Quechua: Main Points of Interest For Linguists

There are fascinating aspects to Quechua for linguists with interests in any of the fields covered here.

This page is just intended as a quick overview and reference guide;
fuller details on some of these issues are on my main page on Quechua linguistics.

Contents

Phonetics and Phonology

Phonological System (for Cuzco-Bolivian dialect)

More on Phonology

is available on another page of mine: click here to go there.

Lack of Voiced Obstruents

Quechua has no voiced obstruents (except through influence from Spanish, or the substrate in Ecuador). However, they have a series of five voiceless stop phonemes by location of stricture (or four, with the palatal one being affricated). Moreover each location of stricture has three contrasting phonemes at each: unaspirated, aspirated and ejective. This gives a total of fifteen voiceless stops. These generally exist only in the southern varieties, i.e. those in contact with Aymara, from which it is assumed they acquired these distinctions.

Ejective and Aspirated Stops

In compensation, you might say, certain varieties of Quechua, particularly the Cuzco and Bolivian variety, still ensures it has a sizeable inventory of obstruents by making not a two-way voiced vs. voiceless distinction, but a three-way distinction between normal (unaspirated) stops vs. aspirated stops vs. ejective stops. All of these are voiceless, and there is the full series of five locations of stricture for all three types (though the ‘palatal’ is usually pronounced as a [tS] affricate, not a true palatal stop).

Note that most other dialects do not have either aspirated or ejective stops, though some, such as Ecuador Quechua, have occasional aspirates (the distinction is, I understand, of dubious phonemic status), or by now a voiced stop series, probably largely through the influence of Spanish. Given their distribution among the dialects of Quechua, the origin of the three-way distinction is hotly disputed: original to Proto-Quechua or influence from contact with Aymara on the part of, precisely, only the southern dialects of Quechua which do show the distinction – see below for more on this.

The basic variant of these remains the unaspirated one, however. It is much more common than the other two, not least because there are clear restrictions on the occurrence of the aspirated and ejective stops: one per root, only in root lexical morphemes, not in grammatical ones, and so on. There are also very curious relationships between both these types of sound, and with the segment [h]. (There are some parallels here with the system of ‘laryngeals’ proposed for Proto-Indo-European.)

In being subject to these restrictions, and in other ways, these two ‘non-standard’ forms often seem to have much in common, so much so that many Quechua linguists use the term ‘glottalised’ as a cover term to refer to both types together (in my view, this is a bad term, I’ll explain why later).

For instance, there are several pairs of alternate forms for some words, such as allpa / hallp’a [aλpa / haλp’a] (earth) or irqi / h’irqi [erqe / herq’e] (young child). Here, where a variant has a normal unaspirated stop later in the word, the whole word begins with a vowel; but in a variant where that stop is either an aspirated or an ejective, no such word can begin with a vowel, but must start with a preposed [h] instead. (This is all a bit reminiscent of the standard modern German pronunciation of having no words that phonetically start with a vowel, since all are actually pronounced with a preposed glottal stop – this is a large part of what gives German its ‘punchy’ pronunciation, generally perceived as not the sweetest-sounding language…).

Vowels

Quechua essentially has the ‘classic’ three‑phoneme vowel system, using only the ‘most articulatorily distinct’ vowels [i], [u] and [a]. However, some dialects, including the main Cuzco‑Bolivian one, have allophonic lowering of [i] and [u] towards [e] and [o] (or even more open variants), conditioned by the context of /q/, /q'/ and /q^h/, and [h] derived from these, where these are contiguous to the vowel, or (for some speakers) separated from it only by a single continuant.

There is variation in the actual amount of lowering, depending on the precise conditioning environments (/q/ before or after the vowel, which sounds ‘block’ the lowering when they fall between the /q/ and vowel in sequences like Vrq, Vnq, Vsq, etc.), and there is also by region, morphological rules, and indeed idiolect in which of these have how much lowering effect. Native speakers show huge variation, not least in attempting to produce Spanish vowels, including massive hypercorrection, putting in [e] everywhere for Spanish [i], and [o] for [u]. All in all, a classic case of conditioned allophonic variation, and clearly not a phonemic distinction. I have yet to come across a single valid minimal pair to show that the [i] vs. [e] and [u] vs. [o] distinctions are phonemic.

