Glottalized and Aspirated Stops in Cuzco Quechua

by Steve Parker, David Weber
Glottalized and Aspirated Stops in Cuzco Quechua
Steve Parker, David Weber
International Journal of American Linguistics
Start Page: 
End Page: 
Select license: 
Select License


    Introduction. Classical analyses of Cuzco Quechua phonology, such as CusihuamAn (1976b), ascribe phonemic status to both aspirated and glot- taliaed stops, contrasting them at the segmental level with each other and with voiceless unaspirated plosives. However, the aspirated and glottalized stops are severely restricted in their distribution compared to the other con- sonant phonemes. The purpose of this paper is to argue that these "laryn- gealized" stops2 are best analyzed not as independent, contrastive segments in their own right, but rather as the manifestation of a floating root-level autosegment. In 4 we propose a formal analysis expressed in terms of com- monly accepted theoretical devices. As we shall demonstrate, a great number of facts relating to the restricted distribution of glottalized and aspirated stops can be accounted for by a small set of formal phonological mechanisms.

    Overview of Quechua phonology. According to traditional descrip- tions (Rowe 1950 and CusihuamAn 1976b:291), the inventory of Cuzco Quechua consonant phonemes consists of twenty-six individual segments:

(I)     p?         t?         k?     q7
    ph        th     Clh     kh     qh
    p    t     c     '     k     9
            s    S             h

mnii I X



IWe are indebted to Pete Landerman (personal communication) for first pointing out to us the restricted distribution of glottalized and aspirated stops in this language. We also thank Andy and Cheryl Black, Andy Eatough, Barbara Hollenbach, Peter Ladefoged, Steve Marlett, David Rood, and an anonymous referee for significant input on previous versions of this paper. We alone are responsible for any remaining deficiencies. The authors' names are listed alphabetically.

Throughout this paper we employ the term STOP in a generic sense to include the affricates [El, [Eh], and [E7]. However, we exclude nasals from this category, so STOP in this context should be understood as implying oral, nonfricative obstruents. Furthermore, for want of a bet- ter label, we use the cover term LARYNGEALIZED to refer to both ejectives and aspirated stops as a group, and not in the traditional sense of creaky voice quality. Two precedents for using

[IJAL, vol. 62, no. I. January 1996. pp 70-851 01996 by The University of Chicago. All rlghts reserved 0020-707119616201-0004$01 00

Of these, the top row consists of glottalized (ejective) noncontinuant ob- struents. Contrasting with these at the segmental level is a series of corre- sponding aspirated stops, as well as their plain, unmodified counterpart^.^

Cuzco Quechua also has five vowel phonemes: /iea~u/.~

Primary stress normally falls on the penultimate syllable of each word. However, there are some words in which stress falls on the ultima, con- trasting with analogous forms bearing penultimate stress (Cusihuamin 1976b:57).

With respect to syllable structure, the great majority of Quechua words can be exhaustively parsed with the relatively uncomplicated maximal tem- plate [CVC]. This gives rise to the following syllable types (and no others): CV V CVC, and VC.5 Consequently, word-medial consonant clusters are heterosyllabic, e.g., /pisqa/ 'five':

Vowel-initial syllables occur only word-initially (within a phrase), whereas all word-internal syllables contain an onset, in keeping with the Obligatory Onset Condition noted for other languages (It6 1986 and Mohanan 1989).6 Therefore, a possible word may have the structure CVCVCV or CVC.CV but not *CVVC.CV

LARYNCEALIZED in this sense for Quechua are Orr and Longacre (1968) and Carenko (1975). For a more general reference, see Maddieson (1984).

'E', Eh, and E are alveopalatal affricates. q', qh,and q are voiceless postvelar or uvular stops. iC is an alveopalatal lateral. The voiced stops b, d, and g also occur in Cuzco Quechua, but only in Spanish loanwords. They will be largely ignored in this paper.

"rior to its contact with Spanish, Quechua had only three vowel phonemes: Iiaul. There now exist minimal pairs demonstrating a contrast between high and mid vowels, e.g., [waEu-] 'to plow, furrow' vs. [wai-0-1 'to fornicate' (Segundo Villasante O., personal communication). Cusihuaman (1976b:291) also analyzes e and o as phonemic in Cuzco Quechua.

