6 INSTITUTE OF SOCIAL ANTHROPOLOGY—-PUBLICATION NO. 8 
3.4. eand e are mid, front, unrounded. In most 
cases rather open, they become close when pre- 
ceding y. 
dem urine, hé-pe tree gourd, dapméyfa? ocean (archaic 
form), ?éya also 
3.5. a and a‘ are low, central, unrounded, and 
constant in value. They have the approximate 
value of a in German “Mann.” 
pak bone, ?4-i uncle, é&hka deaf mute 
3.6. uand wu are high, back, moderately close, 
rounded. They are somewhat more open preced- 
ing a nasal. 
mu’k grass, huki cigarette, tum one 
3.7. A and a’ are central, tending to back, open, 
unrounded, slightly higher than double o in 
English ‘‘took.”’? The lip position is more spread 
than for any English vowel. 
tak house, nA? water, wastén two 
3.8. o and o- are mid, back, semirounded. In 
free syHables they are rather close unless in nexus 
with k, when they are open; in checked syllables 
they tend to be open unless in nexus with y when 
they are close. 
?6ma fog, 6°mo woman, s6’ki snail, k6°so knee, hos cave, 
*ho't scratch, gdgoy spirit, h6yma tomorrow 
3.9. Vowel distribution. There are no initial 
vowels in syllables. 
All vowels may occur in word medial and final 
positions. 
There are no vowel clusters. When through 
morphological processes two vowels fall in con- 
tiguous positions, an h arises to maintain vocalic 
identity. 
hihi roar (*hu: roar + -i nom. suf.) 
yo'Sahi work (*yo'Sa° work +- -i) 
ya?éahi necessity (*ya?éa° suffer + -i) 
yoSiha work! (*yo°S4° work + -a imp. suf.) 
3.10. Limitations between consonants and vow- 
els in syllables. 
With the consonant in syllable initial position: 
d does not precede i; d, 1, and r have not been 
observed preceding e; | and r have not been 
observed preceding a. 
With the consonant in syllable final position: 
¢, s, y, 1, and r have not been observed following 
i; t, é, w, r, and | have not been observed following 
e; t, 8, and w have not been observed following 
a; | and r have not been observed following 0; 
i, 1, and r have not been observed following u; 
t, fi, w, l, and r have not been observed following a. 
Consonantal clusters and the glottal stop do not 
follow long vowels. 
With rare exceptions which seem to follow no 
rule, the alveolars are not found in nexus with 1 or 
y; their place is taken by the palato-alveolar forms. 
worni girl, éikin a skin disease (Sp. ‘‘pinto’’), 
tif excrement, *?it be, SiS bull, éf-éa cradle ® 
Examples of alveolars in nexus with i are as 
follows: makti a supernatural female wood spirit, 
titi civilized person, (Sp. “gente de razon’), 
Sikuhti a type of bee, ma?ksi beforehand, figimat 
enchanted place. Spanish loan words often show 
alveolars in nexus with i, e. g., anfmat animal, 
sintirén cinturon, asi asi. 
4. MORPHOPHONEMICS 
4.1. On the morphophonemie level Popoluca is 
characterized by extensive sandhi phenomena. 
Most of this is of the type known as internal or 
word sandhi, i. e., morphophonemic changes 
which occur when grammatical elements are com- 
bined into words. Jn some cases the sandhi 
phenomena are also external, or phrasal in nature, 
i. e., the morphophonemic changes occur both 
when grammatical elements are combined into 
words and when words are combined into phrases. 
The sandhi processes most commonly used in 
Popoluea are metathesis, syllabic division changes, 
palatalization, consonant changes, consonant syn- 
cope, vocalic changes, and vocalic syncope. The 
strictly internal phenomena are first discussed, 
and then those that are also external. 
INTERNAL SANDHI 
4.2. Changes in syllabic division. Certain mor- 
phological and phonological processes cause 
changes in the syllabic division of many words. 
Thus, when morphemes of the pattern -V, -VC, 
6 When fi (or n[?]) follows i preceding a palato-alveolar, or when it is the 
only slightly palatalized, but may vary from almost no palatalization to com- 
plete palatalization. This phonetic range causes some question as to how 
p in such cases should be interpreted phonemically. Since i preceding an 
alveolar tends to be very weakly palatalized, and in some cases not at 
all (4.5), it appears that with respect to palatalization fi is phonetically 
“weaker” than the other palato-alveolar forms, to which some of the palataliz- 
ing quality assimilates. Hence, it is assumed that (n]~[fJ is /fi/. 
nénéi/fi/ varies from nénéi[n} to nanéi |i} 
pA‘si/fi/ varies from pA-si[n] to pA‘si[f] 
ti/fi/ varies from ti{n] to tiff} 
k{/i/éi varies from k{[n]éi to ki[a]éi 
