Palli is a laguage spended from Thaoa. Although originally an indepewdnent society, many Andanese spilled over and soon became the new majoreity. But their religion did not trasnger over. Because of the strong Andfanese influence, /ŋ/ is romanized as g here. (There is no /g/ anuyway)
The phoneme inventory of Palli can be analyzed as being exactly the same as that of Late Andanese, and the two languages share the same set of alphabets, which is the smallest ever recorded in the world. However, in isolation, the phonology of Palli can be analyzed as differing very slightly from that of Late Andanese.
The Palli dialect of the Thaoa language had never had any voiced stops, unlike the northern dialect that led to Sakhi. Thus proto-Palli began with a two-way distinction between plain voiceless stops and voiceless aspirated stops. This distinction soon disappeared, as the aspirated stops merged with a set of preexisting voiceless fricatives. Later, the number of fricatives was reduced. Apart from these two shifts, however, the other changes to the consonant inventory were minor and often polyconditional. No new consonants were evolved. Thus the consonant inventory of Palli in its classical stage around the year 4175 was
/p m t n s l k ŋ q h/
If /w/ and /j/ (spelled y) are analyzed as allophones of the vowels /u i/, then the consonant inventory of Palli is the same as that of Andanese but for the addition of the somewhat rare uvular stop /q/. However, those teachers who believe that the two languages have the same consonant inventory account for the existence of [q] by analyzing it as a cluster /hk/, which otherwise does not exist in the language. However, there is no special relationship between the sound [q] and this cluster that would suggest that the two could be united; it simply fills a convenient gap in the distribution of sounds, and therefore can be used as a digraph without causing ambiguity.
Allophones of consonants
Palli allows closed syllables, unlike Late Andanese. Word-internally, syllables can end in any of /l h k n/.
The final consonant /l/ never assimilates in any way to a following consonant.
The final consonant /h/ is a spelling convention used to represent the inherited phoneme /ʔ/, which could only appear in syllable-final position, and therefore had no minimal pairs with consonants that never appeared in final position. This glottal stop phoneme always assimilates to a following consonant, matching it exactly, and since it can never appear before a vowel, its underlying representation is a matter of a teacher's discretion. /h/ was chosen because of spelling conventions in related languages.
The final consonant /k/ assimilates in manner, but not in position, to a following consonant. Therefore, there is a merger between the underlying phonetic sequences /hk kk/, as [kk], which frees up the use of the spelling hk to represent /q/, whose doubled form is very uncommon.
The final consonant /n/ assimilates in position, but not in manner, to a following consonant. Thus, there is a three-way merger of the underlying sequences /hŋ kŋ nŋ/ as [ŋŋ], and a two-way merger for the other two nasals.
The vowel inventory of Palli is
/a i u/
Unlike the situation with the consonants, here there is not even a slight disagreement between the vowel inventories of Palli and Andanese. The two languages have the same three-vowel setup and neither language has any allophones resulting from vowel sequences or any other construction which could be analyzed as additional phonemic vowels.
The derivation of the Palli three-vowel system from Thaoa's six-vowel system occurred in two simple unconditional changes. First, the somewhat rare vowel /y/ (IPA /ɨ/) changed to /i/ in all positions. Later, the mid vowels /e o/ changed to /a/ in all positions. This last change made /a/ a very common vowel in Palli, accounting for more than half of the occurrences of vowels in an average text. The fact that /i u/ contract to [j w] before another vowel whereas /a/ remains a full vowel in all positions makes the [a] vowel even more salient than it would otherwise be.
Allophones of vowels
The high vowels /i u/ are lowered to near [e o] when in a closed syllable of any kind. This shift postdated the mergers of consonant clusters such as /hŋ/~/nŋ/ and therefore the allophones never became phonemic.
The low vowel /a/, in all positions, can be moved towards [ə] in fast speech. Also in rapid speech, the high vowels /i u/ can become so short as to be inaudible, instead being heard as a modification of the preceding consonant. This is especially common before fricatives. For example, /tuha/ may sound like [tʷha].
Stress and phonotactics
All words are stressed on their final syllable, which may end in a vowel, l, or n. Word-internally, before a consonant, syllables may also end in h or k; however, the /h/ phoneme changes drastically depending on the consonant that follows.
Relationship to Andanese
On the other hand, the native vocabulary changed quite a lot because there were many sound changes which drastically reduced the phonemic inventory. Early on, most nouns were replaced with their oblique forms, which always ended in vowels. Thus words started out in early Palli by actually getting longer instead of shorter. This is analogous to what happened in Vulgar Latin, where for example costūmen "custom" developed into Spanish costumbre.
- NOTE, DUE TO MY REWORKING OF THAOA, A LOT OF THE INFO HERE NO LONGER APPLIES. PREVIOUSLY I WROTE AS IF PALLI STARTED DEVELOPING C.4200 AD, WHICH IS AFTER THE DEATH OF THE ANDANESE NATION. I'VE SINCE MOVED IT BACK TO ~2668 AD. INSTEAD OF SPREADING THE SOUND CHANGES OVER 4500 YEARS (4200 - 8700 AD), THEY WILL BE COMPRESSED SO THAT THE FINAL DATE OF CLASSICAL PALLI WILL BE VERY NEAR TO THAT OF CLASSICAL ANDANESE.
