Taos phonology

From Wikipedia, the free encyclopedia

Main article: Taos language

The main description of Taos phonology was contributed by George L. Trager in a (pre-generative) structuralist framework. Earlier considerations of the phonetics-phonology were by John P. Harrington and Jaime de Angulo.^[1] Trager's first account was in Trager (1946) based on fieldwork 1935-1937, which was then substantially revised in Trager (1948) (due in part to the inclusion of juncture phonemes and newly collected data in 1947 in the analysis). The description below takes Trager (1946) as the main point of departure and notes where this differs from the analysis of Trager (1948). Harrington's description (although from a different period) is more similar to Trager (1946). Certain comments from a generative perspective are noted in a comparative work Hale (1967).

1 Segments
- 1.1 Consonants
  - 1.1.1 Alternations
- 1.2 Vowels
2 Prosody
3 Loanword phonology
4 See also
5 Notes
6 Bibliography

[edit] Segments

The two following sections detail phonetic information about Taos phonological segments (i.e., consonants and vowels), as well as their phonological patterning in morphophonemic alternations.

[edit] Consonants

Trager (1946) lists 27 consonants for Taos, although in his later analysis he posited 18 consonants.^[2]

		Bilabial	Dental		Alveolar	Palatal	Velar		Glottal
		Bilabial	central	lateral	Alveolar	Palatal	plain	labial	Glottal
Stop	voiced	b	d				g
	unaspirated	p	t			tʃ	k	kʷ	ʔ
	aspirated	pʰ	tʰ
	ejective	p’	t’			tʃ’	k’	kʷ’
Fricative		(f)		ɬ	s		x	xʷ	h
Nasal		m	n
Flap					(ɾ)
Approximant				l		j		w

The flap /ɾ/ is a borrowed phoneme (< Spanish [ɾ]) that occurs in loanwords from New Mexican Spanish that were borrowed relatively recently.^[3]
The fricative /f/ only occurs in Spanish loanwords in a syllable-initial cluster /fɾ/ and may be labio-dental or bilabial [ɸ]. It is briefly mentioned in Trager (1946) and ultimately excluded from the phonological description (along with the other loanword clusters).^[4]
Voiceless stops /p, t, tʃ, k, ʔ/ are very slightly aspirated.
The ejectives /p’, t’, tʃ’, k’/ are weakly glottalized.^[5]
In some speakers, /pʰ/ may have an assimilated bilabial fricative [ɸ]: [pɸ]. This can also vary with a deleted stop closure. Thus, /pʰ/ has the following free variation: [ph~pɸ~ɸ].
The stops /b, d, g/ are voiced [b, d, g] intervocalically and voiceless, unreleased, and with a long closure duration [p̚ː, t̚ː, k̚ː] syllable-finally. Syllable-initially, they only occur in loanwords where Trager describes them as "less voiced".^[6]
There a neutralization of the contrast between the labialized consonants /kʷ, kʷ’, xʷ/ and their non-labial counterparts /k, k’, x/ before the (labial) high back vowels /u, ũ/.^[7] In this environment, Trager assumes the non-labials are present.
Fricative /x/ has weak frication, unlike the stronger frication found in other languages (such as, the closely related Picuris language).
The dental consonants /t, d, n, l, ɬ/ are phonetically denti-alveolar.
Voiceless /tʃ/ is phonetically an affricate and usually post-alveolar [tʃ]. However, it can freely vary with a more forward articulation ranging from post-alveolar to alveolar: [tʃ ~ ts]. Trager (1946) notes that some speakers tend to have more forward articulations [ts] before the vowels /e, o/. Ejective /tʃ’/ also has variation between post-alveolar to alveolar.^[8]
Fricative /s/ tends to have a post-alveolar allophone before high vowels /i, u/ (especially the high front vowel /i/).
The lateral liquid /l/ is velarized at the end of syllables: [ɫ].
The labial glide /w/ is labio-velar.
The glides /w, j/ are phonetically short high vowels [u̯, i̯] no closer than Taos high vowels, which are very close as well. When they occur after nasal vowels, they are nasalized: [ũ̯, ĩ̯].

[edit] Alternations

Please help improve this section by expanding it.
Further information might be found on the talk page or at requests for expansion.