Cusihuaman (1976b:55-56) notes the following exceptions: Under highly restricted cir- cumstances, a heavy coda results in the syllable type CVCC, as in certain onomatopoeic words. Additionally, large-scale borrowing from Spanish has introduced complex onsets in loanwords such as Itriyul, from trigo 'wheat'. Since these facts are not germane to our central arguments, we ignore them in the remaining discussion.

60f course, whenever a vowel-initial word is pronounced in isolation, it is preceded by a phonetic [?I. Due to this nearly universal process, virtually all surface syllables contain on- sets. The only cases of syllables without phonetic onsets arise when a vowel-final word is im- mediately followed by a vowel-initial word in the same phrase. See 4 for more discussion.

3. Distribution of glottalized and aspirated stops. In Cuzco Quechua, the plain (nonlaryngealized) consonants listed in (1) are quite unrestricted in distribution, as would be expected given their phonemic status as contras- tive segments. Thus, any of the sounds IptEkqs Shmnii 1Lrw y/ may occur in the onset position of any syllable of the word, as well as in codas (with perhaps a few exception^).^ Furthermore, these segments may occur more than once in the same root, and in suffixes as well as in roots (again, with perhaps a few ex~e~tions).~

(Quechua does not have any prefixes.) In contrast to this normal pattern, however, glottalized and aspirated stops are subject to extreme phonotactic restrictions:

    They occur only in roots, never in suffixes.
    They occur only in the onset position of the syllable, never in codas.
    When laryngealized consonants do occur, they are always the first syllable-initial stop of the root. Thus, forms such as the hypothetical *[poq?a], in which the second stop is glottalized (rather than the first one), are systematically prohibited.
    They may occur only once per root. No Quechua root contains two as- pirated stops or two glottalized stops. Furthermore, glottalization and aspiration are mutually exclusive in the sense that they never cooccur in the same root. Thus, a word may have an aspirated stop or a glot- talized one, but not both. The generalization is that laryngealization is restricted to one occurrence per root. Therefore, we might say that laryngealization in Cuzco Quechua is culminative in nature and can thus help to demarcate word boundaries (cf. Carenko 1975:14).
    Aspirated stops never occur in roots that begin with an /h/. Perhaps this constraint could be expressed more strongly as follows: aspirated con- sonants and /h/ are mutually exclusive with respect to roots. That is, if a root contains an aspirated stop, it may not contain an /h/, and vice versa.9

Cf) Words containing glottalized stops always begin with a consonant. Whenever an ejective occurs in a reflex of a Proto-Quechua root that began with a vowel, in Cuzco Quechua the word begins instead with [h]. For example, the Proto-Quechua form for 'how many?' is *uyku, while in Cuzco it is hayk7a. Likewise, the Proto-Quechua root for 'uri- nate' (*iSpa-) has evolved in Cuzco as hisp7u-.

For example, a number of historical changes have affected syllable codas: *p > q, *S > s, etc. See Mannheim (1991) for more discussion.

For example, no Quechua suffix contains an /h/.

9This restriction, confirmed by a careful inspection of all h-initial words in Cusihuamln (1976a). is at odds with Mannheim's (1991:207) claim that "It is perfectly acceptable to have aspiration in a stem that begins with h. . . . " Unfortunately, he does not give any unambigu- ous examples to illustrate this alleged phenomenon, although he does list hathun as an alter- nate form of hatun 'large' (1991: 196).

Concerning these last two points (e and f ), we wish to clarify that /h/ is indeed a contrastive phoneme in Cuzco Quechua, despite its epenthetic na- ture in forms such as [hayk7a]. Thus, not all words that begin with an [h] contain a glottalized stop later in the root; for example, /hatun/ 'big; tall', lhuEal 'guilt; sin'. Furthermore, /h/ also occurs intervocalically within roots in a few scattered forms, such as /sehe/ 'barn; grain loft', /muhu/ 'seed', etc. (CusihuamBn 1976~).

