Adelaar Willem, Muysken Pieter. The languages of the Andes

Подождите немного. Документ загружается.

84 2 The Chibcha Sphere

palatalised stop [t

] (cf. Constenla Uma˜na 1984: 75–9).

It should be noticed, however,

that Lugo uses both the symbols <c

−

h> and <ch> (cf. Ostler 1995: 134). In spite of

some obvious misspellings and inconsistencies in the use of the two symbols at issue,

Lugo shows a clear preference for the use of <ch> before (e.g. in chi- ‘our’, chiˆe

‘we’) and of <c

−

h> before any other vowel (e.g. in c

−

huen



ha ‘bad’, c

−



‘priest’).

Ostler notes a number of exceptions in which <c

−

h> precedes , such as c

−

hic

−

hua ‘to

learn’ and c

−

hibc

−

ha ‘Chibcha’. Still, it may be argued that <ch> could have been a palatal

affricate (as in Spanish), whereas <c

−

h> represented some other sound, for instance, [t

Lugo’s symbol <



h> is replaced by <z> in the anonymous grammars (where Lugo’s

−

huen



ha becomes chuenza). It probably represented an alveodental affricate [t

], as

Constenla Uma˜na (1984: 79–85) argues.

Lugo’s symbol <> represented a vowel reportedly intermediate between [e] and [i].

On the basis of comparative considerations and the phonetic descriptions of the time,

Constenla Uma˜na (1984: 93–7) assigns it the value of a high central vowel []. In the other

sources <

> has been replaced by <y>; thus, Lugo’s c

−



becomes chyquy.

When

/y> follows a labial consonant in writing, a symbol is always inserted; e.g. in



sca / muysca ‘human being’. It may have represented a non-syllabic labial element

subsidiary to the realisation of the high central vowel (/mska/ [m

ska]). The fact that

Lugo chose to use a special symbol for the high central vowel turned out felicitous in

that it avoids confusion with the high front vowel [i] and its non-syllabic correlate [y].

Both could be written as well as <y> in colonial Spanish sources, as in fact they

were in Lugo.

Muisca had three simple stop consonants in the labial, alveolar and velar articulatory

positions: [p], [t] and [k], respectively. For the [k] three different spellings were in use:

<c>, <q> and <qu>. Lugo used <q> before <

> and before <h> (e.g. c

−



‘priest’,

qhic

−

ha ‘ten’), but, conforming to the Spanish orthography, he used <qu> before <e>

and , and <c> elsewhere. Before other vowels <qu> represented a labialised velar

sound [k

], as in





squˆa ‘I do’. In the National Library grammar <qu> [k] also

appears before <y> (Lugo’s <

>), and <q> does not occur by itself; for its treatment

of Lugo’s <qh> see below.

The passage (our translation) reads: ‘The pronunciation of the syllables cha, che, chi, cho, chu

should not be done with the tongue as a whole but just with the tip of it.’

The awkwardness of Lugo’s representation of the complex obstruents [t

] and [t

]isreminiscent

of the way in which la Carrera sought to describe comparable sounds in Mochica (by means of

the symbols <cɥ> and <tzh>, respectively; cf. section 3.4).

The word chyquy ‘priest’ originally denoted representatives of the native religion, but was

subsequently also used for Roman Catholic clergymen. Its Spanish corruption jeque ‘sheikh’

presents an interesting case of linguistic interaction. In colonial New Granada native priests

were usually referred to as moh´an.

2.9 The Muisca language 85

In addition to the three plain stops, Muisca had a complex obstruent, both in a plain

and in a labialised version. It was usually written <pq(u)>,asinpqua [pk

a] ‘tongue’,

pqueta [pketa] ‘silly’ and pquyquy [pkk] ‘understanding’. The plain version occurred

before [e], [i] and [], the labialised version before [a] and [o]. This sound may have

represented a coarticulated stop ([pk], [pk

]) and is mentioned in the National Library

grammar as one of the special sounds (pronunciaciones particulares)ofthe language

(Gonz´alez de P´erez 1987: 72; cf. also Constenla Uma˜na 1984: 74). Notwithstanding the

fact that many other exotic consonant clusters were found in Muisca, there are good

reasons to assign a special status to this combination. At least in some lexical items

the presumed coarticulated stop was a reﬂex of *ku-,asMuisca has pk

(w)