[edit] Vowels

[edit] Monophthongs

Taos has six vowels with three contrastive vowel heights and degrees of vowel backness — five of these have an oral-nasal contrast, which persists even before a nasal consonant coda (i.e. the syllables CVN and CṼN contrast, where C = any consonant, V = any vowel, N = any nasal consonant).^[9]

Oral Vowels^[10]
	Front		Back
	oral	nasal	oral	nasal
High	i	ĩ	u	ũ
Mid	ɛ	ɛ̃	ɤ
Low	æ	æ̃	ɑ	ɑ̃

Vowels /i, u/ have lowered variants [ɪ, ʊ] in closed syllables and when unstressed.^[11]
The vowel /u/ has a variant [o] with very narrow lip rounding before /l, m, n/. The vowel /u/ has phonetic "inner rounding", which is presumably a type of lip compression.^[12]
The mid vowels /ɛ, ɤ/ are phonetically somewhat centralized: [ɛ̈, ɤ̈]. Although both are mid, /ɤ/ (upper-mid) is phonetically higher than /ɛ/ (lower-mid).
The mid-front vowel /ɛ/ predominantly occurs in suffixes.
Nasalized /ɛ̃/ is phonetically lower than its oral counterpart: [ɛ̞̃].
In unstressed syllables, mid /ɤ/ is phonetically a central vowel [ə]. In stressed syllables, it is phonetically mid-back [ɤ]. In contrast to the high back vowel /u/, mid /ɤ/ is unrounded. Unlike the other vowels, /ɤ/ has no nasal counterpart.
Nasalized /æ̃/ is phonetically a central vowel [ã]. (Oral /æ/, however, is phonetically front.)
The vowel /ɑ/ has a slightly rounded variant [ɒ] after labials /p, pʰ, p’, b, m, w/ and also before /p, pʰ, p’, b, m/. Before nasals /m, n/, this vowel is centralized: [ɑ̈] (before /n/), [ɒ̈] (before /m/). Before glide /j/, the vowel is raised, approaching [ʌ].^[13]
Nasalized /ɑ̃/ is phonetically slightly rounded and higher than its oral counterpart: [ɔ̃].^[14]
Vowel length allophony:
1. The duration of vowels varies according to stress. Vowels in syllables with primary stress are relatively long and somewhat shorter in syllables with secondary stress. Unstressed syllables have short vowels.
2. The presence of a consonant coda also affects vowel length. Vowels are short in closed syllables (but not as short as unstressed syllables) and long in open syllables.
3. There is also an interaction between tone and vowel length. Vowels with a mid tone are long while with a low tone are "pulsated". Trager mentions further interaction but does not report the details.^[15]

[edit] Diphthongs

In addition to these monophthongs, Taos has five (native) vowel clusters (i.e. diphthongs) that can function as syllable nuclei and are approximately the same duration as the single vowels:

/iɛ, iæ, ĩɛ̃, uɑ, ɤɑ/

Unlike diphthongs in several other languages, each component of the vowel cluster has an equal prominence and duration (i.e. there are no offglides or onglides). The cluster /iɛ/ is rare. Additionally, the vowel cluster

/uɛ/

is found in Spanish loanwords.

The monophthongs can be followed by high front and high back offglides, but these are analyzed as glide consonants in a coda position. Trager notes that in these sequences the glides are not as prominent as the vowel nuclei but that the difference is not very marked, and, in fact, Harrington (1910) describes these as diphthongs on par with Trager's "vowel clusters". The following vowel + glide sequences are reported in Trager (1946):

Vowel nucleus	/j/ offglide	/w/ offglide	Vowel nucleus	/j/ offglide	/w/ offglide
/i/	–	iw	/u/	uj	–
/ĩ/	–	ĩw	/ũ/	ũj	–
/ɛ/	–	–	/ɤ/	ɤj	ɤw
/ɛ̃/	ɛ̃j	–
/æ/	æj	æw	/ɑ/	ɑj	ɑw
/æ̃/	æ̃j	æ̃w	/ɑ̃/	ɑ̃j	–