With respect to point (f) above, Parker (1969:175) posits a diachronic process to account for the interdependency between word-initial [h] and glottalized stops in Cuzco Quechua: "an epenthetic *h emerges before a word-initial vowel if any medial syllable of the stem begins with a glottal- ized stop." He formalizes this rule as follow^:'^

(3) 0 + h 1 # -V (C1 V) C' (where C' stands for an ejective)

This rule implies that the glottalization of word-internal stops in Cuzco Quechua developed prior to and independently from the insertion of word- initial h's, raising an interesting question: Why was this epenthetic [h] in- troduced in vowel-initial words containing a glottalized stop? Although both Carenko (1975) and Mannheim (1991) make note of this insertion pro- cess, neither offers a formal analysis which accounts for its phonological motivation. This is what we propose to do in 4 below.

The following data illustrate the distributional patterns of laryngealized stops described in this section:
(4)     p7atay     'to bite (viciously)'
    p ha tay    'to explode, blow up'
    t7anta    'bread'
    thanta    'old; used up; worn out'
    E7aEu    'treacherous, tricky'
    EhaEu    'ragged, tattered'
    k7anka    'rooster'
    khanka    'slimy, clammy'
    q'a ta    'turbid, muddy'
    qhata    'mountainside'
    hap7iy    'to light (a fire); grab, hold'
    wirp7a    'lip'
    suphu    'body hair'

"There is ample evidence to refute an analysis whereby word-initial h's are posited for Proto-Quechua and then deleted for those dialects in which they do not occur synchronically. It would take us beyond the scope of this paper to summarize such evidence here. One fact which is consistent with Parker's (1969) analysis is the existence of assimilated Spanish loan- words such as [hasut7i], from azote 'whip', in which the initial [h] in the Cuzco form is clearly epenthetic.

musp hapakuy 'to be delirious'
,(a t7a y 'to unplug; uncork'
hunt7a 'full; complete'
rakt7a 'thick' (referring to flat objects)
it ha 'species of small insect'


mlc u 'mixed, jumbled'
moqE?ikuy 'to wash, rinse out the mouth'
sunE?u 'yellow aster'
aEhuy 'to postpone, delay'
anEhayna 'thus, so'
hak 'u 'floury'
rank7ukuy 'to get twisted up'
akhakaw 'how hot it is!'
muskhiy 'to smell'
naq70 'dented, bruised'
hanq7as 'species of lupine plant'
hawqq7ey1' 'to dig up, unearth'
leg he 'rotten, putrid'
rang ha 'forgetful'
warak7a 'sling made of wool'
hamawt7a 'learned, wise'

Naturally, words formed by compounding two roots might initially appear to violate some of the restrictions illustrated above. Some examples (taken from Cusihuaman 1976a) are as follows: [qhete-E7uspi] 'mosquito', [phaka- qhata] 'groin', [maki-t7anta] 'palm (of hand)',12 [hisp7ay-p7uru] 'bladder', etc.13

Further examples of words illustrating these distributional limitations are listed in Appendix A. Several previous works have noted some or all of these phonotactic restrictions on laryngealized segments (Rowe 1950, Orr

"The coda sequence wq in words such as these is in free variation with p in some dialects, for historical reasons (Cusihuaman 1976b:55-56). Since the sequence wq in this position is normally pronounced as [xW], we do not consider such forms to be exceptions to the general- ization that the maximal coda in Quechua is limited to one timing slot.

l2 Forms such as [maki-t7anta] are unusual in that the glottalization does not occur on the first syllable-initial stop of the entire compound word.

l3 Other apparent exceptions to these patterns of distribution have appeared from time to time in the literature. For example, Cusihuaman (1976~) gives /Eaqhay/ as an alternate form of /haqay/ 'that one'. Similarly, Condori Mamani (1977) contains forms such as /t70kha-1 'to burst, explode' and /phoqEhiEeq/ 'to dam up; to overflow'. However, some native speakers reject these as incorrect. Cusihuaman's (1976~) entries for these two verbs are /t70ha-/ and /phoqEi-1, re- spectively. Finally, we note that orthographically transcribed forms such as qhaphra 'fragile' do not contain an aspirated /ph/. Rather, in this position the graphemeph represents a syllable-final If/. Cusihuamin (1976~) gives the variants qhawqra and qhapra for this word.

and Longacre 1968, Carenko 1975, and Mannheim 1991). An adequate pho- nological description of Cuzco Quechua should explain these patterns. The traditional analysis, in which glottalized and aspirated stops are ascribed phonemic status as contrastive units at a segmental level, fails to do so.

