Ghost language

This article is about the Lamuan languages descended from Khulls and spoken on the planet Teppala. There exists another conlang called Lamu which is not part of this world.

Lámū is a name for the southern branch of the Khulls family, including the languages which are spoken inside Pabap territory. Lamuans are not immigrants; they are an indigenous minority of Khulls speakers who never moved away from their homelands even as they were outgrown by the Pabaps.

Shared features are few because of the early date of divergence. A properly designed family tree of Khulls and its descendants would show the Lamuan languages as several different families, each as diverse from the others as whole families are from other families in the West. Nevertheless all Lamuan languages share:

• The early unrounding of all vowels, so that /a e i o u/ turned into /a e i ɤ ɨ/, which can be spelled as {a e i ə y} or {a e i ø y}. All languages later grew new rounded vowels, generally arising in part from vowels next to labialized consonants.

• Loss of distinctive labialization except before a vowel.

• The devoicing of all voiced stops, unconditionally.

• Change to a vertical vowel inventory consisting of /a ə ɨ/.

• The change of all non-aspirated stops into ejectives (in stressed syllables) and later often into geminates, thus eliminating the "three series" system of the parent language with its latent instabiliity due to the rarity of the voiced and voiceless non-aspirate series.
• The simplification of the tone system, such that although proto-Lamu did still have three tones, the distinctions were weak and the whole system was quick to decay into a length/stress distinction.

• With regards to labialized consonants, Lamuan languages fall into the "K" group, meaning that they neither palatalized the plain velars (as did "C" languages such as Moonshine) nor changed the labiovelars into labials (as did "P" languages such as Ogili). Some languages did lose distinctive labialization, but did so at a late stage of development and generally in a way that did not add any new phonemes to the language. e.g. kʷ pʷ > k p, not p p.

Note, the above may be repurposed for a different language family. It could be said that I'm not abandoning the idea so much as I am changing the name "Lamu" to refer to a new language family with completely different sound changes.

Proto-Lamuan phonolpogy

Vowel shifts

/a e i o u/ > /a ɜ ɨ ɜ ɨ/, probably unconditionally with no effects at all on the surrounding phonemes. This is because /e/ and /u/ were much rarer than /o/ and /i/, and had in many cases arisen in tandem with consonant shifts, so that, for example, /čo/ never existed in Khulls, since /čo/ > /če/ in an earlier shift.

Consonant shifts

Palatalization of non-coronals

The language palatalized most of the non-coronal consonants that were not labialized, and then delabialized the labialized ones. However, there were exceptions:

• /p ṗ b/ changed to /f f v/ instead of becoming palatalized. This shift also occurred in Moonshine and Poswa much later, but was unrelated. However, this shift did not happen when they occurred before another consonant (a situation that came to be seen as analogous to labialization).
• The nasals /m n ŋ/, which had no labialized counterparts, also did not palatalize.
• /ḳ ḳʷ/ shifted to /q q/, and resisted palatalization.

• note, this last shift MUST be conditional, or else /q/ would be more common than /k/.

• Also there will be a gap of almost no /či/, but there will be plenty of /ki/, which would be unusual. So perhaps /kʷi/ > /ćʷi/ > /ći/, part of a pushchain shift that also causes /qi qʷi/ > /ki kʷi/ and probably the rest of the /q/ > /k/ later on.

The phoneme /gʲ/ quickly merged with /j/ (spelled /y/ in Khulls).

Meanwhile, those coronal consonants that were postalveolar came to behave as if they were palatalized, and those that were alveolar came to behave as if they were labialized. The sole exception was /l/, which behaved as if it were a dorsal consonant, and therefore became palatalized in all environments to /lʲ/. This treatment of /l/ as dorsal is shared with many languages of the area such as Andanese and Khulls itself. /ʕʷ/ came to be seen as the labialized version of /l/, which is true to its Khulls derivation, and its pronunciation shifted roughly to ʀ, which still contrasted with the voiced velar fricative /g/.

The old phoneme /ǯ/ came to behave as if it were /rʲ/. However, it is likely that /r/ and /z/ soon merged, and took /ǯ/ and /ž/ with them, creating a merged phoneme pair of /r rʲ/ which would be [r ž].

Loss of ejectives

The ejectives /ṗ ṭ/ had survived the shift that had removed /ḳ/, but they soon merged in with the corresponding aspirated voiceless stops unconditionally. (Unless they triggered gemination instead.)

Loss of syllabic consonants

All syllabic consonants are lost. If followed by a vowel, they simply become plain. If not, they change according to the formula /ḷ ṁ ṅ ŋ̇ ṡ ṣ̌/ > /il im in iŋ is iš/. However, this shift occurred later than the shift that changed consonants such as /p/ into /f/ when not before another consonant. Thus Khulls ʕʷŏpṡ "sun" became proto-Lamu ʀɜpis, not *ʀɜfis.

Tone shifts

The high tone "crystallized" into low tone plus /ʔ/, as in Thaoa. This caused gemination of the following cosnonant except before a fricative, in which case it formed an affricate. This allophonic gemination/affrication was already part of standard Khulls, such that, .e.g. in Khulls /àx/ was always pronounced [àk], but in Lamu it became truly phonemic because the tone contrast disappeared.

ā and á merged as ā, which lost its tonal contrast but retained its length contrast.

Pharyngealization survived, but was no longer considered a tone. Instead, it was considered to be /g/ in the syllable coda. This was an orthographic choice, although there were still a few words in the language where /g/ and the pharyngeal vowel marker /ʕ/ appeared to derive from each other.

NOTE, this doesnt work because a literal /g/ can also occur in the syllable coda and the two are unlikely to merge. /R/ would not work either.

Thus, all tones had been eliminated.

Summary of phonology

Thus the phonology of proto-Lamu was:

CONSONANTS

p b m f v t n s r č š ž ʀ l ć ǵ ś j k ġ ŋ x g q ʔ

VOWELS

a ɜ ɨ

There are no diphthongs.

The consonants that could occur word-finally were /p m n ŋ s l ʔ ʕ/, plus those that had resulted from delabialization. Thus, all except some of the palatals.

Consonant that cannot occur wordfinally: /t r č ž ć ǵ ś/, but note that word-final /s/ was allophonically [ʔs] all along and might change to a simple [t].

/q/ might be a mirage, if it shifts to /k/.

Descendant languages

Some shift the vowels to /a i u/, since there was no palatalization or labialization to pull the vowels to the edges of the vowel space. This shift essentially restores the original system that the Gold language had had more than 4000 years earlier, which makes these languages look especially conservative, but this is false. It is reminiscent of the situation where Moonshine appears to be the only Khulls daughter language that preserves the Khulls á tone, when in fact it lost that tone and then redeveloped it from sequences of other tones.

For consistency, if this shift happens in *all* daughter languages (Lamuan languages share many areal traits), or even most of them, the vowels of the parent language could also be written as /a i u/ with the explanation that the parent language /u/ was very close to schwa.