Editing Wichita language (section)

==Phonology==
The phonology of Wichita is unusual, with no pure [[labial consonant]]s (though there are two [[Labialized velar consonant|labiovelars]] /kʷ/ and /w/). There is only one nasal (depending on conflicting theory one or more nasal sounds may appear, but all theories seem to agree that they are allophones of the same phoneme, at best), and possibly a three vowel system using only height for contrast.<ref name="Rood, 1975, p. 315-337" />

===Consonants===

Wichita has 10 [[consonant]]s. In the [[Americanist phonetic notation|Americanist]] orthography generally used when describing Wichita, {{IPA|/t͡s/}} is spelled {{angle bracket|c}}, and {{IPA|/j/}} is {{angle bracket|y}}.

{| class="wikitable" style=text-align:center
|-
! rowspan="2" |
! rowspan="2" |[[Alveolar consonant|Alveolar]]
! colspan="2" | [[Dorsal consonant|Dorsal]]
! rowspan="2" | [[Glottal consonant|Glottal]]
|-
! <small>plain</small>
! <small>[[Labialization|labial.]]</small>
|-
! [[Plosive]]
| {{IPA link|t}}
| {{IPA link|k}}
| {{IPA link|kʷ}}
| {{IPA link|ʔ}}
|-
! [[Affricate]]
| {{IPA link|t͡s}}
|
|
|
|-
! [[Fricative]]
| {{IPA link|s}}
|
|
| {{IPA link|h}}
|-
! [[Sonorant]]
| {{IPA link|ɾ}} ~ {{IPA link|n}}
|
|
|
|-
! [[Semivowel]]
|
| {{IPA link|j}}
| {{IPA link|w}}
| 
|}

Though neither Rood nor Garvin include nasals in their respective consonant charts for Wichita, Rood's later inclusion of nasals in phonetic transcription for his 2008 paper ("Some Wichita Recollections: Aspects of Culture Reflected in Language") support the appearance of at least {{IPA|/n/}}.<ref name="Rood, 2008, p. 395-405" />
* [[labial consonant|Labials]] are generally absent, occurring in only two roots: '''kammac''' ''to grind corn'' and '''camma:ci''' ''to hoe, to cultivate'' (⟨c⟩ = {{IPA|/t͡s/}}).
* Apart from the {{IPA|/m/}} in these two verbs, [[Nasal stop|nasals]] are allophonic. The allophones {{IPA|[ɾ]}} and {{IPA|[n]}} are in [[complementary distribution]]: It is {{IPA|[n]}} before alveolars ({{IPA|/t, ts, s/}} and in geminate {{IPA|[nn]}}) and initially before a vowel, and {{IPA|[ɾ]}} elsewhere. Thus its initial consonant clusters are {{IPA|[n]}} and {{IPA|[ɾ̥h]}}, and its medial & final clusters are {{IPA|[nts], [nt], [ns], [nn], [ɾʔ], [ɾh]}}.
* Final ''r'' and ''w'' are voiceless: {{IPA|[ɾ̥], [w̥]}}
* Glottalized final consonants: One aspect of Wichita phonetics is the occurrence of glottalized final consonants. Taylor asserts that when a long vowel precedes a glottal stop (ʔ), there is no change to the pronunciation. However, when the glottal stop is preceded by a short vowel, the vowel is eliminated. If the short vowel was preceded by a consonant, then the consonant is glottalized. Taylor hypothesizes that these glottalized final consonants show that the consonant was not originally a final consonant, that the proto form (an earlier language from which Wichita split off, that Taylor was aiming to reconstruct in his paper) ended in a glottal stop, and that a vowel has been lost between the consonant and glottal stop.{{sfn|Taylor|1963}}

{| class="wikitable" style="text-align:center"
|-
! Original word ending
! Change
! Result
! Wichita example
|-
| {{IPA|[Vːʔ#]}}
| No change
| {{IPA|[Vːʔ#]}}
|
|-
| {{IPA|[VːVʔ#]}}
| -{{IPA|[V]}}
| {{IPA|[Vːʔ#]}}
| {{IPA|[hijaːʔ]}} (snow)
|-
| {{IPA|[CVʔ#]}}
| -{{IPA|[V]}}
| {{IPA|[Cʔ#]}}
| {{IPA|[kiːsʔ]}} (bone)
|}

