Afansevan/Phonology

Consonants
* <ŋ> is backened when next to  or a stressed or secondary stressed back vowel. ** Voiceless plosives become aspirated whenever they occur in a monoconsonantal onset in stressed syllables. In nobles that come from mountainous regions, these aspirated plosives become ejectives.  *** The dental fricatives are allophones of the dental laminal alveolar sibilants, but both sounds can be pronounced whatever you want them to be when or  occur.  **** /w/ only occurs in foreign loanwords. *****  only represents /x/ when it is used to distinguish /sx/  from /ʂ/ , and  is used everywhere else to represent /x/.

Vowel Allophony

 * Vowels tend to be realized with glottal stops before them at the beginning of words.
 *  and <ö> tend to be realized as [ʉ].
 * Vowels become nasalized when followed by a coda nasal consonant.
 * Vowels other than  are lowered when followed by a coda , so  becomes [e̞ʁ], <er> becomes [æʁ], <yr> becomes [ø̞ʁ ~ əʁ], <ör> becomes [ɶʁ ~ əʁ], <ur> becomes [o̞ʁ], and <or> becomes [ʌ̞ʁ].

Tone
Afansevan has developed a six-tone system following the loss of plosives in coda position (other than when there are two plosives in the coda, where the first one is lost, which makes tone phonemic). Note that the vowel in this example is <a>, and that the tone occurs at the end of the nucleus. In especially formal vocabulary that was loaned from Sanskrit, a high tone can be used to denote syllables that were stressed in the original language

Pseudo-Diphthongs
Imperial Afansevan allows a set of pseudo-diphthongs, which is formed by combining any vowel with an approximant <j> or <w>. Diphthongs ending with -w are only allowed to occur in loanwords, while the pseudo-diphthong <ij> is not allowed to exist in native words at all, often being used to transcribe long /i/, especially in words from Sanskrit and Nahuatl.

Phonotactics
Afansevan's syllable structure is (C)(C)V(C)(C), where C is any consonant, and V is any vowel, <r>, or <l>. Consonant clusters made of two plosives are not allowed in the coda. All permitted consonant clusters in the onset are any obstruent (other than <r>) followed by <r> or <l>, or a sibilant followed by an obstruent. <R> and <l> can serve as the nucleus of a syllable, in that case they are pronounced as syllabic consonants. Voiced and voiceless obstruents (other than <r>) are not allowed to go together in coda-onset clusters, in that case, the last one assimilates the other ones in voicing (e.g. /pzf/>/psf/). Also, voiced obstruents (other than <r>) are not allowed to end a word, as they become voiceless in that environment, but voiced obstruents are still represented in the romanization, the orthography, and the IPA in phonemic transcriptions (Devoicing at the end of words doesn't exist in the dialect spoken by the Californian royal family). Obstruents of the same place of articulation and manner of articulation of different voicing placed next to each other is also not allowed.

Stress System
Stress always falls on the first syllable, while secondary stress falls on the last syllable. Unstressed and non-secondary stressed vowels are reduced towards the schwa. However, vowel reduction is more complicated in dialects spoken by nobles that come from the extreme north of OTL's California.