[edit] Reduplicative patterning

Noun stems that end in a vowel have a suffixation-reduplication process in absolute forms that attaches a glottal stop /ʔ/ and a reduplicant consisting of a reduplicated stem-final vowel to the noun stem (which is, then, followed by an inflectional suffix):^[16]

STEM- + -ʔV- + -SUFFIX (where V = a reduplicated vowel)

If the stem-final vowel is an oral vowel, the reduplicated vowel is exactly the same as the stem vowel. However, if the stem-final vowel is nasal, the nasality is not copied in the reduplicant — that is, the nasal vowel will be reduplicated as that vowel's oral counterpart. For example, the stem /pɑ̃-/ appears with the reduplicated vowel /ɑ/ in the form

/ˌpɑ̃̄ˈʔɑ̄nɛ̄/ "earth (nonsingular)" (< pɑ̃- + -ʔɑ- + -nɛ)

In stems that end in a vowel cluster, only the second vowel of the cluster is reduplicated:

/ˌʔìæˈʔǣnɛ̄/ "corn (nonsingular)" (< ʔìæ- + -ʔæ- + -nɛ, stem: ʔìæ-)

[edit] Prosody

[edit] Stress

Please help improve this section by expanding it.
Further information might be found on the talk page or at requests for expansion.

Trager analyzes Taos as having three degrees of stress:^[17]

primary
secondary
unstressed

Trager describes Taos stress in terms of loudness; however, he also notes in several places where stress has effects on vowel length.

All words must have a single primary stress. Polysyllabic words can, in addition to the syllable with primary stress, have syllables with secondary stress, unstressed syllables, or a combination of both unstressed and secondarily-stressed syllables.

Trager (1946) states that the primary and secondary stress levels are in complementary distribution in low-toned and high-toned syllables. However, his later analysis rejects this.

[edit] Tone

Taos three tones:^[18]

high (symbol: acute accent ´)
mid (symbol: macron ¯)
low (symbol: grave accent `)

Trager describes the tones as being distinguished by pitch differences. The mid tone is by the most commonly occurring tone; high tone is limited to a few stems and suffixes; the low tone is relatively common in stem syllables. The high tone is described as "higher and sharper" than the mid tone while the low tone is "distinctly lower and drawling".

There is no tonal contrast in unstressed syllables, which have only mid tones. Trager (1946) found the stress level to be predictable in syllables with high and low tones; however, Trager (1948) finds this to be error with the addition of new data and different theoretical outlook. (See stress section above.)

[edit] Syllables and phonotactics

The simplest syllable in Taos consists of a single consonant in the onset (i.e. beginning consonant) followed by a single vowel nucleus, i.e. a CV syllable. An onset and nucleus are obligatory in every syllable. Complex onsets consisting of a two-consonant cluster (CC) are found only in loanwords borrowed from New Mexican Spanish. The nucleus can have optionally two vowels in vowel clusters (V or VV). The syllable coda (i.e. the final consonants) is optional and can consist of up to two consonants (C or CC). In other words, the following are possible syllable types in Taos: CV, CVV, CVC, CVVC, CVCC (and in loanwords also: CCV, CCVV, CCVC, CCVVC, CCVCC, CCVVCC). This can be succinctly represented in the following (where optional segments are enclosed in parentheses):

C₁(C₂)V₁(V₂)(C₃)(C₄) + Tone

Additionally, every syllable has a tone associated with it. The number of possible syllables occurring in Taos is greatly limited by a number of phonotactic constraints.

Onsets. A single onset C₁ can be filled by any Taos consonant (except the borrowed /f/) — that is, /p, p’, pʰ, b, m, w, t, t’, tʰ, d, n, l, ɬ, tʃ, tʃ’, s, ɾ, j, k, k’, kʷ, kʷ’, g, x, xʷ, ʔ, h/ are possible onsets. The single onsets /b, d, ɾ, g/ are only found in Spanish borrowings. In a loanword two-consonant C₁C₂ cluster, C₁ can be filled only by voiceless stops /p, t, k, f/ while C₂ can be filled only by /ɾ, l/ in the following combinations:^[19]

	C₂
C₁	ɾ	l
p	pɾ	pl
t	tɾ	–
k	kɾ	kl

Of the onsets, /p, p’, pʰ, t, t’, tʰ, tʃ, tʃ’, k, k’, kʷ, kʷ’, ʔ, ɬ, x, xʷ, h, pɾ, pl, tɾ, kɾ, kl, fɾ/ can only occur as onsets (and not as codas).