: {{IPA|Vː}} - long vowel
: V - short vowel
: C - consonant
: # - preceding sound ends word
* Taylor also finds that previous phonetic transcriptions have recorded the phoneme {{IPA|/ts/}} (aka {{angbr|c}}), as occurring after {{IPA|/i/}}, while {{IPA|/s/}} is recorded when preceded by {{IPA|/a/}}.{{sfn|Taylor|1963}}
* The {{IPA|*kʷ, *w, *p}} merger; or Why Wichita Has No {{IPA|/p/}}:
** In Wichita the sounds {{IPA|/kʷ/}} and {{IPA|/w/}} are not differentiated when they begin a word, and word-initial *p has become {{IPA|/w/}}. This is unusual, in that the majority of Caddoan languages pronounce words that used to begin with *w with {{IPA|/p/}}. In Wichita, the three sounds were also merged when preceded by a consonant. Wichita shifted consonant initial *p to {{IPA|/kʷ/}} with other medial occurrences of *p. {{IPA|/kʷ/}} and {{IPA|/w/}} remain distinct following a vowel. For example, the word for 'man' is {{IPA|/wiːt͡s/}} in Wichita, but {{IPA|/piːta/}} in South Band Pawnee and {{IPA|/pita/}} in Skiri Pawnee.{{sfn|Taylor|1963}}

===Phonological rules===
* The coalescence of morpheme-final {{IPA|/ɾ/}} and subsequent morpheme-initial {{IPA|/t/}} or {{IPA|/s/}} to {{IPA|/t͡s/}}:
{{interlinear|indent=3|ti-r-tar-s|c1= → ''ticac''|IND-PL-cut-IMPERF|'he cut them'}}
{{interlinear|indent=3|a:ra-r-tar|c1= → ''a:racar''|PERF-PL-cut|'he has cut them'}}
{{interlinear|indent=3|a:ra-tar|c1= → ''a:ratar''|PERF-cut|'he has cut it'}}
* {{IPA|/w/}} changes to {{IPA|/kʷ/}} whenever it follows a consonantal segment which is not {{IPA|/k/}} or {{IPA|/kʷ/}}:
{{interlinear|indent=3|i-s-wa|c1= → ''iskwa''|IMP-you-go|'go!'}}
{{interlinear|indent=3|i-t-wa|c1= → ''ickwa''|IMP-I-go|'let me go!'}}
* {{IPA|/ɾ/}} changes to {{IPA|/h/}} before {{IPA|/k/}} or {{IPA|/kʷ/}}. The most numerous examples involve the collective-plural prefix ''r-'' before a morpheme beginning with {{IPA|/k/}}:
{{interlinear|indent=3|ti-r-kita-re:sʔi|c1= → ''tihkitare:sʔi''|IND-COL-top-lie.INAN|'they are lying on top'}}
* {{IPA|/t/}} with a following {{IPA|/s/}} or {{IPA|/ɾ/}} to give {{IPA|/t͡s/}}:
{{interlinear|indent=3|keʔe-t-rika:s-ti:kwi|c1= → ''keʔecika:sti:kwi''|FUT-I-head-hit|'I will hit him on the head'}}
* {{IPA|/t/}} changes to {{IPA|/t͡s/}} before {{IPA|/i/}} or any non-vowel:
{{interlinear|indent=3|ta-t-r-taʔas|c1= → ''taccaʔas''|IND-I-COL-bite|'I bit them'}}
* {{IPA|/k/}} changes to {{IPA|/s/}} before {{IPA|/t/}}:
{{interlinear|indent=3|ti-ʔak-tariyar-ic|c1= → ''taʔastariyaric''|IND-PL-cut.randomly-repeatedly|'he butchered them'}}
* {{IPA|/ɾ/}}, {{IPA|/j/}}, and {{IPA|/h/}} change to {{IPA|/s/}} after {{IPA|/s/}} or {{IPA|/t͡s/}}:
{{interlinear|indent=3|ichiris-ye:ckeʔe:kʔa|c1= → ''ichirisse:ckeʔe:kʔa''|bird-ember|'redbird'}}<ref name="Rood, 1975, p. 315-337" />

===Vowels===

Wichita has either three or four vowels, depending on analysis:{{sfn|Taylor|1963}}<ref name="Rood, 1975, p. 315-337" /><ref name="Garvin, 1950, p. 179-184" />

{| class="wikitable" style=text-align:center
|-
!
! [[Front vowel|Front]]
! [[Back vowel|Back]]
|-
! [[Close vowel|High]]
| ɪ ~ i ~ e
|
|-
! [[Mid vowel|Mid]]
| ɛ ~ æ
| (o/u)
|-
! [[Open vowel|Low]]
|
| ɒ ~ a
|}

These are transcribed as {{angbr|i, e, a, o/u}}.

