1. Pseudo-Indo-European (very similar to PIE languages in terms of vocab and grammar, but there are no regular sound changes)
2. Focus on a nice sound
3. Strongly morphological inflection, including ablaut, infixes, stem changes and anlaut mutations
4. Matriarchalic sprachbau, i.e. the primary form is always the feminine one, the masculine one is marked
Phonetics/Phonology
This section deals with the phonetics of Paatherye.
Consonant inventory
Nothing too special in here, except maybe for the labialized bilabial approximant /ʋʷ/ that opposes /w/ but is realized as [v] word initial. The romanization is in brackets (native script coming later on)
/p pʰ t tʰ k kʰ/ <p ph t th k kh>
/t͡s d͡ʒ/ <c j>
/f ð s ʒ ɣ h/ <f d/dh s ș g/gh h>
/m n/ <m n>
/ʋʷ w j/ <v w y>
/r l/ <r l>
Vowel inventory
Paatherye distinguishes only four different vowels: /a e i u/, as well as two diphtongs: /ai̯ ei̯/. Vowel length is a distinctive feature.
/a aː/ <a ā>
/e eː i iː/ <e ē i ī>
/u uː/ <u ū>
/ai̯ aːi̯ ei̯ eːi/ <ay āy ey ēy>
Phonotactics
The maximum syllable structure is (C)(C)V(N). The coda consonant can only be one of /l m n r s/, and /s/ can only occur after /a/ and /e/. If the vowel changes due to ablaut, the consonant changes too: It becomes /r/ (Rhotacism)
Example: ma khēlir (the money, nominative), but mā khēlisan (the money, accusative): The inherent /s/ was changed to [r] in the nominative form, where it is the anlaut of the subsequent syllable in the accusative form.
Only the first syllable of a word can have two anlaut consonants. Word-intern, only one consonant is allowed, so there are anlaut mutations occuring if, for example, a word starts with a prefix. Example:
twan (big) -> stāwen (biggest) | Dissimilation of /tw/ to /w/.
A complete list of all possible two-consonant anlauts can be found behind the spoilers.
Native alphabet
Paatherye uses a slightly modified version of the Devanagari alphabet with 39 letters. It is an abugida, i.e. every consonant has an inherent vowel /a/ that can be modified accordingly. What's special about that writing system is that /m n r l s/ have alternative forms when they occur in coda, hence it is always clear which consonant serves as anlaut, and which as coda. If a word starts with a vowel, a "dummy consonant" is used: श doesn't bear any phonetic value. Similar to most Indic scripts, the alphabet is sorted after place of articulation, with the vowels coming first, then back consonant to the front, suborded from "hard" consonants to "soft" consonants. Early Elvish writers didn't have exact knowledge of the places of articulation, so there are some "irregularities", as /h/ is placed at the front consonants, for examples. Last come the special coda forms of the consonants.
श शा शै शाग़ <a ā ay āy> [a aː ai̯ aːi̯]
शे शो शौ शौग़ <e ē ey ēy> [e eː ei̯ eːi̯]
शि शी शु शू <i ī u ū> [i iː u uː]
ख क घ ग व <kha ka gha ga wa> [kʰa ka ɣa ɣa wa]
ज ष य न ल <ja șa ya na la> [d͡ʒa ʒa ja na la]
ठ त च ड़ ध द <tha ta ca ra dha da> [tʰa ta t͡sa ra ða ða]
फ प फ़ व्व म ह <pha pa fa va ma ha> [pʰa pa fa ʋʷa ma ha]
शं श़ शर शऌ शज़ <am an ar al as> [am an ar al as]
Note that the distinction between घ and ग (gh/g) as well as between ध and द (dh/d) is purely etymological and doesn't bear any phonetic value.
Next: Phonotactics, Prosody and Stress.