4. Phonological analysis. Given that a root in Cuzco Quechua may con- tain at most one laryngeal feature, and given that the "docking site" of these features is completely predictable (the first syllable-initial stop of the word), we should not speak of glottalized and aspirated PHONEMES but rather glot- talized and aspirated ROOTS, as Carenko (1975) and Mannheim (1991) point out. While this move is a step in the right direction, it does not directly account for the curious distributional characteristics of [h] vis-8-vis laryn- gealized roots. In this section, we posit a formal analysis which explains these facts in a simple and straightforward way.

We propose that the inventory of Cuzco Quechua consonant phonemes does not include aspirated or glottalized stops. Rather, in the lexicon a given root may (optionally) be assigned, in addition to its linear segments, one of two Laryngeal Node features: for glottalization the feature constricted glottis ([cg]), and for aspiration and /N the feature spread glottis ([sg]).14 These autosegments are required to dock to the first [-son, -cont] segment in an onset position (scanning the word from left to right). Obvi- ously, the motivation for invoking these steps is distributional, based on the defective patterns outlined in 3. To illustrate this analysis, the words [tanta], [t7anta], and [thanta] have the following lexical entries:15
(5)     [tanta]:     Itantal     'collection; combination'
    [t7anta]:    Itanta, [cgll     'bread'
    [thanta]:    Itanta, [sgll     'old; used up; worn out'

In order that these root-level features surface phonetically on a particular segment, we need to posit an autosegmental association rule. In formalizing

I4For the sake of restricting our analysis as much as possible, we posit that the features [cg] and [sg] are privative in this language. As far as we can tell, no details of our presenta- tion hinge crucially upon this assumption.

Is As pointed out by an anonymous reviewer, the fact that these laryngeal autosegments are limited to one per root must be stipulated as a language-specific constraint; it does not fall out directly from any universal principle. That is, while the two features [cg] and [sg] are mu- tually exclusive if associated to the same segment, there is no a priori reason they should not both be available in the same root. A parallel with tone would be helpful here. Autosegmental melodies like HL or LHL do occur at the morpheme level, even though these pitch specifica- tions would be contradictory if realized simultaneously. Thus, hypothetical root "melodies" such as [cg] [sg] [cg] cannot be automatically discounted. Nevertheless, we are not aware of any autosegmentalized features other than tone which have been claimed to exhibit contour melodies of this sort.

this process we use the parametric model recently introduced by Archangeli and Pulleyblank (1994). In their format, phonological rules are expressed as a convergence of four binary parameter settings: function, type, direction, and iteration. The default or unmarked type element is path, which corre- sponds to traditional association lines. The nondefault setting for type (what gets inserted or deleted) is F-element, or feature. Figure 1 shows a complete rule chart with all of the potential settings filled in to illustrate the model. In figure 1, A and T stand for Argument and Target, respectively. We can now formalize the rule of Laryngeal Docking, as shown in figure 2.

This rule specifies that a floating [sg] feature is associated with the left- most stop of the word, producing aspiration. For glottalization, a very simi- lar chart would be required, but the Argument would be CG rather than SG.