The Psychological Status of the Phoneme

All this makes for a classic case study in the psychological status of the phoneme, particularly under language interference among bilinguals. For Spanish of course does have a clear five-phoneme system: /i/ /e/ /a/ /o/ /u/. Such is the extent of bilingualism with and borrowing from Spanish, that it is arguable that Quechua has begun to take on board phonemic /e/ and /o/ – though this really only applies to bilinguals, so it’s a debatable call to say it really is now Quechua, when it still isn’t for monolinguals. Anyway, it is this that has given rise to the great and bitterly divisive issue of whether to write Quechua with three or five vowels, and all the bluster that goes with it. Click here to see my article on this.

Syllable Structure, Phonotactics and Stress

Stress falls very regularly on the last but one syllable. The only exceptions are a number of ‘emotive’ suffixes, particularly for surprise or insistence.

Bolivian Quechua also shows a pattern of dropping certain final single-syllable morphemes, such as the accusative/adverbial marker ‑ta, and the yes/no question marker ‑chu, but not shifting back the stress so that it thus does fall on the final syllable, such that this non-standard stress pattern effectively becomes the marker for accusative/adverbial or yes/no questions.

The Effects of Language Contact on the Phonological System

Quechua is a fascinating case study for this, in various areas, most particularly:

• There is a great debate in the historical and comparative linguistics of Quechua as to whether language contact alone has been responsible for the wholesale ‘importation’ of the full ejective and aspirate series of five stops/affricates from Aymara into only those Quechua dialects which are in direct contact with it. (The alternative, ‘first sight’ explanation is that the three-way opposition was there in Proto-Quechua, and has been lost in all dialects but those in contact with Aymara). Some fascinating linguistic methodology has been thought up to try to elucidate the problem, though it still remains an open question. For more details, see my section on this on my What is Quechua Like page.

• The impact of Spanish on Quechua, through centuries of dominance and massive bilingualism, has meant that there are now varieties of Quechua that have acquired a voiced-voiceless distinction in stops that Proto-Quechua did not have, and there is an ongoing argument as to how far the Spanish system of five vowel phonemes has turned the originally allophonic distinctions of [i] vs. [e] and [u] vs. [o] into phonemic ones in modern Quechua too. See the above section on Vowels.

Back to Contents - Skip to Next Section: Semantics Lexis and Derivational Morphology

Inflectional Morphology

Quechua is a classic agglutinating language, showing many of the typical characteristics of such languages. (For a very good introduction to morphological typology, and the terms used below such as segmentability, fusion, invariance, incorporating and so on, see Comrie (1981: ch.2.1)). There is even a case for saying Quechua has certain elements of an incorporating morphological type, see the section below on derivational morphology.

Above all Quechua typically has long words built up from a basic root meaning followed by strings of suffixes, which are generally all very ‘clear-cut’, in that they show:

A very high degree of segmentability (or very low ‘fusion’): so while there is a ‘case system’, this is only in the standard agglutinating sense, i.e. where a word-form bears a combination of three meanings (or ‘functions’) – say number-plural, person-third, case-accusative – each individual meaning can be identified with a separate suffix, unlike in ‘fusional’ languages like Latin, Russian, and even German, where all three meanings can be ‘fused’ on a single ‘unsegmentable’ morpheme.

