Adelaar Willem, Muysken Pieter. The languages of the Andes
Подождите немного. Документ загружается.
314 3 The Inca Sphere
-sa become unaspirated, even when separated from -sa by another morpheme (Hardman
1966: 99).
(227) kawki-ta-w-ta-sa
{kawki-t
(h)
a-w(a)-ta-sa}
what.place-AB-CV-2S-IR
‘Where are you from?’
(Hardman 1966: 98)
The additive suffix -psa is reminiscent of Quechua -pas.Italso has similar functions.
Along with its usual meaning ‘also’, ‘too’, it is used with interrogative roots to indicate
indefiniteness.
(228) qaˇci-psa
who-AD
‘whoever’, ‘anyone’
(Belleza Castro 195: 135)
(229) qaˇci-psa hal-ur-p-t
y
i
{qaˇci-psa hal(a)-ur(u)-p(
h
a)-t
y
i}
who-AD fall-IW-3S.IM-NE
‘Let no one enter!’
(Belleza Castro 195: 135)
Other independent suffixes are -aˇsi ‘maybe’, -ha ‘of course!’, ‘now I see!’, -il
y
i
‘emphatic’, -iˇsi ‘I remember’, -kasa ‘already’, -k
h
a ‘furthermore’, -mna ‘they say’, -ra
‘still’, and -sk
h
a ‘once again’. The suffixes -aˇsi and -iˇsi are only found after other
independent suffixes (which then lose their final vowel). The suffix -k
h
a is frequently
found after -r(a)- (230).
(230) mam-
ŋŋa sa-w-utu-r-k
h
a-wa “iqu, hani ma-ta-t
y
ik
h
uw-ha”
{mam(a)-
ŋa sa-w-utu-r(a)-k
h
a-wa “Iqu, hani ma-ta-t
y
ik
h
uw(a)-ha”}
mother-1P say-PV-3S.1O-AN-AD-AF girl, don’t go-2S-NE
that.place-AC
‘My mother said to me furthermore: “Girl, don’t go there!” ’
(Belleza Castro 1995: 158)
The Jaqaru lexicon has undergone a considerable influx of Quechua loans, most of
which are from the central Peruvian dialects. An example of Quechua influence are the
numerals. In accordance with the rules of Aymaran word structure, the Jaqaru numerals
for ‘seven’, ‘eight’ and ‘nine’ present an added final vowel, which is not found in Quechua
and which betrays the latter as the lending language.
(231) qanˇc
.
isi ‘seven’ Central Peruvian Quechua: qanˇc
.
is
pusaqa ‘eight’ Quechua: pusaq
isqu˜na ‘nine’ Quechua: isqun
3.3 The Aymaran language family 315
The other Jaqaru numerals are either cognate to the Aymara ones (maya ‘one’,
paha ‘two’, puˇsi ‘four’), or shared with both Aymara and Quechua (kimsa ‘three’,
piˇcqa ‘five’, suhta ‘six’, ˇc
.
u
ŋ
ka ‘ten’, paˇc
.
aka ‘hundred’, wara
ŋ
qa ‘thousand’). Units
are added to tens with the possession marker -ni,asinAymara; e.g. ˇc
.
u
ŋ
k-maya-ni
‘eleven’ (< 10 + 1). Complex numbers are subject to different sorts of internal vowel
suppression.
An interesting phenomenon is the presence of remnants of sound-symbolism, which
is also visible in Quechua loans. In the word t
y
ahˇsa ‘small’ (from Central Peruvian
Quechua takˇsa) and in ut
y
ut
y
ul
y
qu ‘goblin’ (compare Central Peruvian Quechua uˇcuk
ul
y
qu ‘little man’) the palatal character of the alveopalatal stop appears to indicate small-
ness, as is also the case in other sorts of palatality in the neighbouring Quechua dialects
(cf. section 3.2.5).
In addition to basic vocabulary shared with other Aymaran languages, there are also
a substantial number of words which are exclusive to Jaqaru (and possibly Cauqui).