In more traditional geometric terms, the rule of Laryngeal Docking can be formalized as follows:

(6) SG

In order to ensure that the two laryngeal autosegments associate only with onset consonants, and not codas, we posit a filter which prohibits con- sonants in syllable-final position from licensing a Laryngeal Node:

(7) Coda Filter

d Root Node

I! Laryngeal Node

This device, which correctly prevents the features [cg] and [sg] from docking to coda consonants, is made possible because, with the exception of a few borrowed words, voiced stops are systematically absent from syllable codas. Since all the remaining obstruents in Cuzco Quechua are predictably voiceless, and since all sonorants except /h/ are redundantly voiced, the only underlying segments which would need to be lexically specified for any Laryngeal Node features are /h/ ([sg]) and the class /bdg/ ([+ voice]). The filter (7) correctly excludes them from occurring syllable-finally. Assuming that this filter is turned off postlexically, the default rule [a son] +[a voice] would then supply all other Laryngeal Node specifications necessary for implementing the speech string phonetically. The Coda Filter posited above is in no sense unusual since languages naturally tend to maximize consonant contrasts in onset position while neutralizing many of them syllable-finally.16 This filter constrains the operation of the Laryngeal Docking rules by causing them to skip over the first stop in a root if it happens to occur in coda position.

l6 In fact, this tendency is more extreme in Cuzco than in any other dialect of Quechua, as Mannheim (1991) notes.

We now address the formal issues raised by the distribution of [hl in la- ryngealized roots. The prohibition against aspirated stops and /h/ occurring in the same root can be accounted for by the filtering action of the Obliga- tory Contour Principle (OCP). Given lexical representations such as those posited above, two hypothetical /h/-initial forms would have these underly- ing configurations:

(8) [hanta]: / h ant a/


When the floating [sg] in (9) docks, an OCP violation results because at that stage in the derivation the representation contains adjacent occurrences of the same feature on the same tier:

If this hypothetical form were now to surface phonetically, it would be ungrammatical. One way in which this anomaly might be resolved is by delinking one of the two occurrences of [sg]. If we assume that an /h/ which has been disassociated from its underlying [sg] feature will be de- leted (perhaps by Stray Erasure), one of the following two results would be produced:

(1 1) hanta or

Either of the options in (11) would be an acceptable phonetic form in Cuzco Quechua. As a matter of fact, the prediction that fluctuation of this nature might arise is confirmed by cognates such as huqari-'to lift; rise' in Cuzco vs. uqhari-in Bolivian Quechua and hurqu-'to remove, take out' (Cuzco) vs. urqhu-(Bolivia) (Carenko 1975: 1 I)."

"Indeterminancy of this nature has also resulted in fluctuation between synchronically attested variants such as [haqay] vs. [taqhay] for the word meaning 'that one' (Cusihuaman 1976~).In this case, if we assume that the initial [h] of [haqay] is somehow more basic than the

What is more, by appealing again to the OCP, we can develop a formal explanation for the appearance of word-initial epenthetic h's in glottalized roots. Recall from n. 6 in 2 above that in many languages, words (or phrases) which are underlyingly vowel-initial actually begin with a phonetic glottal stop; this is a low-level, automatic reflex to fulfill the requirement for an onset (cf. the Obligatory Onset Condition). For example, a word such as Iasikuyl 'to laugh', which contains no ejective, is pronounced phrase-initially or in isolation as [?asikuy]. In words containing a noninitial glottalized stop, however, the application of this process leads to another violation of the OCP, as in the partial derivation of the word [hayk7a] 'how many? below:

(12) Underlying Representation I a y k a ,[cg] 1

Laryngeal Docking

i' Epenthesis i'ayk7a


Resulting Form *[?ayk7a]

At this stage in the derivation, the intermediate form *[?ayk7a] violates the OCP on the [cg] tier and so is ungrammatical (in Cuzco Quechua). To resolve this clash, we posit that the initial [?] is converted into an [h] by a process of dis~imilation,'~

thereby explaining the motivation for Parker's (1969) diachronic rule in (3) above. Synchronic evidence for the continuing productivity of this rule is offered by variants such as [abpa] -[habp7a] 'dirt, ground, soil' and [usuta] -[husut7a] 'sandal' (Cusihuamin 1976~). In sum- mary, the process of onset epenthesis inserts a C-slot which is then filled by a laryngeal glide, subject to the OCP. Thus, we see that by relying on a hand- ful of basic, uncontroversial mechanisms of current phonological theory, all the facts relating to the interaction of laryngeal features in Cuzco Quechua are accounted for simply and elegantly.

corresponding [El of [i-aqhay], we can account for why the laryngealization "skips over" the first syllable-initial stop in the latter form, in violation of one of the restrictions previously noted. Another possible explanation of the form c'aqhayis that it resulted from a merger of the words c'ay 'that' and haqay 'that (distal)' (Pete Landerman, personal communication). (The latter is the current form in Bolivian Quechua.)

l8 We thank Barbara Hollenbach (personal communication) for pointing this out to us. Carenko (1975) offers a very similar explanation.