A very high degree of invariance (or ‘regularity’), because:

• Markers do not, as in ‘fusional’ languages like Latin, Russian, and even German, vary depending on different lexical classes. Quechua has no lexical classes at all, i.e. no different verb conjugations or noun/adjective declensions, no gender or classifier system (an no agreements other than number); no irregular verbs, nouns or adjectives. (For more on lexical classes, gender and classifier systems in general, see Corbett (1991))

• Nor does the marker of one meaning vary by the context of the other meanings it is found in combination with: for example a given person marker remains the same, irrespective of whether the number is singular or plural. (This happens ‘automatically’ in fusion, but also in some agglutinating or isolating languages)

• The only significant variance is by purely phonological conditioning in order to maintain the syllable structure within the limits of the language’s phonotactics. So almost always the choice is at most between two allomorphs (usually identical except for one having a vowel that the other does not: e.g. ‑s vs. ‑si; ‑m or ‑n vs. ‑mi), depending simply on whether the stem to which that morpheme is to be attached ends in a consonant or vowel.

Also typical of agglutinating languages, Quechua has very powerful derivational morphology. In fact part of the wealth of derivation is owed also simply to the fact that the distinction between parts of speech is not very strict.

• There is great flexibility in re-interpreting noun and verb stems as of the other part of speech (see the next section on these two points).

Moreover, there are very strong regularities and parallels (‘iconicities’ if you prefer) between the markers used on noun and verb stems. So for example the possessive person and number markers on nouns are generally identical or very similar to the subject person and number markers on verbs: e.g.

3^rd person possessive + plural ‘their’ = -n-ku their house wasi-n-ku
3^rd person subject + plural ‘they’ = -n-ku they eat qhawa-n-ku

A final typical feature of agglutinating languages to mention is that in Quechua, subordination is almost entirely by nominalisation. Fuller details on this later.

All this regularity, by the way, certainly makes Quechua easy to learn. Its case system, for instance, is a far cry from the complexities of such languages as Latin, Russian, Greek and German, and no harder, indeed than English prepositions.

Back to Contents - Skip to Next Section: Syntax

Semantics, Lexis and Derivational Morphology

Very Rich Derivational System

Quechua, also typically for an agglutinating language, has a hugely powerful derivational system, all by suffixation (there are no prefixes in the language).

Many of the suffixes blur the distinction between inflectional and derivational morphology (if you ever thought there was a clear-cut one), with some suffixes nicely straddling the borderline. The -chi ‘causative’ suffix, for instance, turns verbs like qhaway (to see) and wañuy (to die) into what in English are new lexemes: qhawachiy (to show, i.e. ‘to make [somebody] see’) and wañuchiy (to kill, i.e. ‘to make [somebody] die’). This should be derivational then, in theory. But -chi also corresponds to English ‘to have/get something done’: wasichay is ‘to build a house’, and wasichachiy is ‘to have a house built [by someone else, for you]’. This is more like something that is inflectional, as it certainly is in many languages.

Wasichay, by the way, itself is derived from the noun wasi, house, plus the –cha ‘factive’ or ‘making’ suffix, plus the –y infinitive ending. Unlike -chi, which derives new verbs from verb stems, this one derives verbs from noun stems.

In any case, the distinction between parts of speech is very flexible in Quechua. A great many verb roots are identical to the noun roots (as indeed in English). Likewise, almost any adjective can stand as a noun: puka is the adjective red, or the noun the red one, taking all the normal noun case suffixes. So as in English, you can get words of the same form which are at once a noun, an adjective, and a verb stem (like green in English). This flexibility is also crucial to the whole subordination system of the language (see below).

Even with the suffixes seen above, the fact that Quechua can say whole sentences like He is making me a house with just one word starts to make it look like not only an agglutinating language, but something of a real incorporating one. Another example is the ‘desiderative’ suffix -naya, which when added to a noun stem can transform it into a verb meaning ‘to want/feel like x’, where x is the noun. So for example aycha is the noun for meat, and aychanayashani is a form of the derived verb which means I feel like eating [some] meat.

Lexis

Native Quechua vocabulary is entirely alien to European languages, of course, though there has been a very heavy influx of Spanish loan-words.

There is a fascinating and very puzzling commonality of lexis between the different varieties of the Quechua and Aymara (also known as the Jaqi or Aru) language families. Whether this is this due just to the clearly intense contact between the language families over millennia, or also to an ultimate distant common origin, is a hotly debated topic, calling for some very interesting methodologies in historical and comparative linguistics.

