided what the labial obstruent should be phonetically. Doesn't really matter anyway.

Branch 1

The conservativ one.

Proto-stage

• P → p, conditionally pʰ
• Stop phonation shift the 1st
  • Voiced → aspirated

Insert non-inventory-disrupting changes here.

Until modern stage

• Stop phonation shift the 2nd
  • Aspirated → voiceless
    • except pʰ → f
  • Voiceless → voiced
• l → r

Result

	t	ʧ	k
b	d	ʤ	ɡ
f	s	ʃ	x
v	s	ʒ
m	n	ɲ
w	r	j

Branch 2

More mixing going on in here.

Proto-stage

• P → ɸ → h
• POA shift the 1st
  • Labiodentogenesis (dentalogenesis? dentigenesis?)
    • s → f
    • w → v
  • Sibilants: postalveolar > alveolar
    • also ʧʰ → ts (unaspirated)

Until classical stage

• 1st spirantization
  • v d ɡ → b~β d~ð ɡ~ɣ (B D G)
    • spirant allophones intervocalic
    • additionally: j h → D G or zero
• POA shift the 2nd
  • Conditional labialization of velars
• ts → θ
• z → r
• ɣ ɣʷ → j w

(So where am I going to put a Classical Language on Earth'? Don't feel like doing India, and this doesn't match the Central Asia Sprachbund. Maybe Middle East, this could be Classical Egyptian… There's something in the complex spirant system resembling Ouiqál, so this might indeed fit in the East Mediterranean Sprachbund even if it doesn't have voiceless nasals. That would however mean having to relocate the EMSB relativs of Dork Elvish, maybe to eastern Anatolia. Or am I overdoing the separation? If Tsonboran dies around -4ka at latest, there ought to be plenty of room for this in Persia, too.)

Post-classical stage

• Fricativ voicing messes up & phonemicizes the B D G allophony
  • Check history of English for details?
• 2nd spirantization
  • Aspirate stop → spirant
  • Messes up the previous even better… & we probly get some fairly tasty orthographical arcaisms in the process
• ɲ → j (restoring it as an initial)

Until modern stage

• POA shift the 3rd
  • Labialization contrast of velars shifted to palatalization contrast
    • …after which the palatal ones decay to postalveolar affricates
    • CF Ethiopian'!
    • Contemporary vowels should count too; perhaps via chains like kʷu → ku → ky → ci → ʧi
  • Also in the same palatalization hullabaloo: s z → ʃ ʒ
• Kill interdentals
  • θ → s
  • ð → n, r?
• Remaining ɣ → ɦ

Branch 3

Batfuck.

Proto-stage

• P → f
• x → h (initial)
• ɲ → ŋ (coda), n (initial), nj (medial)
• kʰ → kx
• ɡ → ɣ
• Stop phonation shift the 1st
  • voiceless → mild ejectiv (CF Klusterax B?)

Old stage (still well preceding the subfamily's breikup & Classic #2)

• Schirantism: (k)x ɣ → (t)ʂ ʐ
• l → r intervocally
• Dental weakening
  • lT → lː → l
  • lenited initially
    • tʰ tʼ n → h ʔ n̥ (CF Telefax? last one seems to point towards W. Mediterranean)
    • /d/ remains (a la Classic, or because backed to alveolar) vs. weakens to [ð], then null (a la Telefax)??

The point here is to get a single phonological [+POSTERIOR] series, constituting of a velar ejectiv, retroflexen (fleces? :b), and heng.

Middle stage

• ʐ → ɻ → ɽ
• Stop phonation shift the 2nd
  • d ʤ → d̥ ʤ̥ / #_ (possibly also in some medial positions)
  • aspiration lost medially
  • so we have initially ejectiv <> aspirate <> lenis; medially ejectiv <> tenuis <> voiced
• Sibilant → (lenis) affricate initially
  • further ts → d̥ (via [θ] probably)

Until modern stage

• (t)ʂ → (t)s
• ejecticity <> aspiration ceases to be contrastiv
  • tʼ tʰ lenited, ʧʰ originally rare, kʰ routed previously → no biggie
• ɽ → r

Result

	t	ts	ʧ	k	ʔ
	d		ʤ
f		s	ʃ		h
m	n		ŋ
w	l	r	j

This one needs to stay in a non-/p b/ Sprachbund. Hmm, first, there's that one branch of Mivizian, but that would be too far south in Africa. Besides, duplicating RL areal feature locations is lame. NW India? Maybe not. Greece ~ Balkan? Even Caucasus might not be out of question, and it's a greit place for Sprachbunds (can we get more than one in there?) Ouiqál isn't a problem, as it's a newcomer in the region.

I like how we get contemporary /r/ out of three completely different original consonants - /l ʤ ɡ/ ;)