Rimes. Within the syllable rime, any single Taos vowel — /i, i, ĩ, ɛ, ɛ̃, æ, æ̃, ɑ, ɑ̃, ɤ, u, ũ/ — may occur in the nucleus. In complex nuclei consisting of vowel clusters, the following combinations are possible:

Vowel nucleus clusters
	Final component
Initial component	ɛ	ɛ̃	æ	ɑ
i	iɛ	–	iæ	–
ĩ	–	ĩɛ̃	–	–
u	uɛ ^†	–	–	uɑ
ɤ	–	–	–	ɤɑ

† - only in loanwords

The /uɛ/ cluster was found only in a single word /ˈpūɛlɑ̄næ̃̄/ "frying pan" (from Spanish puela probably from French poêle).^[20]

A subset of Taos consonants consisting of voiced stops and sonorants — /b, d, g, m, n, l, ɾ, w, j/ — can occur in coda C₄ position. There is a restriction that high vowels cannot be followed by a homorganic glide (i.e., /ij, uw/ do not occur). Not all VC combinations are attested. The attested sequences of V + glide are listed in the vowel diphthong section above. Additionally, /s/ may appear in coda position in loanwords.^[21]

In complex two-consonant C₃C₄ codas, Trager (1946) states that the final consonant C₄ can consist of a voiced stop /b, d, g/ and be preceded by a consonant C₃ consisting of a non-liquid sonorant /m, n, w, j/.^[22] However, Trager (1948) states that the following are the only attested coda clusters:

/ng, lg, jg/

Trager does not discuss the combinatory possibilities between segments and tones, although he does for stress and tone.

[edit] Intonation

Please help improve this section by expanding it.
Further information might be found on the talk page or at requests for expansion.

[edit] Loanword phonology

Trager (1944) indicates the type of phonetic/phonological changes that New Mexican Spanish loanwords undergo when being adapted to the Taos language.^[23] Different degrees of nativization occur in Spanish loanwords: earlier borrowings have greater differences while later borrowings (borrowed by speakers who are probably increasingly bilingual) have much greater similarity with the Spanish forms. The chart below lists some of the correspondences. The inflected nouns in the table are in the absolute singular form with the inflectional suffix and any reduplicant separated from the initial noun stem with hyphens.

New Mexican Spanish phoneme(s)	Taos phoneme(s)	Example
		Taos		Spanish
		word	gloss	Spanish
/b/ [b] (initial)	/m/	ˌmūlsɑ̄-ˈʔɑ̄-nǣ	"pocket"	< bolsa
/b/ [β] (intervocalic)	/b/	ˈxʷǣbǣsī	"Thursday"	< jueves
/d/ [d] (initial)	/t/	tūˈmĩ̄kū	"week"	< domingo
/d/ [ð] (intervocalic)	/l/	ˈsɑ̄bɑ̄lū	"Saturday"	< sábado
/d/ [ð] (intervocalic)	/d/ (after Taos /l/)	tɛ̄lɛ̃̄ˈdūnɛ-nɛ̄mæ̃̄	"fork"	< tenedor
/dɾ/ [ðɾ]	/jl/	kūˌmǣjlī-ˈʔī-nǣ	"one's child's godmother"	< comadre
/f/	/pʰ/	kɑ̄ˈpʰǣ-nɛ̄	"coffee"	< café
/x/ [h]	/h/ (initial)	ˈhǣlgɑ̄-næ̃̄	"rug"	< jerga
/x/ [h]	/x/ (intervocalic)	ˈmīlxīnæ̃̄	"the mother Virgin"	< virgen
/xu/ + V ([hw] + V)	/xʷ/ + V	ˈxʷǣbǣsī	"Thursday"	< jueves
/nd/ [nd]	/n/	sūˌpɑ̄nɑ̄-ˈʔɑ̄-nǣ	"bed spring"	< sopanda
V + /ng/ [Ṽng]	/Ṽk/	tūˈmĩ̄kū	"week"	< domingo
/ɲ/	/j/	kǣˈjūn-ɛ̄-nǣ	"canyon"	< cañón
/ɾ/	/l/	ˈlǣj-nǣ	"king"	< rey
/r/	/ld/	ˌmūldū-ˈʔū-nǣ	"donkey"	< burro
/i/ [i]	/i/	mɑ̄lˌtījū-ˈʔū-nǣ	"hammer"	< martillo
/i/ [j] (before V)^[24]	/j/	ˌjǣwɑ̄-ˈʔɑ̄-nǣ	"mare"	< yegua
/ie/ [jɛ]	/iæ/	ˈmīælnæ̃̄sī	"Friday"	< viernes
/u/ [u]	/u/	ˈlūnæ̃̄sī	"Monday"	< lunes
/u/ [w] (before V)	/w/	ˌjǣwɑ̄-ˈʔɑ̄-nǣ	"mare"	< yegua
/e/ [ɛ] (stressed)^[25]	/æ/	ˌwǣltɑ̄-ˈʔɑ̄-nɛ̄	"garden"	< huerta
/e/ [e] (unstressed)^[26]	/i/	ˈmīælkūlīsī	"Wednesday"	< miércoles
/o/	/u/	ˌtūlū-ˈʔū-nǣ	"bull"	< toro
/a/ [ɑ]	/ɑ/	ˌmūlsɑ̄-ˈʔɑ̄-nǣ	"pocket"	< bolsa