Examples are karma(ha) ‘man’ (Aymara ˇcaˇca), uhara ‘maize’ (Aymara tunqu), wasa-
‘to go’ (Aymara sara-), il
y
a- ‘to see’ (Aymara un
y
a-), palu- ‘to eat’ (Aymara manq’a-),
and many others. The interrogative pronouns in Jaqaru present considerable differences
vis-`a-vis their Aymara counterparts, as in qaˇci ‘who’ (Aymara k
h
iti), quwa ∼ qusa
‘what’ (Aymara kuna), qaˇcwira ‘which’ (Aymara kawk(n)i:ri), alongside similarities, as
in qamiˇsa ‘how’ (Aymara kamisa).
3.3.11 Jaqaru sample text
The following animal story was collected by J. M. B. Farf´an in the late 1940s from a
storyteller called Vicente Casanova (Farf´an 1952: 80–1). It is entitled ‘The guinea-pig
and the fox.’ Farf´an collected Jaqaru narratives long before the first modern study of
the language appeared. Like most early observers of the Jaqaru language, he did not
succeed in rendering the complex sounds of the language consistently. In what follows,
we shall try and supply a modernised version which follows the original text as faithfully
as possible.
1. miˇs-uru-wa tiŋŋku-w-i atuqu k’uyt
y
u-wˇsqa may qa:q-na
{miˇs-uru-wa ti
ŋku-w-i atuqu k’uyt
y
u-wˇsqa may(a) qa:q(a)-na}
one-day-AF meet-PV-3S fox guinea.pig-IS one rocky.peak-L
‘One day a fox and a guinea-pig met on a rocky peak.’
The element miˇs-isacombinatorial variant of maya ‘one’ which is used in a compound
with uru ‘day’; when preceding a noun as a modifier, maya is realised as may by vowel
suppression. The long vowel in qa:qa (from Quechua qaqa)isour rendering of a long
vowelinBelleza Castro (1995) and a plain vowel in Hardman (1983a); the last vowel of
316 3 The Inca Sphere
qa:qa is lost before the locative marker (-na)infinal position. The root ti
ŋ
ku-isaloan
from Quechua.
2. atqu-wa katu-w-i k’uyt
y
-
h
a sa:ma-t
h
a pal-nuˇs-p-taki
{at(u)qu-wa katu-w-i k’uyt
y
(u)-ha sa:ma-t
h
a pal(u)-nuˇs(u)-p(
h
a)-taki}
fox-AF catch-PV-3S guinea.pig-AC back-AB eat-FN-3P-B
‘The fox caught the guinea-pig by the back in order to eat him.’
Concerning pal-nuˇs-p-taki:any suffix loses its vowel and aspiration before -taki;
-nuˇsu loses its final vowel before the person marker -p
h
a; root-final vowels are lost
before -nuˇsu.
3. k’uyt
y
u-qa sa-w-i atuq-
h
a “hani hayˇc-k
h
-uta-t
y
i
{k’uyt
y
u-qa sa-w-i atuq(u)-ha hani hayˇc(a)-k
h
(a)-uta-t
y
i}
guinea.pig-TO say-PV-3S fox-AC don’t beat-M-2S.1O-NE
‘The guinea-pig said to the fox: “Don’t kill me!”’
The combination hayˇc-k
h
(a)- (lit. ‘to beat suddenly’) has the meaning ‘to kill’. Farf´an
uses the form atuxa,which we have interpreted as atuq-
h
a; the accusative marker -ha
merges with the last consonant in the root after suppression of the latter’s final vowel,
resulting into an aspirated -q
h
-.
4. pan-sa hira aq
h
i-ru
{pan(i)-sa hira aq
h
i-ru}
two.people-4P let’s.go cave-AL
‘“Let us go the two of us to a cave!’
The form pani is a special form for ‘two people’, from paha ‘two’ and possibly the
possession marker -ni;itloses its final vowel before the person marker -sa. The form
hira functions syntactically as a verb form, hence the complement in -ru.