V where other

Chibchan languages have k

V or kuwV (Constenla Uma˜na 1989: 43–4). Compare, for in-

stance, Muisca pqua, Chimila k

´a:

, Cuna k

apinni [k

a:bin], Uw Cuwa (Tunebo) k`uwa

‘tongue’; Muisca pquaca,UwCuwa k

´ıka ‘arm’; Muisca pquyhyxio,UwCuwa k

as´aya

‘white’; Muisca pquyquy ‘understanding’, Cuna k

ake [k

a:ge] ‘heart’; Muisca fapqua

‘chicha’, Uw Cuwa b´ak

a (Huber and Reed 1992). The presence of a coarticulated stop

is limited to Muisca and Duit. It has not been found in other Chibchan languages.

Six symbols that were used in the transcription of Muisca have been interpreted as rep-

resenting fricatives: , <f>, <s>, <x>, <g> and <h>. Constenla Uma˜na (1984:

74) observes that and <f> were never used contrastively. These symbols may have

referred to voiced [β] and voiceless [ϕ] bilabial fricatives, respectively, presumably al-

lophones of a single phoneme. It should be observed that <f> was only used before

avowel symbol, whereas occurred in any position, before another consonant as

well as word-ﬁnally. Lugo practically limited the use of <f> (as an alternative for )

to the position before (e.g. fuc

−

ha ‘woman’, but cubun ‘language’), whereas the

other sources extended the use of <f> to other prevocalic positions as well (e.g. fac

‘outside’, ﬁhista ‘breast’). As it seems, some orthographic normalisation occurred to the

extent that <f> became used mainly (but not exclusively) in root-initial position before

avowel.

The identiﬁcation of and <f> as identical elements is challenged by

the existence of a word-initial cluster <bf>,which is found in the word bfue ‘beam’.

Possibly, this cluster may have to be reinterpreted to contain an internal vowel [βϕue].

The symbols <s> and <x> represented sibilants (presumably [s] and [ˇs], respec-

tively). Constenla Uma˜na (1984: 86) considers them not to be contrastive. The symbol

<x> was used optionally before and <y> (e.g. xie ‘who?’, bxy ‘I carry’), whereas

<s> could occur in all positions (e.g. saca ‘nose’, soco ‘bring!’, sisy ‘this’, muysca

‘human being’). Although in many lexical items (e.g. sie ∼ xie ‘river’, ‘water’; also

‘who’) the two symbols can be used interchangeably, in other cases there seems to be

a preference for either one. Word pairs such as siu ‘rain’ versus xiun ‘sweat’, and sua

Today f is widely used in place names of Chibchan origin, such as Facatativ´a and Fusagasug´a.

86 2 The Chibcha Sphere

‘sun’ versus xua ‘dew’ do suggest an opposition, albeit of a limited functional load. In

Lugo <x> is mainly, and fairly consistently, found before .

The symbol <g> is interpreted by Constenla Uma˜na (1984: 88–90) as a voiced velar

fricative [γ ], historically derived from a Chibchan voiced velar stop *g; e.g. gye [γ e]

‘excrement’ (cf. Chimila ga:, Kogui gai). The assumption that <g> wasafricative is

based on a probable symmetry with the labials (<b/f>) and on the pronunciation of the

corresponding symbol in Spanish. As in Spanish, <gu> rather than <g>,was written

before <e>, and, in Lugo, also before <

>; e.g. in gue ∼ gu



‘to be’, guity- ‘to

whip’.

Even though the regular realisation of <g> was allegedly a fricative, a stop

allophone may have occurred instead after a nasal, for instance, in the verbal future

ending -nga.