Although NM Spanish /a/ is usually borrowed as Taos /ɑ/, it is nativized as /æ/ when it precedes the Taos glide /j/, which is the nativization of NM Spanish /d/ in the cluster /dɾ/ (/dɾ/ > Taos /jl/). Because Taos /ɑ/ when followed by /j/ is typically raised (i.e. /ɑj/ is phonetically [ʌj]), Taos /æ/ is phonetically a closer match to NM Spanish low /a/. Thus, NM Spanish compadre is borrowed as /kūmˌpǣjlī-ˈʔī-nǣ/ "one's child's godfather (absolute, singular)" (with /adɾ/ [ɑðɾ] > /æjl/).

Taos /æ/ is a better match than /ɛ/ for NM Spanish /e/ (phonetically [ɛ]) because Taos /ɛ/ is restricted to affixes in native Taos words.

Another common process is the insertion of /i/ after in New Mexican Spanish words ending in /s/, as native words in Taos cannot have syllables ending in /s/.

The other NM Spanish phonemes are nativized as similar phonemes in Taos: NM Spanish /p/ > Taos /p/, NM Spanish /t/ > Taos /t/, NM Spanish /g/ > Taos /g/ (but see above for NM Spanish sequence /ng/), NM Spanish /ʧ/ > Taos /tʃ/, NM Spanish /s/ > Taos /s/, NM Spanish /m/ > Taos /m/, NM Spanish /n/ > Taos /n/, NM Spanish /l/ > Taos /l/.

Later borrowing, which has been subject to less alteration, has led to the development of /ɾ/, word-initial voiced stops /b, d, g/, syllable-final /s/, and consonants clusters /pl, pɾ, tɾ, kɾ, kl, fɾ/. The word-internal cluster /stɾ/ is reduced to /st/ in Taos, as in NM Spanish maestro > Taos /ˌmɛ̄stū-ˈʔū-nǣ/ "teacher" — the cluster was reduced further to just /t/ as /ˌmɛ̄tū-ˈʔū-nǣ/ in one speaker, a reflection of the older pattern where /s/ cannot be syllable-final.