5. k
h
uwa-n-k-i-wa yaŋŋ-ŋŋayak-nuˇsu-ŋŋ-taki k’uyt
y
u ANTCHAQAXA”
{k
h
uwa-n(a)-k(a)-i-wa yaŋ(a)-ŋa yak(a)-nuˇsu-ŋ(a)-taki k’uyt
y
u
}
that.place-L-LV-3S-AF companion-1P give-FN-1P-B guinea.pig very
more. robust-AC
‘There are my companion(s), so that I can give you a more robust
guinea-pig.”’
Our subdivision of this sentence differs from that in Farf´an, who has a comma after
k
h
uwa-n-k-i-wa;itseemed more logical to connect this form with ya
ŋ
-
ŋ
a, than with the
3.3 The Aymaran language family 317
word referring to the cave. As we have seen, the element -k- can either be interpreted as a
locative verbaliser -k(a)-,asamarker of simultaneousness -k-,orasasequence of both;
here we have opted for the first solution. The nominaliser -nuˇsu retains its final vowel
before the first person possessive marker -
ŋ
a. The form transcribed as antchaqaxa and
translated as ‘m´as fuerte’ (‘more robust’) by Farf´an could not be analysed. It may either
contain the elements anca ‘much’ and -ˇc
.
aqa ‘more’, or the word anˇc
.
aˇc
.
i ∼ anˇc
.
h
aˇc
.
h
i
‘(too) many’, followed by the accusative marker -ha (as suggested by Cerr´on-Palomino,
personal communication). In either case the quantifier would have been extracted from
the noun phrase.
6. atqu-qa anc’-ˇsu-w-i k’uyt
y
-
h
a katu-t-
h
airp-nuˇs-p-taki qunca-p-n
uk-
h
a
{at(u)qu-qa anc’(a)-ˇsu-w-i k’uyt
y
(u)-ha katu-t(a)-ha irp(a)-nuˇs(u)-
p(
h
a)-taki qunca-p(
h
a)-n(a) uk(a)-ha}
fox-TO let.go-OW-PV-3S guinea.pig-AC seize-SN-AC lead-FN-3P-B
brother.of.male-3P-G there-AC
‘The fox set free the guinea-pig that had been captive, so that he might
lead him to the place where his brothers were.’
The form katu-t-
h
a is here interpreted as accusative, but an ablative interpretation
(katu-t-t
h
a ‘from a captive state’) would also make sense. The genitive–locative marker
-na in final position suppresses the vowel and the aspiration of the possessive third person
marker -p
h
a;thevowelof-na itself is suppressed by the initial vowel of the root uka
that follows it; the combination -n(a) uka refers to ‘the place of’; uka with a short (or
extra-short) vowel refers to a place (‘there’).
7. u:ka-t-qa k’uyt
y
u-qa hal-ru-w-i uˇc
.
uŋŋsa-ru haypta-w-i
{u:ka-t(
h
a)-qa k’uyt
y
u-qa hal(a)-(u)ru-w-i uˇc
.
uŋsa-ru haypta-w-i}
that-AB-TO guinea.pig-TO fall-IW-PV-3S hole-AL disappear-PV-3S
‘Then, the guinea-pig entered into a hole and disappeared.’
According to Hardman (1966: 7, 1983a: 131), the demonstrative pronoun ‘that’ is
u:ka (or uka if extra-short vowels are accepted), it being distinguished from uka ‘there’
by vowel length; Belleza Castro (1995) observes no length distinction between the
two demonstratives; instead of u:ka-t-qa, u:ka-t
h
a (without topic marker) would be an
alternative reading for Farf´an’s ukatxa. The derived verb hal-ru-isusually found as
hal-uru-; the interpretation ‘to enter’ betrays the original meaning of hala-, ‘to run’, still
found in Aymara. The verb haypta- can possibly be analysed as derived from haya ‘far’
318 3 The Inca Sphere
and -pta- ‘upward movement’, ‘to begin’; since this is not a productive formation with
noun roots, we treat haypta-asanunderived root.
8. atqu-qa haruqa-w-i naray-k-uˇsu
{at(u)qu-qa haruqa-w-i naray(a)-k-uˇsu}
fox-TO remain-PV-3S wait-SM-SS.SU
‘The fox remained waiting.’