Word-final vowels are [[phonation|devoiced]].

Though Rood employs the letter {{angbr|o}} in his transcriptions,<ref name="Rood, 2008, p. 395-405" /> Garvin instead uses {{angbr|u}}, and asserts that {{IPA|/u/}} is a separate phoneme.<ref name="Garvin, 1950, p. 179-184" /> However, considering the imprecision in vowel sound articulation, what is likely important about these transcriptions is that they attest to a back vowel that is not low.

Taylor uses Garvin's transcription in his analysis, but theorizes a shift of *u to {{IPA|/i/}} medially in Wichita, but does not have enough examples to fully analyze all the possible environments. He also discusses a potential shift from *a to {{IPA|/i/}}, but again, does not have enough examples to develop a definitive hypothesis. Taylor finds {{IPA|/ɛ/}} only occurs with intervocalic glottal stops.{{sfn|Taylor|1963}}<ref name="Garvin, 1950, p. 179-184" />

Rood argues that {{IPA|[o]}} is not phonemic, as it is often equivalent to any vowel + {{IPA|/w/}} + any vowel. For example, {{IPA|/awa/}} is frequently contracted to {{IPA|[óː]}} (the high tone is an effect of the elided consonant). There are ''relatively'' few cases where speakers will not accept a substitution of vowel + {{IPA|/w/}} + vowel for {{IPA|[o]}}; one of them is {{IPA|[kóːs]}} 'eagle'.{{clarify|reason=Is this saying that [kóːs] would NOT be understood to mean "eagle"?|date=June 2021}}

Rood also proposes that, with three vowels that are arguably high, mid, and low, the front-back distinction is not phonemic, and that one may therefore speak of a '[[vertical vowel system|vertical]]' vowel inventory (see below). This also has been claimed for relatively few languages, such as the [[Northwest Caucasian languages]] and the [[Ndu languages]] of [[Papua New Guinea]].

There is clearly at least a two-way contrast in [[vowel length]]. Rood proposes that there is a three-way contrast, which is quite rare among the world's languages, although well attested for [[Mixe language|Mixe]], and probably present in [[Estonian language|Estonian]]. However, in Wichita, for each of the three to four vowels qualities, one of the three lengths is rare, and in addition the extra-long vowels frequently involve either an extra [[morpheme]], or suggest that [[prosody (linguistics)|prosody]] may be at work. For example,

::{{IPA|nɪːt͡s.híːːʔɪh}} 'the strong one'
::{{IPA|nɪːːt͡s.híːːʔɪh}} 'the strong one'''s''''

::{{IPA|hɛːhɪɾʔíːɾas}} 'let him find you'
::{{IPA|hɛːːhɪɾʔíːɾas}} 'let him find '''it''' for you'

::{{IPA|háɾah}} 'there'
::{{IPA|háːɾɪh}} 'here it is' (said when handing something over)
::{{IPA|háːːɾɪh}} 'that one'

(Note that it is common in many languages to use prosodic lengthening with demonstratives such as 'there' or 'that'.)<ref name="Rood, 1975, p. 315-337" />

This contrasts with Mixe, where it is easy to find a three-way length contrast without the addition of morphemes.<ref name="Rood, 1975, p. 315-337" />

Under Rood's analysis, then, Wichita has 9 phonemic vowels:<ref name="Rood, 1975, p. 315-337" />

{| class="wikitable" style=text-align:center
|-
!
! [[Vowel length|Short]]
! [[Vowel length|Long]]
! [[Vowel length|Overlong]]
|-
! [[Close vowel|High]]
| ɪ
| ɪˑ
| ɪː
|-
! [[Close-mid vowel|Mid]]
| ɛ
| ɛˑ
| ɛː
|-
! [[Open vowel|Low]]
| a
| aˑ
| aː
|}

===Tone===

There is also a contrastive high [[tone (linguistics)|tone]], indicated here by an acute accent.

===Syllable and phonotactics===

While vowel clusters are uncommon (unless the extra-long vowels are clusters), [[consonant cluster]]s are ubiquitous in Wichita. Words may begin with clusters such as {{IPA|[kskh]}} ({{IPA|kskhaːɾʔa}}) and {{IPA|[ɾ̥h]}} ({{IPA|ɾ̥hintsʔa}}). The longest cluster noted in Wichita is five consonants long, counting {{IPA|[ts]}} as a single consonant {{IPA|/c/}}: {{IPA|/nahiʔinckskih/}} 'while sleeping'. However, Wichita [[syllable]]s are more commonly CV or CVC.