[edit] See also

Taos language

[edit] Notes

^ de Angulo's work includes an unpublished grammatical sketch and analyzed texts (housed at the American Philosophical Society).
^ The main difference with respect to consonants in the 1948 analysis is that aspirated consonants /pʰ, tʰ/, ejectives /p’, t’, tʃ’, k’/, and labialized consonants /kʷ, kʷ’, xʷ/ were considered consonant sequences, respectively /ph, th, pʔ, tʔ, tʃʔ, kʔ, kw, kʷʔ, xw/. Similar alternate analyses of C + ʔ vs. C’ have been proposed in other languages in region, e.g., Zuni syllable-initial consonants (see: Zuni consonants). This difference affects his analysis of syllable structure and phonotactics. The sequences /pʔ, tʔ, kʔ/, then, contrast with /bʔ, dʔ, gʔ/ sequences. Intervocalically, /pʔ, tʔ, kʔ/ are interpreted as syllable onsets while /bʔ, dʔ, gʔ/ are split by syllable boundaries /b.ʔ, d.ʔ, g.ʔ/. Trager admits that this analysis may not be ideal for the labialized segments since /kwʔ/ would be the only native three-consonant cluster that appears in syllable onsets (other three-consonant clusters appear in Spanish loanwords). Even so, he rejects a phoneme sequence /kʷʔ/ because the exclusion of loanword phonology is "very bad methodology". Unlike his decomposition of other Trager (1946) consonants, Trager (1948) retains /tʃ/ as a unitary segment. A final difference is the interpretation of the cluster /fɾ/ found in loanwords.
^ Earlier borrowings with Spanish [ɾ] were nativized as Taos /l/. (Note that the Spanish trill [r] was nativized as a Taos sequence /ld/.)
^ Trager (1948), which does not exclude loanword phonology, interprets the cluster as /phɾ/.
^ Harrington (1910) notes the glottalized stops have weak glottalic releases like the Georgian language or Mayan languages but unlike other strongly ejective releases found in other American languages (such as Kiowa).
^ Note that there is only a contrast between voiced and voiceless stops in word-initial position. If the Spanish borrowings are excluded, the voiced and voiceless stops are in complementary distribution and only a single plain stop series may be posited with predictable voiced allophones between vowels and voiceless allophones elsewhere. Additionally, although voiced stops are reconstructed for Proto-Kiowa-Tanoan, all reconstructed voiced stops have merged with other consonants in Taos in stem-initial position.
^ That is, there is /k/ - /kʷ/ neutralization, a /k’/ - /kʷ’/ neutralization, and a /x/ - /xʷ/ neutralization.
^ There is some discrepancy between the descriptions of this variation. Trager (1946) states that /tʃ’/ has a post-alveolar articulation [tʃ’] before front vowels and an alveolar articulation [ts’] elsewhere; however, Trager (1948) states "c and c’ are more often [č] and [č’] than [tˢ] and [tˢ’]...". And in his short description in Trager (1943), he states: "... c is [č] in Taos before i, u, į, ų, and [ts] elsewhere...".
^ Harrington (1910) does not note this oral-nasal contrast in the context of a following nasal stop.
^ The IPA symbols used here are equivalent to the following phonemic symbols of Trager: i = Trager i, u = Trager u, ɛ = Trager e, ɤ = Trager ə, æ = Trager a, ɑ = Trager o.
^ The general tendency is for unstressed vowels to move away from the periphery of the vowel space.
^ Trager (1946): "... the rounding is often more what might be called an inner rounding than one caused by the kind of puckering of the lips found in European language u-vowels...".
^ Trager (1944: 152)
^ Trager (1946) remarks that this vowel is rather similar to the American English vowel in the word hunt.
^ Specifically Trager (1946) states:

"All Taos vowels are long in free syllables when loud-stressed [= Wikipedia primary stress], and are rather short in weak-stressed [= Wikipedia unstressed] syllables. Medial-stressed [= Wikipedia secondary stress] vowels are slightly shorter than loud-stressed ones. Vowels with normal tone [= Wikipedia mid tone] and loud or medial stress are usually monophthongal longs, those with low tone are pulsated (‘reduplicated’) whether the stress is loud or quiet (thus t‘ˈə̀t‘o “by the day” is [t‘ˈə̀ᵊt‘α] and t‘ˌə̀ʔˈəna “day” is [t‘ˌə̀ᵊʔˈə·na]); vowels with medial stress and high tone are rather short, those with loud high accent are longer. There are differences in the length of loud-stressed vowels depending on the vowel of the following syllable, but it would lead into too much detail to go into them in a limited description such as the present one. The exception to the length of loud-stressed vowels is when they are followed by a plain stop, especially p, t, k, when the stop is long and ambisyllabic and the vowel is quite short. In checked syllables the vowels are always short, but less so with the low tone than otherwise, and never as short as weak-stressed vowels."