The verb haruqa-isrecorded as harwaqa-inBelleza Castro (1995).
9. k’uyt
y
u anc haya haypta-w-i
{k’uyt
y
u anc(a) haya haypta-w-i}
guinea.pig very long.time disappear-PV-3S
‘The guinea-pig disappeared for a very long time.’
Note that anca (Quechua anˇca) belongs to the common lexicon of Aymaran and
Quechuan. The usual interpretation of haya is ‘far’, but it also has a temporal meaning.
10. u:ka-t
h
a anc’-ˇsu-w-i atuq naray-k-iri-ru qal-u˜na uk-sana-t
h
a
qa:qa-tha
{u:ka-t
h
a anc’(a)-ˇsu-w-i atuq(u) naray(a)-k-iri-ru qal(a)-u˜na
uk(a)-sana-t
h
a qa:qa-t
h
a}
that-AB let.go-OW-PV-3S fox wait-SM-AG-AL stone-DI
there-topside-AB rock-AB
‘Then, from a rock up there, he dropped some little stones on the fox,
whowas waiting.’
The suffix -u˜na suppresses the final vowel of the preceding root and expresses the
concept of a diminutive; as a full form, u˜na means ‘pup’ or ‘recently born animal’ both in
Jaqaru and in Central Peruvian Quechua. The spatial root -sana enters into composition
with a preceding noun and means ‘above’, ‘on top of’; when a case marker follows it,
-sana eliminates the preceding vowel; -sana is also found in Quechua (for instance, in
the Huarochir´ı manuscript; see sections 3.1 and 3.2.11).
11. k’uyt
y
u ar-k-i atuq-
h
a sa-ˇsu “qa:qa-wa hal-k-i”
{k’uyt
y
u ar(u)-k-i atuq(u)-ha sa-(u)ˇsu qa:qa-wa hal(a)-k-i}
guinea.pig call-SM-3S fox-AC say-SS.SU rock-AF fall-SM-3S
‘The guinea-pig called to the fox and said: “The rock is falling.”’
In the word sa-ˇsu ‘saying’, the root-vowel is preserved as a consequence of its minimal
CV shape; as after most roots, -uˇsu loses its preceding vowel; sa-ˇsu accompanies verbs
of communication in order to introduce direct speech.
3.4 The Mochica language 319
12. sa-k-i atuq-
h
a kat-nuˇs-p-taki qa:qa
{sa-k-i atuq(u)-ha kat(u)-nuˇs(u)-p-taki qa:qa}
say-SM-3S fox-AC hold-FN-3P-B rock
‘He tells the fox to hold the rock.’
13. u:ka-t
h
a k’uyt
y
u ma-w-i ay-iri l
y
uq’a ˇcahl
y
a kat-nuˇs-p-taki
{u:ka-t
h
a k’uyt
y
u ma-w-i ay(a)-iri l
y
uq’a ˇcahl
y
a kat(u)-nuˇs(u)-p(
h
a)-taki}
that-AB guinea.pig go-PV-3S carry.sticks-AG stick lathing
sustain-IS-3P-B
‘Then, the guinea-pig went to fetch sticks and ladders in order to
support it.’
Both Belleza Castro (1995) and Farf´an (1961) have l
y
uq’i and ˇcahra for l
y
uq’a and
ˇcahl
y
a, respectively; the word ˇcaql
y
a is also found in Aymara and in Quechua; its most
common meaning refers to the lathing used on roofs.
14. atqu-qa kat-k-uˇsu-wa haruqa-w-i qa:q-
h
a
{at(u)qu-qa kat(u)-k-uˇsu-wa haruqa-w-i qa:q(a)-ha}
fox-TO sustain-SM-SS.SU-AF stay-PV-3S rock-AC
‘The fox stayed behind, supporting the rock.’
After the present-tense marker -k- the subordinator -uˇsu retains its initial vowel.
Instead of haruqa-, Farf´an has haruyqa-. Note the position of qa:q-
h
a outside the scope
of the subordinate clause to which it grammatically belongs.