Quechua also has very many fantastic onomatopoeic words, especially making use of its ejective and aspirated stops, particularly the uvulars. It also makes considerable use of reduplication. So you end up getting beauties like this ‘sentence’:

Q’ultin q’ultin q’ultin! = He downs it in one! (i.e. drains his glass)

Back to Contents - Skip to Next Section: Typology and Universals

Syntax

Which Categories does Quechua Grammaticalise?

Number is grammaticalised to a fairly large extent, though in many cases it is only optional to mark it, or cannot be marked at all when the context is clearly plural (e.g. after numerals and quantifiers). Also, when Quechua marks both the subject and (direct or indirect) object of a verb as suffixes on that verb, often the marker of plural is ambiguous as to whether it refers to the subject or (direct or indirect) object: so in quwanku, segmented qu-wa-n-ku, the root is qu- give, -wa is the first person indirect object, and -n is the third person subject. -ku is a plural marker, but it is not fixed whether it refers to the subject, indirect object person, or both, so quwanku can mean any of:

plural refers to subject person only: they give me [paykuna ñuqaman] quwanku
plural refers to indirect object person only: he/she gives us (exclusive) [pay ñuqaykuman] quwanku
plural refers to both subject and indirect object persons: they give us (exclusive) [paykuna ñuqaykuman] quwanku

The alternative, fuller formulations alongside show how any ambiguities can be resolved, if desired, by using the full pronouns for the persons involved, with the appropriate subject or object markings, and the plural marker only for those that are plural.

Key: pay = 3^rd person; nuqa = 1^st person; -man = indirect object; -kuna = plural; -yku plural (exclusive).

As you can tell from the above, Quechua grammaticalises an inclusive vs. exclusive distinction in the first person plural. (Aymara languages, by the way, have a fascinating person system, where 2^nd person takes precedence, and number is not grammaticalised for persons, and may be optionally on nouns).

As noted above, Quechua has no noun classes of any sort (neither a ‘grammatical gender’ nor a classifier system), nor indeed does it even mark sex (‘biological gender’), not even on pronouns – pay means any of he, she or it – so only lexical means are used to distinguish male from female). Likewise, animacy and humanness are not grammaticalised in Quechua (this is unlike the situation in the Aymara languages, where humanness is very importantly marked – see Hardman (1966), Hardman (1983) and Hardman (2000).)

Generally, Quechua has extensive and obligatory marking of both topic and focus, while it does not mark definiteness. In at least one dialect, however, (Junín-Huanca), the topic marker -qa has turned into a definiteness marker. Also, focus is being lost in some environments in some dialects. For more details on these points, see the sub-section below on Topic and Focus Systems.

Topic and Focus Systems

Quechua has extensive and obligatory topic and focus systems, though there is considerable variation in these between the various dialects. Moreover, the markers of each interact heavily with each other, and with certain other categories, in ways which I have never seen satisfactorily analysed and described in any literature on Quechua. A good Ph.D. topic for someone?

The topic markers are, simplifying things a lot, -ri in questions and -qa in statements. However, -qa is often found on several constituents at once, and at least in certain dialects, appears to work often as little more than a marker of constituent boundaries.

The focus markers are also the evidential/validational markers. In fact the contrasting phonological forms of these markers (-mi, -si, -cha, etc) is what marks the various evidential/validational values, while it is only the position of whichever of them is used in a sentence (i.e. which constituent it is suffixed to) which actually marks the focus of that sentence. Among other things, these markers are particularly important in signalling both yes/no questions and negations.

The Junín-Huanca dialect has turned the topic marker -qa into (unique for Quechua) a definiteness marker, outwardly akin to postposed, bound definiteness markers as in Romanian or the Scandinavian languages. See Cerrón‑Palomino (1976a).

Many varieties of Bolivian Quechua have all but lost the evidential/validational marker is the direct evidence/affirmative -mi, taking this evidential/validational state as default with zero explicit marking. What this means, though, is that focus marking is therefore completely lost when the evidential/validational context is direct evidence/affirmative.