Trager's symbols:
- the grave accent < ` > represents either a low tone and primary stress or a low tone and secondary stress phonemically and (apparently) represents low tone phonetically;
- the beginning single quote < ‘ > represents aspiration;
- the upper vertical line < ˈ > represents a mid tone and primary stress phonemically and (apparently) primary stress phonetically;
- unmarked vowels are unstressed (and phonetically mid-toned);
- the lower vertical line < ˌ > represents mid tone and secondary stress phonemically and (apparently) secondary stress phonetically;
- the middle dot < · > represents long vowel length.
Although Trager is obviously using the symbols with different values in phonemic and phonetic representation, he does not ever explicitly state their phonetic values (apparently stress and tone are represented with separate symbols). In the phonemic representations, the symbols represent a combination of tone and stress. The acute accent has two possible tone-stress combinations because Trager found these to be in complementary distribution; however, in his later analysis (1948) he finds this to be false.
^ Martin Haspelmath refers to this type of element with the term duplifix.
^ Trager uses the following terminology: loud (= Wikipedia primary), normal or quiet (= Wikipedia secondary), weak (= Wikipedia unstressed).
^ Trager uses the term medial instead of mid.
^ Phoneme /f/ is actually excluded from Trager (1946) because it occurs only in loanwords and from Trager (1948) because it is analyzed as /ph/.
^ (The usual development of Spanish [we] is into Taos cluster /uɑ/, as in /ˌmūɑjǣ-ˈʔǣ-nǣ/ "ox" from < buey.
^ This was not noted in Trager (1946).
^ Trager (1946) uses the term sonant to refer to this non-liquid sonorant class, while the term sonorant refers to the usual natural class.
^ Further details on New Mexican Spanish are in Trager & Valdez (1937).
^ Trager analyzes New Mexican Spanish as not having /w, j/ phonemes (unlike Castilian Spanish). Thus, a word like martillo is phonemically /martíio/, buey(e) /buéie/, cien /sién/, yegua /iéua/.
^ One exception to the regular pattern is stressed NM Spanish /e/ > Taos /ɛ/ in /ˈsíɛnæ̃̄/ "hundred" < cien (cf. Castilian ciento)
^ Exceptions to the regular pattern are unstressed NM Spanish /e/ > Taos /æ/, as in /ˌmūɑjǣ-ˈʔǣ-nǣ/ "ox" < buey [ˈbwɛje], and /e/ > /ɤ/ in /ˈmɑ̄ltɤ̄sī/ "Tuesday" < NM Spanish martes.

[edit] Bibliography

Hale, Kenneth L. (1967). Toward a reconstruction of Kiowa-Tanoan phonology. International Journal of American Linguistics, 33 (2), 112-120.
Harrington, J. P. (1909). Notes on the Piro language. American Anthropologist, 11 (4), 563-594.
Harrington, J. P. (1910). An introductory paper on the Tiwa language, dialect of Taos. American Anthropologist, 12 (1), 11-48.
Nichols, Lynn. (1994). Vowel copy and stress in Northern Tiwa (Picurís and Taos). In S. Epstein et al. (Eds.), Harvard working papers in linguistics (Vol. 4, pp. 133-140).
Trager, George L. (1936). The language of the pueblo of Taos. Maître Phonétique, 56, 59-62.
Trager, George L. (1939). The days of the week in the language of Taos Pueblo, New Mexico. Language, 15, 51-55.
Trager, George L. (1942). The historical phonology of the Tiwa languages. Studies in Linguistics, 1 (5), 1-10.
Trager, George L. (1943). The kinship and status terms of the Tiwa languages. American Anthropologist, 45 (1), 557-571.
Trager, George L. (1944). Spanish and English loanwords in Taos. International Journal of American Linguistics, 10 (4), 144-158.
Trager, George L. (1946). An outline of Taos grammar. In C. Osgood (Ed.), Linguistic structures in North America (pp. 184-221). New York: Wenner-Green Foundation for Anthropological Research.
Trager, George L. (1948). Taos I: A language revisited. International Journal of American Linguistics, 14 (3), 155-160.
Trager, George L.; & Valdez, Genevieve. (1937). English loans in Colorado Spanish. American Speech, 12 (1), 34-44.