15. u:ka-t
h
a k’uyt
y
u ma-w-q
h
-i akiˇs-kama-ya
{u:ka-t
h
a k’uyt
y
u ma-w-q
h
(a)-i akiˇs(a)-kama-ya}
that-AB guinea.pig go-PV-RP-3S now-LI-LS
‘And the guinea-pig went away again until today.’
The affix -ya in akiˇs-kama-ya is limited to this particular form; it can be left out.
3.4 The Mochica language
Mochica is the only language of the Peruvian coastal region that has survived long enough
to become documented in a substantial way. Its linguistic area was centred around the
modern towns of Chiclayo and Lambayeque, and the historic town of Za˜na, in the coastal
plain of northern Peru. According to Fernando de la Carrera Daza (1644), parish priest
of Reque and the author of the only colonial grammar of the language that has been
preserved, there were also groups of Mochica speakers in the highlands east and north
of the nuclear region, in the modern departments of Cajamarca and Piura, including
a colony of mitimaes (see the introduction to this chapter) near the town of Balsas in
the Mara˜n´on river valley. Toward the end of the nineteenth century, the existence of the
320 3 The Inca Sphere
Mochica language had become reduced to two coastal villages in the neighbourhood of
Chiclayo: Eten and Monsef´u. When the German scholar Middendorf stayed in Eten in
the 1880s, he still had the opportunity to work with bilingual, as well as monolingual
speakers. About the middle of the twentieth century it was no longer possible to obtain
more than fragmentary data from semi-speakers. At present, at the beginning of the
twenty-first century, the language is almost certainly extinct.
The Mochica language has been known by several names, most of which are ambigu-
ous or misleading. Carrera (1644) and Mart´ınez Compa˜n´on (1985 [1782–90]) called it
Yunga,which is a Quechua word for low altitude areas with a temperate climate, for the
populations living there and for their languages. Middendorf (1892) opted for the names
Muchik (mentioned by the Augustinian chronicler Antonio de la Calancha in 1638) and
Chimu with a reference to the kingdom of Chim´uorChimor, which had its capital at
Chanch´an, just north of Trujillo, and which was subjugated by the Incas in about 1470
(Rowe 1948). The name Mochica or Muchic is reminiscent of the name of the indige-
nous community of Moche south of Trujillo. However, Cerr´on-Palomino (1995b: 41)
argues against a relationship between the language name Mochica and the town name
Moche. The surroundings of Trujillo, including Chanch´an and Moche, belonged to the
linguistic area of Quingnam, another language mentioned by Calancha. Quingnam was
also referred to as the ‘Fisherman’s language’ (la lengua pescadora or la lengua yunga
pescadora).
108
Very little is known about this language of which neither a grammar, nor
a dictionary has been preserved.
109
Calancha suggests that it was in use all along the
central Peruvian coast as far south as Carabayllo (near Lima), which was as far as the
former kingdom of Chimor had extended.
In a detailed analysis based on Carrera’s and Calancha’s affirmations and on a docu-
ment of 1638 published by Ramos Cabredo (1950), Torero (1986) defines the linguistic
area of the Mochica language as the coastal region extending between the R´ıo la Leche
and the town of Motupe, to the north, and the Chicama river valley with the town of Paij´an,
to the south. The southernmost part of this region, situated between the Jequetepeque (or
Pacasmayo) and Chicama rivers apparently was a contact area where both Mochica and
Quingnam competed due to a northward expansion of the Chim´u kings. On the north side,
Mochica bordered on the Sechura language and the language of the oasis of Olmos, the
latter known from specific mentions in colonial sources (Cabello Valboa 1586; Calancha
1638). The Sechura language survived until the nineteenth century (Rivet 1949).
108
It has been suggested that the ‘Fisherman’s language’ was a language distinct from Quingnam
(Rabinowitz 1983), but this view is rejected in Torero (1986) on the basis of an analysis of the
phrasing in Calancha’s text.
109
The Trujillan scholar Zevallos Qui˜nones (1989, 1992) has studied the lineage names of the
indigenous elites of Lambayeque and of the Trujillo region. His data show a marked lexical
and phonetic contrast between the two areas.