With the complex symbol <gu> we touch upon an insufﬁciency of the colonial Muisca

orthography because an identical sequence was used to represent a bilabial glide [w] or a

labialised voiced velar fricative [γ

], e.g. in gue [γ

e ∼ we] ‘house’, gui ‘wife’ [γ

i ∼

wi]. In such cases the morphophonemic behaviour of the lexical item in question must be

considered in order to recover the correct form. For instance, the ﬁrst-person possessed

form of ‘house’ is zue [t

we]‘my house’, not *zgue [t

γ e], as might be expected if the

velar element were prominent. Lugo, who used diacritics, appears to have indicated the

presence of a labial element by a circumﬂex accent on the following vowel (guˆe, guˆı ),

although not consistently. The question whether the velar element was pronounced at

all in such cases remains open for discussion. Lugo occasionally used the spelling -guˆa

for the interrogative sufﬁx [wa]. Some Chibchan cognate relations, such as Muisca gua

‘ﬁsh’ (Cuna ua, Chimila wa:

gra:, Kogui uaka), suggest that a velar fricative was not

necessarily always represented in <gu>.

All colonial sources agree that the symbol <h> represented an aspiration. In con-

tradistinction to <gu>, the combination <hu> was not normally used to denote a labial

approximant [w], as in the orthography of so many other Amerindian languages during

the colonial period, e.g. hui ‘inside’ was pronounced [huy ∼ huwi], not *[wi]. The

Muisca aspiration occurred prevocalically, but characteristically also between two like

vowels. Sequences of identical high vowels separated by <h> may have counted as

single vowels. This can be deduced from the fact that the tense–aspect sufﬁx -squa,

The vocabulary of the National Library grammar mentions a related verb root uity- [wit], which

has the meaning ‘to whip/chastise oneself.’ This unique correlation does not prevent us from

following the general opinion that guity-was pronounced [γ it], not [γ

it] nor [wit]. The shape

of the preﬁxes that accompany guity- supports this view.

Constenla Uma˜na (1984: 98) mentions the example of the root gua- ‘to kill’ as a case where a

velar fricative could have been retained (by comparison with Guatuso kua:, also Kogui g

´aˇsi).

However, the root for ‘to kill’ in Muisca is gu-, rather than gua-. The confusion is probably due

to Lugo (1619: 77–8), who translates the verb for ‘to feed’ gua-squˆa as ‘to kill’, ‘to say’, glosses

which correspond to gu-squˆa;but see p. 70, where he translates it correctly. In the Muisca stem

(b)gu- ‘to kill’ a velar fricative was indeed present.

2.9 The Muisca language 87

otherwise only occurring after monosyllabic roots ending in a vowel, was also regularly

found after roots that ended in such a sequence, e.g. b-chuhu-squa ‘to wash’, b-chihi-

squa ‘to write’, ‘to paint’, tyhy-squa ‘to sit down’.

This situation suggests a special

class of aspirated vowels. In Lugo’s grammar the <VhV> sequence is relatively in-

frequent, omission of its ﬁrst vowel being the rule, as, for instance, in shˆu



hˆa ‘eight’,

qhˆuma ‘big’, th



pqua- ‘to wound’ (in the vocabulary of the National Library grammar:

suhuza, cuhuma, b-tyhypqua-, respectively) and bhacˆuca ‘needlessly’ ( fahacuca in the

vocabulary published by Quesada Pacheco 1991). In some items Lugo hesitates between

notations of the types <hV> and <hVhV>,asinthe postposition bhˆohˆo



hˆa ∼ bhˆo



hˆa

‘with’ (bohoza), or inserts an <h> where others do not have it, as, for instance, in th





‘to love’ (b-tyzy-).

The nasals <m> and <n> were both very frequent. The symbol <r>,which presum-

ably represented a vibrant, is occasionally found in Lugo, but it is nearly absent in the

other sources, who replace it with <ch>; e.g. (Lugo)



erq



[erk], (Gonz´alez de P´erez

1987: 286) yechyc [et

k] ‘other’.

At least in borrowings from Spanish it survived, as

can be deduced from a case like raga ‘dagger’ (Spanish daga). There were no laterals

in Muisca.

In addition to the vowel <

/y> (see above), the sources of the Muisca language

record ﬁve vowels that were similar to their Spanish counterparts: a, e, i, o, u. The vowel

symbols and can refer both to syllabic [i], [u], and to non-syllabic [y], [w].

An analysis of the lexical and grammatical data brings to light the necessity to distinguish

between the two possibilities in writing, as their realisation is not entirely predictable.