One of the strengths of this formal analysis is that it directly accounts for the fact that epenthetic [h] and [?I are in complementary distribution in Cuzco Quechua: a phonetic [?I is inserted to fulfill the requirement that all syllables have onsets, except in words with ejectives, which surface with an initial [h] instead. (An important difference between these two "allophones" is that [?I insertion only applies phrase-initially, whereas [h] epenthesis ap- plies within a phrase as well.) These findings are theoretically significant in that they extend the domain in which the OCP is operative. That is, it is typically claimed that the OCP is active in root constraints but is turned off in the postlexical phonology (McCarthy 1986). Word-initial epenthesis in this language, however, has postlexical properties: it inserts [?I, which is not a phoneme, and thus violates Structure Preservation. Nevertheless, this phonetic [?I is blocked from surfacing if another [cg] specification occurs later in the word. Thus, here is an apparent postlexical process which is sur- prisingly subject to the 0CP.19

5. Conclusion. Previous treatments of glottalized and aspirated stops in Cuzco Quechua, such as Carenko (1975) and Mannheim (1991), have de- scribed the tight restrictions on laryngeal segments in this language, but without directly accounting for their distribution in formal terms. In his concluding section, Carenko (1975:13) postulates that Quechua dislikes sequences of two identical elements, be they vowels, consonants, or laryn- gealized articulations, while Mannheim and Newfield (1982) and Mannheim (1991) concern themselves primarily with the spread of iconic laryngealiza- tion in this variety of Quechua via processes of onomatopoeia, semantic ex- tension, and lexical association.

The contribution which our analysis adds to these works is that we pro- vide a phonological explanation of the data utilizing a minimum of formal machinery. By making use of well-established phonological principles and constraints, we show how the distribution of aspirated and glottalized stops can be accounted for in a compact way. The result is that a large number of seemingly unrelated facts fall out directly from the joint operation of a few simple devices: a root-level autosegmental prosody, a coda filter, the Oblig- atory Onset Condition, and the Obligatory Contour Principle.


The forms listed in this appendix, when combined with the examples given throughout the text, constitute a nearly exhaustive list of all the morphemes in

l9 Thanks to an anonymous referee for pointing this out

CusihuamBn (1976a) which contain either a glottalized or an aspirated stop in a noninitial syllable. For some representative (but nonexhaustive) examples of la- ryngealized stops in word-initial position, see list (4) in the body of the paper. This appendix is organized as follows: all words containing Ip7/ as the onset of the second syllable of the word are listed first, followed by words exhibiting /ph/ in that position, then /t7/, /th/, /E7/, etc. A further distinction is also made as to whether the initial syllable is open or closed. Within each of these subsections the words are alphabetized according to the first letter in the Quechua form.

lap70tiyay kip7iqey rap 'a rap7i siP7u hakp$ hakp7iy hasp7iy hisp7ay hump7i kamp7u h'usP7iy mikp7uy mukp7uy samp7a sikp'uy simp7a kut'ay mat7ay mat7i mot7e fiit7iy fiut7u rit7i sat7iy sit'i sut7i sut7u wit7uY hawkt7ay lapt7ay h'ant7a kipt7a kiwkt7a maqt7a mastLy mayt70

'to walk in the mud'

'to blaze, flame'


variant of rap7a 'leaf'

'folded, bent, wrinkled up'

variant of akpa 'dirt, ground, soil'

'to dig up' 'to scratch' 'to urinate' 'perspiration, sweat' 'soft' 'to slip out, escape' variant of mukp7uy 'to swallow' 'tame; peaceful' 'to knot up' 'plait, braid; tress'

'to swallow'

'to cover up (a hole)'

'to slice, carve (meat)'

'hard; forehead'

'boiled corn'

'to squeeze, crush'