3.4 The Mochica language 321
For the time being, the Mochica language must be considered as a language isolate,
notwithstanding the fact that several authors have tried to connect it to other languages.
Jij´on y Caama˜no (1941: 6) has proposed a genetic relationship with the extinct (and
virtually undocumented) Ca˜nar and Puruh´a languages of the Ecuadorian highlands.
110
Relations of Mochica with Mayan, with Mapuche, and with Uru-Chipaya (Stark 1972a)
have been suggested as well.
The Mochica language is well known for its unique exotic sound system, which has
generated even more exotic orthographies (especially Carrera Daza’s). From a structural
point of view, it is notoriously different from both Quechua and Aymara, the influence in
either direction being limited to a few loan words from Quechua and vice-versa (Cerr´on-
Palomino 1989c). It lacks both the morphological complexity and the rigid regularity of
the major Andean languages.
3.4.1 The sounds of Mochica
The reconstruction and the recovery of the Mochica sound system are problematic.
The various scholars of the language have developed and used different notations. Both
Carrera Daza and Middendorf went a long way towards explaining the value of their
symbols, but neither of them succeeded in eliminating all doubts as to the pronuncia-
tion of these symbols, their observations often being very far apart from each other. A
very helpful comparison of the sources, enriched with personal observations obtained
in 1929, is found in Lehmann’s notes, published by Schumacher de Pe˜na (1991). The
much awaited field notes collected by Br¨uning in 1904–5 are kept in the Ethnographic
Museum of Hamburg in unpublished form (Cerr´on-Palomino 1995b: 66–7). An addi-
tional complication for the interpretation of the original Mochica sound system is the
fact that several crucial phonological developments occurred between the seventeenth
and nineteenth centuries, which makes it hazardous to use the more recent data for the
interpretation of the older materials. Furthermore, there remains the question whether
the dialect that survived in the Eten–Monsef´u area was representative of the language
as a whole.
The seventeenth century Mochica sound system has been the object of several mod-
ern reconstructions (Hovdhaugen 1992; Cerr´on-Palomino 1995b; Torero 1997), but
the results are not at all concurrent.
111
Since the exact pronunciation of the Mochica
sounds remains a matter of speculation, the examples in the following pages are repre-
sented in their original orthography. In the section on grammar, grammatical forms and,
110
The only suggestive similarity which Jij´on y Caama˜no indicates between Mochica and Ca˜nar
is a single word, Mochica nech [neˇc] ‘river’, recorded as necha in Ca˜nar.
111
Anew comparative study of the Mochica lexicon preserved in the different sources is Salas
(2002). It contains a discussion of the sound system.
322 3 The Inca Sphere
occasionally, lexical items will be given in Carrera Daza’s seventeenth century orthogra-
phy, followed by a slash ‘/’ and Middendorf ’s transcription whenever there is a difference.
The oldest known specimens of the Mochica language are found in a religious text-
book, Rituale seu Manuale Peruanum,byGer´onimo de Or´e (1607). Its spelling does
not cast much light on the complexities of the language. Carrera Daza’s Arte de la
lengua yunga, best known through Altieri’s commented re-edition of 1939, is funda-
mental for understanding the phonological and morphosyntactic characteristics of the
language. Villareal (1921) incorporates Carrera Daza’s grammar and presents the data
in a reorganised but often erratic way. Middendorf (1892) is based on contemporary
data, although Carrera Daza’s Arte provided the framework for part of the elicitation.
Carrera Daza distinguishes six vowels in Mochica: <a>, < e>, <i>, <o>, <u>
and <œ>.
112
The latter is the object of an impressionistic description to the effect that
“it begins as an e and ends as an u,insuch a way that there are two vowels in one”.