Modality System

Quechua has a fascinating modality system, very different to European languages. More details on this soon, but if you’re interested in any of the following issues, Quechua could provide a very useful case study for you:

• Quechua has a crucial interaction between negation and interrogative mode: the same marker -chu is used to mark yes/no questions, and also compulsorily accompanies the negator mana.

• Quechua has a very rich evidential/validational system. There is a long-running debate on whether these should be analysed as evidential or validational markers (my own view is that a mixture of the two is needed, but for more details see the Ph.D. thesis by Martina Faller).

• The evidential/validational system also has interesting interactions with tense. Quechua has two past tense markers: the first (-rqa) is for past events directly experienced; the other (-sqa) for those not directly experienced, including before one’s lifetime, or even when one was asleep, dreaming, unconscious, drunk, or too young to realise what was going on. The latter is also used for surprise. Many of its functions have been copied over to the use of the Spanish pluperfect in the Andes, even now among Spanish monolinguals. See Lipski (1994).

Subordination System

Subordination is almost entirely by nominalisation, another typical feature of agglutinating languages. Fuller details on this later.

Back to Contents - Skip to Next Section: Historical and Comparative Linguistics

Typology and Universals

• Quechua blatantly defies the proposed universal tendencies on the ordering of morphemes, closer to or further from the root depending on which category they grammaticalise (e.g. as suggested in Matthews (1991)). Here’s a quick summary of the order of inflectional morphemes on Quechua verbs, showing how jumbled it is, the suffix morphemes being added sequentially going down this list. One line per morpheme, [+] indicates the two meanings are fused on that morpheme. The colour and font format is: [category]–Value_on_that_Category. This analysis is taken from Cusihuamán (2001a: 176) – certain of these meanings do allow of different segmentation analyses, but the jumbled order remains.

• verb STEM

• [person of verb object]–1^st

• [aspect]–Progressive

• [tense]–Past

• [tense]–Future + [person of verb subject]–1^st

• [mood]–Imperative + [person of verb subject]–3^rd

• [person of verb subject]–All others

• [person of verb object]–2^nd

• [tense]–Future + [person of verb subject]–3^rd

• [number of verb subject and/or object]–Plural

• [mood]–Conditional

I still have to write up the remaining themes I propose to cover in this section I’m afraid. For now, just an inkling of the themes I’ll talk about here:

• The interrelations and interference between evidential and validational systems.

• The interrelations and interference between topic and focus systems, and between focus and modality systems.

• How definiteness marking can develop historically out of topic marking.

• ‘Linguistic postulates’: how the categories a language grammaticalises can have a big role in shaping how speakers see the world in general. There are articles on this for the Aymara language family, by Martha Hardman, discussing the importance in the language of humanness, 2^nd person, and always stating your evidence for everything you say; the last one is also highly relevant to Quechua. She also discusses the serious sociolinguistic problems this can lead to in the Andes in acquisition of Spanish by Aymara speakers, and again the same goes for Quechua. References to the articles are: Hardman (1972) in Spanish; and Hardman (1978) and Hardman (1988) in English

Back to Contents - Skip to Next Section: Sociolinguistics (including Bilingual Education)

Historical and Comparative Linguistics

There is a great deal of research still to do here! For a brief rundown of the dialect differences within Quechua, see my other pages for a brief survey, and a map of dialects. Note that there are by now plenty of holes in the former Quechua dialect continuum, where the language has died out in the face of Spanish. There’s also a big break across the far south of Peru and northern Bolivia where Aymara intervenes.

I’ve recently completed a whole new website on regional diversity in pronunciation, so you can now hear online the Sounds of the Andean Languages at www.quechua.org.uk/Eng/Sounds/ and read a layman’s account of their origins at . For for a briefer overview click here for a few more details on this.

My existing pages do have some details on some of these questions: see the history section on my page on Introduction to the Indigenous Languages of the Americas and the relationships with other languages section on my page Quechua - What's it Like?

I will eventually add full subsections to this page about the main outstanding questions in:

• historical development of the Quechua language family

• dialect classification and geography

• ‘genetic’ relationships with other languages

• contact relationships with other languages

• particular phenomena in (historical) language change.

