Proto-Human language

Proto-Human was the most recent common ancestor of the languages of Etrand and former Hulra, being spoken from 1300 BEKE to roughly 800 BEKE in Etrand and Hulra.

The language was poorly attested, and most written records come from High Elven and Wood Elven sources. However, in Hulra, rich traditions of written literature were starting to developed already in the transition period between Proto-Human and Old Hulran, between 850 BEKE and 750 BEKE, while Etrand remained a land of illiteracy until the Kingdom of Etrand was founded in 0 BEKE/AEKE.

Consonants

 * Loss of palatal and post-alveolar consonants
 * The retroflex nasal [ɳ] merged with the retroflex rhotic [ɻ], which in turn eventually merged with the plain alveolar rhotic /r/ (see below)
 * The retroflex series [ʈ͡ʂʰ ʈ͡ʂ ɖ͡ʐ ʂ ʐ ɭ ɻ] merged with the plain alveolar /t͡sʰ t͡s d͡z s z l r/ series.
 * The palatal nasal [ɲ] and lateral [ʎ] merged with the palatal approximant /j/.
 * The voiceless palatal sibilants /t͡ɕʰ t͡ɕ ɕ/ became /ç/. The voiced palatal sibilants /d͡ʑ ʑ/ merged with /j/.
 * The palatal rhotic [ɺʲ] merged with the plain alveolar /r/.
 * Fricatization: the aspirated stops /pʰ tʰ kʰ kʷʰ/ became fricatives /ɸ θ x xʷ/. The aspirated affricate /t͡sʰ/ lost its aspiration and merged with /t͡s/.
 * Fricatization of intervocalic /b d/ to [β ð]
 * Lenition of /g gʷ/ to /ɣ ɣʷ/ everywhere except after nasals
 * Merger of the guttural rhotic /ʁ/ with the plain alveolar rhotic /r/
 * Loss of /ŋ/ and /ŋʷ/
 * Syllable-onset /ŋ ŋʷ/ became /ɣ ɣʷ/
 * Syllable-coda /ŋ/ became /n/

Consonants
The voiced non-sibilant fricatives [β] and [ð] appeared only as post-vocalic allophones of /b/ and /d/. They also served as intervocalic allophones of their voiceless counterparts /ɸ/ and /θ/.

The voiced velar stops [g] and [gʷ] appeared only as allophones of /ɣ/ and /ɣʷ/ after nasals.

The velar nasal [ŋ] was the allophone of /n/ before /k x ɣ/. Do note that /nɣ/ was pronounced as [ŋg], not [ŋɣ].

The labio-velar nasal [ŋʷ] was the allophone of /n/ before /kʷ xʷ ɣʷ/. Do note that /nɣʷ/ was pronounced as [ŋʷgʷ], not [ŋʷɣʷ].

Proto-Human's one and only rhotic phoneme /r/ had a variable pronunciation, ranging from apico-alveolar to post-velar / pre-uvular. It likely depended on the surrounding consonants, and the post-velar allophone likely also had a slight fricatization ''(similiar to real-life Czech <ř> /r̝/, just post-velar / pre-uvular instead of alveolar). ''Likely, [r̺] was used "normally", and [ʀ̟] was used after velar consonants, maybe even before back vowels word-initially.

The palatal fricative /ç/ was a rather unstable phoneme, and it likely merged with /x/ before splitting up into the descendant languages. Likewise, the voiced labio-velar fricative /ɣʷ/ was also an unstable phoneme. In all descendant languages, [ɣʷ] merged with /w/, and /nɣʷ/ [ŋʷgʷ] merged with /mb/. The fact that these mergers happened in both Old Etrandish and Old Hulran seem to imply that the sound shift could have happened at an earlier point, most likely in Late Proto-Human.

Monophthongs
The laxed vowels [ɪ ʊ] only appeared as allophones of short /i u/ in closed syllables where the coda-consonant was a velar consonant. That also extends to [ŋ], which itself was the allophone of /n/ before velar consonants.

While the long /eː/ was a close-mid [eː] and the long /ɛː/ was open-mid [ɛː], the short /e/ was likely truly middle, neither [ɛ] nor [e], but in between the two: [e̞]. Or, according to other theories, it could have been a close-mid [e], like it's long counterpart.