For instance, the spelling ja of the word for ‘ﬁrewood’ (also found as ia, cf. Uw Cuwa

r´eya)inthe vocabulary accompanying the National Library grammar (Gonz´alez de

P´erez 1987: 273) strongly suggests the presence of a non-syllabic approximant [ya].

By contrast, the word for ‘illness’ iu must be interpreted as polysyllabic [iyu] on the

basis of its morphophonemic behaviour.

This option appears to be conﬁrmed by the

1612 Vocabulary, in which the form at issue is transcribed as i¨u, suggesting that the two

vowels were pronounced separately (Quesada Pacheco 1991: 62). In the Royal Palace

Library grammar published by Lucena Salmoral (1967: 67) the past participle of the verb

m-u- ‘to spin’ is explicitly written as ¨uisca,again suggesting that the vowel sequence

The ending -squa is also found after the verbaliser -go-, e.g. in muysy-go-squa ‘to dream’

(cf. Adelaar 1995a). In the Muisca vocabularies all non-stative verbs are presented with either

the ending -squa or -suca.

The word



erq



is found in one of the sonnets that accompany Lugo’s grammar (Ostler 1995:

136).

In word-initial position, is often replaced with its orthographic variant <v>, especially

when a full vowel [u] is represented.

The verb iu- ‘to be ill’ takes the imperfective sufﬁx -suca, not its alternative form -squa. Since

the latter is preferred after monosyllabic roots ending in a vowel, the use of -suca suggests a

polysyllabic structure for the root iu (cf. Adelaar 1995a).

88 2 The Chibcha Sphere

Table 2.6 Inventory of Muisca consonant phonemes

Bilabial Alveolar Palatal Velar Coarticulated labiovelar Glottal

Stops p t k pk

/pk

Affricates c t

/ˇc

Fricatives β / ϕ s(ˇs) γ h

Nasals m n

Vibrant r

Glides w y

Table 2.7 Inventory of Muisca vowel phonemes

Front Central Back

High i  u

Mid e o

Low a

Aspirated versus plain

may have been polysyllabic [uwi]. Finally, it is likely that gui [γ

i ∼ wi] ‘wife’ and ui

[uy ∼ uwi] ‘left side’ were distinct not only in writing, but also in pronunciation.

A putative inventory of the Muisca consonant phonemes is shown in table 2.6. It con-

tains the phonemes proposed in Constenla Uma˜na (1984) expanded with our additions.

The vowels of Muisca are shown in table 2.7.

Consonant clusters frequently occur both in word-initial and -ﬁnal position (cf.

Constenla Uma˜na 1984: 100–1). Most (but not all) of the initial clusters owe their

existence to the presence of a preﬁx, in particular the verbal preﬁx b-/m-,which can

indicate transitivity, and the personal reference markers z- ‘ﬁrst person singular’ and

m- ‘second person singular’, which indicate both possessor and subject. These personal

reference markers are also found as ze-/zy- and um-, respectively, making it clear that

if an additional vowel was not always present, it was at least recoverable.

In (59) the

preﬁx ze-isgiven in its full form, whereas m- is short.

(59) m-hu-za-c ze-guque

2S.SG-come-NE-AL 1S.SG-say.PA

‘I thought that you had not come.’

(Gonz´alez de P´erez 1987: 257)

Lugo writes





- for the ﬁrst-person preﬁx. Remember also that um- is frequently represented

as vm-;inour discussion of Muisca grammar, we shall write um- throughout.

2.9 The Muisca language 89

The preﬁx b-/m- is frequently found before other consonants. Although there seems

to be no vowel that can be recovered, certain combinations with b-,inparticular, are

awkward to pronounce if no vowel is inserted (see also above for the unusual case of

bfue).

Some examples of clusters involving b-/m- are given in (60). The function of

the preﬁx will be treated later.

(60) b-so- ‘to eat (herbs, leaves)’

b-ca- ‘to eat (maize, meat, fruit, etc.)’

b-chichua- ‘to learn’

b-hu- ‘to carry’

m-nypqua- ‘to hear’, ‘to understand’

Word-ﬁnal consonant clusters are frequently found in genitives which precede their

heads. When the genitive consists of a noun with a ﬁnal vowel a, that vowel may be

suppressed. As a result, a consonant cluster can end up in word-ﬁnal position, e.g.

in (61).