'crumbled, broken to pieces'

'snow; ice'

'to put, place inside'


'clear, sharp'

'wet; leaking'

'to cut into pieces' (e.g., a stick)


'to lick up, lap' (said of a dog)

'firewood; cow' (colloquial)

'quinoa seed (ground up to make cereal)'

variant of kipt7a 'quinoa seed'

'young man, youth'

'to stretch out, extend'

'bunch, bundle'


6ust7a walt7ay wist7ay wist7u ithi muthuy wathiyay hat7iyuy hat7uy hit7ay hut7uy

XU? 'U



mac aqway


mac ay


mrc a mut7ay muE7u sat7a sut7i suE7u


wrc zcrcry wit70n wit7utu kawt7iy ka wkt 'ay l~iwkr'iy kunt7isqa kunt7iy riaqt7a riawt7i riukE7u rikt7ay wah't7a wikt7uy athiy hik7iy huk7i huk'uta h'ik7i kuk7iy mak7ah'ikuy mak7as mik7i

'country woman in native dress' 'to wrap up; bandage' 'to throw or cast a sharp stick' 'lame, crooked; bowlegged' 'small, diminutive' 'a type of common shrub with yellow flowers' 'to prepare a barbecue' 'to throw dust at; cover with dust' 'to chew, gnaw' 'to throw, cast; to empty' 'small; short' 'impudent, brazen' variant of mat7aqway 'snake' 'snake' 'cave; shelter, refuge' 'wretched, unfortunate' 'to kiss' 'nape of a four-legged animal' 'bush, shrub; vegetation' 'species of fish found at high altitudes' 'lame; paralyzed from the waist down' 'onomatopoeic sound made by pigs' 'lower part of the arm or thigh' 'elbow' 'to clear, cut down weeds' 'to take something out of the water' 'to tear, rip, break apart' 'painted' (past participle) 'to paint' 'comb' 'cutting edge' (of a knife) 'a type of jungle plant with small red flowers' 'physiognomy; to wake up' 'drivel; slime' 'to throw or cast out; abandon' 'to sneeze' 'to cut up a peeled potato' 'corner, interior angle' 'mouse; small child' (colloquial)

'to hang on to someone' (said of a child)
'a type of pitcher carried on the shoulder'
'humid, damp, moist'

kunk7uy kusk7a mawk7a mink7a misk7akuy misk7i mukk7a rakk7a rawk7ana sank7ayu say k7uc'iy sayk7uy sirk7a sukk$ wank7a wank7iy wark7a wask7a wayk7u wisk7ac'a ankhayna fiuskhu sankha sankhu sunkha waskha


'to grind or crush food with the teeth'

'to behead, decapitate; hardly, difficultly'

'withered' (said of the hand)

'to extract, pull out'

'fallow field'

'orifice, opening'

'lame; unequal in size'

'to limp'

'tendon; thick muscles'

'to intercept, halt, interrupt, cut off'

'how much?; how many?'


'to stand on tiptoe'

'work, labor'

'clay, argil'

variant of kank7i 'clay, argil'

'a thick white liquid which flows from the vagina of certain animals'

'to lick one's plate'

'smooth, slick, slippery'

'old, used, worn out'

'collaboration; participation'

'to stub (a toe)'

'sweet, delicious'

'a small, round stone used to crush condiments'

variant of rakt7a 'thick' (referring to flat objects)

variant of k7akka 'species of cactus'

'to tire, wear out' (transitive)
'to tire, wear out' (intransitive)
'hot pepper seed'
'steep, abrupt'
'to roll up' (e.g., a rope)
variant of warak7a 'sling made of wool'

'vizcacha rodent'
'thus, so, in this way'
'useless' (said of a woman)
'ditch; trench; gully'
'toasted flour mixed with hot water'
'beard; mustache'

wayq'o welq7ay wisq7ay moqho naqha mosqhoy fiusqhon rayqha yarqha yarqhay hasut7i

'to suffocate; drown'

'to wet, drench, douse, soak'

'wet, humid, moist'

'thief' (derogatory)

'crushed, mashed'

variant of leqhe 'rotten, putrid'

'worn-out hat'


'spoiled, rotten' (referring to food)

'withered. wilted: dried out'

variant of naq70 'dented, bruised'

'down, pubescence formed in the water of lakes'

variant of laq70 'crushed, mashed'

'species of small cactus'

'species of small bush'

'hard, tough' (leather, meat, etc.)