113
Furthermore, Carrera Daza uses a diacritic to indicate length. It remains restricted to a
few roots and to the ending -ˆo,which is part of the genitive construction:
(232) ˆutzh
‘big’
(233) aiu-ng-œn-ˆo
that-G-PL-AJ
‘belonging to those’
Torero (1986, 1997) points out that Carrera Daza’s symbol <i> does not always
refer to a full vowel. It may indicate a non-syllabic glide as in example (233), and it
may either indicate or emphasise the palatal nature of an adjacent consonant, as in ˜nai˜n
‘bird’ (/n
y
an
y
/) or ¸cio ‘he, it’ (/s
y
o/). According to this author, vowel sequences and true
diphthongs did not occur in seventeenth century Mochica, given the fact that non-nuclear
<i> can always be interpreted either as a consonant, or as a palatality marker, whereas
non-nuclear <u> is not found.
Middendorf (1892) distinguishes eleven plain vowels, two ‘impure’ vowels and four
diphthongs. Although one can assume that some of these distinctions are liable to be
eliminated by reanalysis, this is not possible in a number of cases where minimal pairs are
provided. Reanalysis leading to elimination may be possible in the case of the diphthongs
(ai, ei, oi, ui)ifone considers the second element to be a palatality marker, as does Torero
(see above). It should be observed, however, that in nineteenth century Mochica, the
112
Villareal (1921) uses the symbol <æ>,whereas Altieri’s 1939 edition of Carrera Daza’s gram-
mar has capital <Æ> alternating with standard <œ>.For reasons of transparency we will only
use the latter symbol in our examples.
113
‘Tiene principio de e yfindeu,demanera que son dos vocales en una’ (Altieri 1939: 11).
3.4 The Mochica language 323
palatality distinction no longer had the significance which it had two centuries earlier.
Middendorf emphasises the distinct, almost separated pronunciation of the elements that
make up a diphthong. As it appears, the fact that i functioned as a palatality marker did
not mean that its presence was merely an orthographic device.
The two ‘impure’ vowels, which Middendorf writes ¨a and ˚u respectively, are discussed
by that author in great detail, with the confession that he never managed to pronounce
them properly. Middendorf’s characterisation of the impure vowels is essentially the
same as that of Carrera Daza’s for the symbol <œ> with the difference that the ‘u’
element is said to be more strongly represented in the vowel ˚u, and more weakly so, if
at all, in the vowel ¨a. The vowel ˚u is relatively infrequent and seems to be restricted to
roots with a preference for initial and pre-labial positions, whereas no such restrictions
apply to ¨a.However, there is at least one minimal pair: ¨ap ‘hot pepper’ versus ˚up ‘salt’
(cf. Torero 1997: 125). These two words are clearly kept apart in Lehmann’s phonetic
notation (Schumacher de Pe˜na 1991): ‘hot pepper’ œˆp, ‘salt’ ¯u˘up. Carrera Daza only
recorded the word for ‘salt’ as œp.
The high frequency of one of the so-called impure vowels (viz. Carrera Daza’s <œ>
corresponding to Middendorf’s ¨a), which occurred in endings presumably unstressed as
well as in roots, leads to the question whether indeed this vowel represented a single
phoneme. Mochica morphophonemics includes a rule of vowel loss in unstressed open
syllables which always affects precisely that vowel. It would not be far-fetched to assume
that the Mochica vowel inventory included a schwa-type vowel which may have been
the product of a neutralisation of several full vowels, rather than an allophone of œ/¨a
alone.
114
This mechanism of vowel loss is illustrated in (234):
(234) ˜nofœn ˜nofn-œr-ˆo
man man-G-AJ
‘man’ ‘belonging to the man’
(235) mit-apœc mit-apc-o-i˜n
bring-AG bring-AG-AR-1S.SG
‘one who brings’ ‘I am in the habit of bringing.’ (lit. ‘I am one who
brings.’)
Middendorf’s inventory of plain vowels (Middendorf 1892: 48–51) includes four
long vowels ¯a, ¯ı, ¯o and ¯u,which are matched by four ‘normal’ vowels a, ˘ı, ˘o,
115
and u.
Additionally, there are two more short vowels ˘a and ˘u. The pronunciation of ˘a is said to
be like a in German before doubled consonants, whereas ¯u is said to be similar to u in
114
Torero (1997: 125) speaks of a vocal de apoyo ‘support vowel’.
115
The symbols ˘ı and ˘o are not consistently employed by Middendorf; in most of his grammar he
replaces them with i and o.