(61) muysc chimy

[muysca ‘human being’]

human.being.G ﬂesh

‘human ﬂesh’

(Gonz´alez de P´erez 1987: 210)

2.9.3 Grammar

In a review of previous syntactical claims concerning the Muisca language Ostler (1994)

concludes that Muisca is a strict SOV language. The auxiliary verb gue/gu



follows the

main verb. Both adjectives and numerals follow the noun they modify, while demon-

stratives and genitives precede the noun they modify. Muisca has postpositions and case

sufﬁxes. All these features are characteristic for most of the Chibchan languages that we

have discussed until now.

From a morphological point of view Muisca is a complex language that uses both

preﬁxes and sufﬁxes. The preﬁxes are mainly related to personal reference and valency

(passive, transitive–intransitive). Nominalisation, verbalisation, negation, interrogation,

topicalisation, mood, tense, aspect, case and degree (of adjectives) are indicated by

means of sufﬁxes. Other morphological processes found in Muisca are root suppletion,

internal vowel change and the suppression of ﬁnal vowels, as in (61). Some parts of the

verbal conjugation, in particular the present and past participles, the so-called second

supine, and the imperative can be highly irregular.

Spanish colonial grammarians often tolerated consonant sequences which their modern suc-

cessors prefer to split by inserting shwa-type vowels; for a similar example in Araucanian see

chapter 5.

90 2 The Chibcha Sphere

The Muisca verbal lexicon comprises a formal division between transitive and in-

transitive verbs. This division is made visible by the conjugational behaviour of the

verbs, and in most cases also by the shape of the verb stem itself. Many transitive verb

stems in Muisca contain the preﬁx b-/m-; see also (60). If the transitive verb stem is

part of a transitive/intransitive verb pair, its intransitive counterpart does not have the

b-/m- preﬁx. Speaking generally, the m- allomorph is found before vowels and nasals,

whereas b- occurs in clusters with non-nasal consonants. It has been argued that, since

m- is the prevocalic allomorph (where no inﬂuence from the phonological environment

is expected), it also constitutes the more basic form of the preﬁx (Ostler 2000: 288).

Transitive–intransitive verb pairs differentiated by b-/m- are exempliﬁed in (62) and

(63). The latter also exhibits a root vowel change.

(62) to- ‘to split’ [intransitive]

b-to- ‘to split’ [transitive]

(63) na- ‘to go’ [intransitive]

m-ny- ‘to take along’ [transitive]

Transitive verb stems that are not paired with intransitives may contain the b-/m-

preﬁx, although many of them do not. All transitive verbs lose their b-/m- preﬁx in the

imperative, and in several participial and gerund forms (64). Some intransitive verbs

also have an initial b-/m- element, which behaves as if it were part of the root and is

never lost. Intransitive verbs of any type take an additional preﬁx a- in the imperative

form (65) and occasionally also in one of the participles.

(64) b-quy- ‘to do’ [transitive]

quy-u ‘do!’ [transitive imperative]

quy-ia ‘having done’ [transitive past participle]

(65) bgy- ‘to die’ [intransitive]

a-bgy-u ‘die!’ [intransitive imperative]

Examples (66) and (67) illustrate the case of a transitive–intransitive pair with an ir-

regular imperative form, where it can be seen that the preﬁx a- serves to avoid homonymy

between a transitive and an intransitive form. The most common imperative ending is

-u (plural -u-ua). For negative commands the negative future is used.

(66) b-ga- ‘to make’, ‘to cause to become’ [transitive]

so ‘make!’, ‘cause to become!’ [transitive imperative]

(67) ga- ‘to become’ [intransitive]

a-so ‘become!’ [intransitive imperative]

2.9 The Muisca language 91

There are other formal strategies to distinguish transitive from intransitive verbs. The

latter may take a sufﬁx -n-, not found in the former, e.g. miu- ‘to crumble (transitive)’

versus miu-n- ‘to crumble (intransitive)’.

A number of Muisca verbs are stative. They refer to existence or position and may

encode number or shape of the theme. Most of these verbs end in -ne, e.g. zo-ne ‘to be lo-

cated (singular)’, puy-ne ‘to be located (of liquids)’ (Ostler and Gonz´alez forthcoming).