'to mark, sign, engrave'

'to put to death by hanging'

'a tumor believed to be caused by supernatural beings'

'to drink, sip; absorb'

'stupid, slow, crazy'

'species of white heron'

'to squeeze, press, wring out'

variant of hawqq7ey 'to dig up, unearth'

variant of erqe 'infant, child'

'cross-eyed, squint-eyed'

'fat, chubby, robust'

'species of worm'

'to carry in one's arms'


'dull, worn-out, used up'

'knot, protuberance'

'to smell, breathe through the nose'



'ravine; canyon'

'to clean out a hole with a stick'

'to close up, cover over'

'promontory, high point of ground'

variant of naha 'previous, anterior'

'sleep; dreaming'
variant of yarqha 'canal; irrigation ditch'

'canal; irrigation ditch'

'to make a canal or irrigation ditch'


humint7a 'corn tamale'
ku~uc'~a 'edible gelatinlike substance found in lakes'
hamak7u     'tick, chigger'
hanuk7a     'weaned'
hanuk7ay     'to wean'
kawkak7u     variant of wakwak7u     'armpit'
wakwak7u     'armpit'     



DIANA, 1994. Grounded Phonology. Cambridge, Mass.: The M.I.T. Press. CARENKO,

E. I. 1975. On laryngealization in Quechua. Linguistics, An International Review


MAMANI, 1977. Autobiografia. ed. and tr. Ricardo Valderrama Fernandez and Carmen Escalante GutiCrrez. Cuzco, Peru: Centro de Estudios Rurales Andinos "Bar- tolomC de las Casas."

CUSIHUAMAN 1976~.Diccionario Quechua: Cuzco-Collao.

G., ANTONIO. Lima: Ministerio de Educaci6n and Instituto de Estudios Peruanos. . 1976b. Gramatica Quechua: Cuzco-Collao. Lima: Ministerio de Educaci6n and In- stituto de Estudios Peruanos.

DENT, LAUREL JANE. 1981. Laryngeal control in the production of three classes of voiceless stops, with occasional reference to Bolivian Quechua. Ph.D. dissertation, University of Pennsylvania.

HOLLENBACH, E. 1984. The phonology and morphology of tone and laryngeals in

BARBARA Copala Trique. Ph.D. dissertation, University of Arizona. HORNBERGER,ESTEBAN,AND NANCYHORNBERGER.

1983. Diccionario Tri-lingiie Quechua de Cusco: Quechua, English, Castellano. La Paz: Qoya Raymi. ITB, JUNKO. 1986. Syllable theory in prosodic phonology. Ph.D. dissertation, University of Massachusetts. [New York: Garland Press, 1988.1 MADDIESON,

IAN. 1984. Patterns of Sound. Cambridge: Cambridge University Press. MANNHEIM, 1991. The Language of the Inka Since the European Invasion. Austin: Uni-

BRUCE. versity of Texas Press. MANNHEIM,BRUCE,AND MADELEINE 1982. Iconicity in phonological change. Pa-

NEWFIELD.pers from the Fifth International Conference of Historical Linguistics, ed. Anders Ahlqvist, pp. 21 1-22. Amsterdam: John Benjamins.


JOHN J. 1986. OCP effects: gemination and antigemination. Linguistic Inquiry 17:207-63. MOHANAN,

TARA. 1989. Syllable structure in Malayalam. Linguistic Inquiry 20:589-625. ORR, CAROLYN, AND ROBERTE. LONGACRE.

1968. Proto-Quechumaran. Language 44:528-55. PARKER,

GARY J. 1969. Comparative Quechua phonology and grammar IV: the evolution of Quechua A. University of Hawaii Working Papers in Linguistics 1:149-204. ROWE,JOHNHOWLAND.

1950. Sound patterns in three Inca dialects. IJAL 16:137-48.

  • Recommend Us