To read on the web a full article by the Jaqaru expert Martha Hardman entitled “Aymara and Quechua : Languages in Contact”, where she explains why she rejects claims of a demonstrable common origin of the Quechua and Aymara (or in her terminology, ‘Jaqi’) language families, click on:

www.aymara.org/biblio/quechua.html

Back to Contents - Last Section in Document

Sociolinguistics (including Bilingual Education)

The sociolinguistic position of Quechua is absolutely fascinating. The language is very much a low status one, but there is something of an attempt at restoring its prestige, though as yet without much success. There are also big differences between the situations in the various countries where it is spoken: Peru, Bolivia, Ecuador and Argentina.

For an introduction to the sociolinguistic situation of Quechua and its speakers, see the sections on Who Speaks Quechua?, Social Situation, The Old World Meets the New, and Survival of Indigenous Languages on my webpage on an Introduction to the Indigenous Languages of the Americas, and my page on How Endangered is Quechua? Note also the big international conference of Americanists coming up in July 2003, which will touch heavily on the sociolinguistics of Quechua – there’s a call for papers out now too.

There is so much to say on the sociolinguistics of Quechua that I’ll be expanding this section here when I can, but for now just briefly here are a few topics for which Quechua can provide some ideal case studies:

• Moves to introduce and promote bilingual education in Quechua have been pretty half-hearted at best, given how political institutions, not least Education Ministries, are generally run by Spanish monolinguals who have very little to do with Quechua-speaking communities, and often a pretty dim view of them.

• There’s also the problem of the Cuzco Academy getting on its high horse and telling speakers of all other dialects of Quechua that theirs is just a debased form of their own ‘pure, imperial’ Quechua. Historical and linguistic nonsense, of course, but damagingly powerful in the Andean environment.

• Another fascinating issue is the massive loss of Quechua in any urban environments, the principal reason for its decline in a world of rapidly increasing urbanisation. I know of a Ph.D. thesis on the loss of Quechua among migrants to Lima, for example, entitled ‘The Language Left in Ticlio’, by Dr Tim Marr.

• Note also that the traditional European dialectology principle that for the truest speakers least affected by the standard language are ‘NORMs’ – i.e. Non-mobile Older Rural Males – is turned on its head in the Andes, at least if one sees the role of Spanish similarly to that of a ‘standard’ form of a language vis‑à‑vis its dialects. In the villages where Quechua is strongest, it is the women who speak the form least influenced by Spanish, precisely because they are often pretty much monolingual in Quechua while men are generally much more competent in Spanish (partly because they tend to be more mobile, and go to school and sometimes military service).

There eventually will be more coverage of Quechua sociolinguistics here, but for now here are some pointers to useful works. The first one contains a whole collection of articles on issues of language and identity in the Andes.

Coronel-Molina, Serafin M. & Linda L. Grabner-Coronel (eds.) (2005)
Lenguas e identidades en los Andes: perspectivas ideológicas y culturales.
Ediciones Abya Yala: Quito
In: Spanish Availability: Abya Yala
Click here for a full presentation of this work

Coronel-Molina, Serafin M. (1999)
Functional Domains of the Quechua Language in Peru: Issues of Status Planning
in: International Journal of Bilingual Education and Bilingualism - 2(3): 166-180
Multilingual Matters

Hornberger, Nancy H. & Kendall A. King (1998)
Authenticity and Unification in Quechua Language Planning
in: Language Culture and Curriculum - 11(3): 390-410
Multilingual Matters

You may also find the position, role and attitudes of the Quechua Language Academy in Cuzco very interesting!

Back to Contents

Back to Homepage

STOPS and	voiceless	p	t	tS	k	q
AFFRICATES	ejective	p’	t’	tS'	k’	q’
	aspirated	pH	tH	tSH	kH	qH
FRICATIVES	voiceless		s	(S ?)			h
NASALS		m	n
LATERALS			l	´
FLAP			r
GLIDES		w		j