Proto-Kát language

Proto-Kát (PKát) is the reconstructed common ancestor of the Kátic branch of the Koyanic languages.

Name
The name Kát is taken from the stem *kát- "individual person, people", following a pattern adopted for naming Koyanic branches after their reflex of the Proto-Koyanic term *qə́x₂c "one" (see also: Proto-Qʼač, Proto-Qak and Proto-Gäj). The original root also continued to be used as a numeral in the Proto-Kát form *thrikãʔ, with the addition of a classifier prefix *thri- (← PKn *tʰl̥ə́x₁s "count") and debuccalisation of the coda *-t to *-ʔ.

The full inflection of *kát- can be seen in the table below: *kát- may also be the source of Proto-Hautaric *qʰát "person", thus also the name Ħautkar (← *qʰá⟨u⟩t-ek-er) and Ħautar delûr (← *qʰá⟨u⟩t-er délu⟨u⟩-er)

Consonants
The consonants highlighted in green were allowed in coda position.

Allophones with *h
The aspirated plosives (pʰ tʰ cʰ kʰ) are mostly retained from Proto-Koyanic, but in Proto-Kát are analysed as clusters with *h, (ph th ch kh). The sibilant *s may also appear in sequence with *h, but the phonetic realisation of *sh is not known. Some have suggested that the sibilant *s was characteristically voiced [z], with the devoiced allophone [s] before *h, similar to the sonorants. Another theory is that these were affricates or truly just fricatives with an aspiration distinction [(t)s (t)sʰ].

Vowels
Proto-Kát features three vowel qualities and a symmetrical set of six tones, for a total of 18 phonemic vowels. The phonetic quality of the tones is not readily reconstructible, but the following symmetries are believed to represent the structure of the tone system:


 * Proto-Koyanic coda continuants tend to result in group A tones if unvoiced and group B tones if voiced.
 * The tones of group A follow inverse contours to group B, e.g. if A3 was a high falling tone, B3 would be a low rising tone.
 * Contour 1 was flat, Contour 2 was likely either a smooth rise or fall and Contour 3 may have been more complex and been pronounced longer.

Syllable structure
Proto-Kát contrasted two different syllable types. Strong syllables occur only in word stems and have the form (O)(h)(S)(G)VT(C):

Stressed syllables in a stem could take all six tones, while other syllables could take only {A1 A2 B1 B2}.
 * Obstruent = {p t s c k ʔ} (Note that ʔh neutralises to h)
 * Sonorant = {m n ɲ ŋ r l}
 * Glide = {y w}
 * Vowel = {i a u}
 * Tone = {A1 A2 (A3) B1 B2 (B3)}
 * Coda = {m n ŋ p t s k ʔ r l}

Weak syllables, occurring only in inflectional affixes, did not carry phonemic tone and had the slightly reduced structure: (O)(S/G)V(C). By convention these are transcribed as unmarked *i *a *u, identical in form to the A1 tones, but it is not known whether these were truly toneless, or whether Proto-Kát had a form of tone harmony in which they carried a tone which agreed with the tone group of the previous syllable.

Pre-Proto-Kát
Pre-Proto-Kát is an intermediate language stage between Proto-Koyanic and Proto-Kát, discernible by internal reconstruction and comparison with historical contact languages. It represents the direct ancestor to Proto-Kát spoken around the 4th millennium BGS, an early diverging Koyanic branch characterised by three key developments:


 * 1) Development of tone from Late Proto-Koyanic vowel phonation
 * 2) Simplification of complex sonorant clusters
 * 3) Rhotacisation of *z

Tonogenesis
Tones in Kát were initially generated from Late Proto-Koyanic vowel phonation in a regular, systematic way, with changes dependent on stress and length: The PKn laryngeals *x₁ *x₂ *x₃ produced the qualities *i *a *u respectively, while the schwa *ə merged with *a, resulting in the following vowel inventory: It is believed that at this stage, contour 3 was simply a lengthened form of contour 2. A full table of correspondence is shown below:

Sonorant cluster simplification
Complex onset and coda clusters consisting purely of sonorants were simplified in many cases:


 * N₁N₂- in onset clusters simplifies to a single nasal N₂-
 * Liquids L following nasals N were lost completely.

Rhotacism of *z
Proto-Koyanic *z rhotacised to *r when next to vowels and glides and merged with *s elsewhere.

Dorsal stops chain shift
The dorsal plosives *c *cʰ *k *kʰ *q *qʰ each shifted forward to the next place of articulation, resulting in a merger of PKn palatal and alveolar plosives.


 * c cʰ > t tʰ
 * k kʰ > c cʰ
 * q qʰ > k kʰ

Rhotacisation of medial *l
Medial *-l- was rhotacised to *-r- when following alveolar consonants, e.g. the PKn root *tʰl̥ə́x₁s "to count" became the numeral prefix *thri-.

Tone shift
Tones shifted according to the phonation of coda consonant clusters, which in Proto-Koyanic were either fortis or lenis.


 * B1/B3 B2 → A2 A1 before a fortis coda (pʰ tʰ s cʰ kʰ h m̥ n̥ ɲ̊ ŋ̊ r̥ l̥ ẙ w̥):
 * B1 {ĩ ã ũ}, B3 {î â û} → A2 {í á ú}
 * B2 {ì à ù} → A1 {i a u}
 * A1/A3 A2 → B2 B1 before a lenis coda (p t z c k m n ɲ ŋ r l y w):
 * A1 {i a u}, A3 {ǐ ǎ ǔ} → B2 {ì à ù}
 * A2 {í á ú} → B1 {ĩ ã ũ}

In fact, this may represent a tone split from four into six tones, since A3 {ǐ ǎ ǔ} and B3 {î â û} may not have been truly contrastive tones until this time. The result of this shift was that a fortis coda may only follow a vowel with an A tone, while lenis codas follow only B tones, as shown in the table below: The tone shift was also an important piece of evidence in the novel reconstruction of Proto-Koyanic, which posits that PKn lenis plosives *p *t *c *k *q were originally voiced [b d ɟ g ɢ]. This would provide a better phonetic motivation for the split in tone quality, since voicedness of coda consonants affecting tone is a well-established phenomenon.

After this change, all voiceless sonorants merged with their voiced counterparts in coda clusters, with *h being lost completely.

Monophthongisation
Stressed:

Unstressed:
 * {a, á, ǎ}y {a, á, ǎ}w → ǐ ǔ
 * {ã, à, ǎ}y {ã, à, ǎ}w → î û


 * ay aw → i u
 * áy áw → í ú
 * ãy ãw → ĩ ũ
 * ày àw → ì ù

Under the generally accepted assumption that contour 3 was pronounced with a longer vowel than contours 1 and 2, the stressed tone shift to contour 3 here might represent an example of compensatory lengthening due to loss of the glide.

Coda cluster simplification
Coda clusters were reduced to single lenis consonants. The cells highlighted in blue are cases where a liquid consonant was lost, triggering the same type 1, 2 → 3 contour shift associated with stress and compensatory lengthening which occurred in monophthongisation. Note that loss of a nasal did not trigger this change:


 * Glide or liquid loss triggers tone shift:
 * *aẙtʰ → *ǐt
 * *al̥t → *ǎt
 * *al̥n̥t → *ǎt
 * Nasal loss does not:
 * *an̥tʰ → *at

Stem
Noun and verb stems consist of strong syllables and always end in a consonant other than *y, *w or *h.

Numerals
Proto-Kát, used a base-10 system of numerals. The ten basic numerals are shown below: Notes: *thri- < *tʰl̥ə́x₁s was incorporated into numbers as a prefix to numerals.