The staircase shift
Palli never had any voiced fricatives. Its voiceless fricatives included an /f/ not present in Sakhi, which mostly corresponded to proto-Sakhi /b/. Later, Palli underwent the "staircase shift" where the four fricatives collapsed into two, dependent on the following vowel. A chart of each consonant paired with each vowel has steps like a staircase, hence the name. The sound change is easier to visualize if the endpoints are /f/ and /s/, with a subsequent universal shift of /f/ > /h/. However, the actual sound change was more complicated than this.
/f/ stayed /f/ always.
/h/ became /f/ when before /a o u/, but became /s/ when before /e i/.
/š/ became /s/ when before /a e i/, but became /f/ when before /o u/.
/s/ stayed /s/ always.
Down to three vowels
Later, there came to be only three vowels, as in ANdanese. The shift was /a e o/ > /a/, unlike Andanese. This happened after the staircase shift.
- /k/ becomes [ʔ] between vowels, and since Palli is mostly CV, this means that [k] is not a particularly common sound in Palli as a phone, unlike Andanese. <----- NOT SURE ABOUT THIS
- Palli is not a mora-timed language, so sequences like /sia/ seem like single syllables to speakers, unlike Andanese this sequence would be pronounced as two (very quickly spoken) syllables. Likewise, the sequence /sii/ sounds like [ši], again behaving as if it were a single syllable. It is thus possible to analyze Palli has having a far larger phonology than Andanese. Put another way, in Andanese, /si/ and /sii/ would sound the same except for length. In Palli /si/ may sound like a single consonant, [š].
Stress accent moves from always-initial to always-final under the influence of Late Andanese.
Palli ported in the entire Andanese vocabulary and added it to their own bocab inheried from Thaoa. Thus the phonemes of ANdfnaes are just a susbeet of the others and are overrrepreseented repsective to Sakhi etc. Thew grammr is mostly still Thaoa-like, e.g. it has noun cases derived from infixes, which means it retains closed syllables, even though the vast majority of the jnouns do not have closed syllables in the nominative case because Andanese didnt have closed syllables at all. Andanese has many synonyms for the same thing with little difference in m,eaning, e.g. latuhi, latuunama, latunuma "boat". Palli takes all of these and oftne uses the longest one .
Palli people do not consider themselves descendants of the disgraced Andanese; instead, they say they served as a refuge for the Andanese while maintaining their own culture.
The early Palli territory resembled Greece physically, in that it had a heavily indented coastline with offshore islands. The climate was cool temperate in the earliest days after the split with Sakhi, but warmed up steadily over time. There were not nearly as many islands in Palli-land as there are in Greece, but the Palli people were nonetheless dominated by their ocean.
Despite being named after a woman, Palli had been invaded by the aggressively virile Andanese people and opposed Sakhi's embrace of radical feminism. When the Pallian leadership learned that Sakhi had been forced to sign a Feminist Compact putting Sakhi firmly under female control and shutting down their military, they contemplated an invasion, planning to use the Sakhi men as slaves and the Sakhi women as prostitutes. But no action was taken, because the two nations were separated by the difficult Hutithasi mountains, and even with their superior male standing army the Pallians knew that such a war would be very difficult.
In some phrases, subjects are circumfixes. That is, the normal noun goes in front, then the verb and other words dependent on it, then an addition word that reflects the noun.
This setup is descended from the proto-language's SOV nouns, which were entire sentences that came to be used as ordinary nouns by removing the inflection from the (always final) verb. Palli differed from the other languages only in that the verb part of the compound became opaque while the object part of the compound remained "alive" and therefore changeable. Thus, it could also be said that objects are infixes within the subjects. However, unlike the SOV words of these other languages, in Palli the compounds are spelled as separable words.
- note, what about the verb? is it also an infix? try to remember the "messy" words like šučupoladaešakšamo and poekšekmek and what the repetitions were for.
For example, if lanami nuhu "orange" is infixed into hasa liha "boy", the result is hasa lanami nuhu liha "a boy ... an orange". Possibly these words would be abbreviated; the final morpheme could be tacked onto the end of the verb, and therefore seen as a sort of verbal noun-class agreement suffix.
The words tend to be divided based on what was once the border between a noun and its verb. These verbs could be thought of as noun classifiers, e.g. liha "talking" applies to humans.
Perhaps instead, all nouns are bipartite, consisting of a root and a suffix. This suffix cliticizes to the last word in the phrase of which it is the topic. If the noun is used in isolation, it cliticizes to its own root and therefore appears as a single word. If it is the subject of a clause, it cliticizes to the verb and therefore appears at the end of the clause (verbs are still always "final") and therefore appears to be a verb inflection. This essentially mimics the Andanese classifier prefix system, except that
- The noun suffix appears only once, and encloses the entire phrase, rather than appearing on each word.
- It is always a suffix, never a prefix. Many Andanese loans have the prefixes intact, but these are treated as part of the stem and will not separate.
To increase efficiency, probably single-syllable verbs will be favored. The word for boy could use -ka instead of -liha, even though animals also use -ka. The language is still by far the most inefficient in its family, however.
- Possibly č > š while this happens.