The majority of Muisca verbs, however, are active. Active verbs are conjugated for tense.

They can take either one of the characteristic endings -squa and -suca,which is the form

by which they are normally found in the colonial dictionaries. The endings -squa and

-suca indicate imperfective aspect in ﬁnite verbs, which can then be interpreted as present

or (ongoing) past-tense forms.

(68) ze-bquy-squa

1S.SG-do-IA

‘I do’, ‘I was doing.’

(Gonz´alez de P´erez 1987: 81)

(69) ze-guity-suca

1S.SG-whip-IA

‘I whip’, ‘I was whipping.’

(Gonz´alez de P´erez 1987: 84)

As we have noted before, the choice of -squa is phonologically determined (see

also Adam 1878). For -suca there is no such restriction. Since it is added to all active

verb roots that cannot take -squa,itappears to have a default distribution. In a few cases

-suca has been found with a root that normally takes -squa, allegedly with a frequentative

meaning (70), (71).

(70) ze-hu-squa

1S.SG-come-IA

‘I come.’

(Gonz´alez de P´erez 1987: 332)

(71) ze-hu-suca

1S.SG-come-IA

‘I come often.’

(Gonz´alez de P´erez 1987: 332)

It may be observed that examples (70) and (71), originally from the National Library

grammar, seem to be incompatible with a statement made in the same source, namely,

(70) and (71) are entries in Gonz´alez de P´erez’s edition of the vocabulary of the National Library

grammar, where they appear as zchusqua and zchusuca, respectively. Elsewhere in the same work

(Gonz´alez de P´erez 1987: 98, 121) we ﬁnd zehusqua with the meaning ‘to come’. We assume

that zchusqua and zchusuca, not attested in any other source as far as we know, are miswritings

for zehusqua and zehusuca, respectively. Uricoechea (1871: 204), who must have used the same

source, writes zuhusqua/zuhusuca in accordance with a pronunciation rule formulated in the

National Library grammar (Gonz´alez de P´erez 1987: 140).

92 2 The Chibcha Sphere

that ze-hu-squa serves as the frequentative counterpart of the defective verb i-xyquy ‘I

come’ (Gonz´alez de P´erez 1987: 121).

A frequentative meaning also obtains when -suca is added to a stative stem (72), (73).

(72) i-sucu-n

1S.SG-be.there-ST

‘I am there.’

(Gonz´alez de P´erez 1987: 261)

(73) i-sucu-n-suca

1S.SG-be.there-ST-IA

‘I am there frequently.’

(Gonz´alez de P´erez 1987: 261)

Some verbs, such as -na ‘to go’, can have a present-tense interpretation, even when

the imperfective sufﬁx is lacking. If there is an imperfective sufﬁx, it has a frequentative

meaning, even though it is -squa.

(74) i-na

1S.SG-go

‘I go’, ‘I went.’

(Gonz´alez de P´erez 1987: 123)

(75) i-na-squa

1S.SG-go-IA

‘I go often.’

(Gonz´alez de P´erez 1987: 123)

The indicative future tense of verbs is formed by adding -nga to those verbs that can

take -squa in the present tense, and -nynga to those that can only take -suca; for a historical

phonological explanation of this phenomenon see Ostler and Gonz´alez (forthcoming).

Stative verbs change -ne to -ne-nga (76). The future form of the verb ‘to be’ is nga.

(76) muysca-c ze-gue-ne-nga

man-AL 1S.SG-be-ST-F

‘I will be a man (person).’

(Gonz´alez de P´erez 1987: 127)

The past tense has no speciﬁc ending. Verb roots which end in -a- normally take

an (optional) sufﬁx -o [w] (77); some roots (including a few in -a-) take -quy (∼-que)

(78); the others remain unaltered.

Before the negative sufﬁx -za most roots remain

unaltered (79).

There was probably no difference in use between the endings -n and -ne (∼-ny) (Ostler and

Gonz´alez forthcoming).

The velar stop in -quy (∼-que)may also appear in other forms of the paradigm, namely in the

imperative and past participle. From a historical point of view, it can be considered as a part of

the root (Ostler and Gonz´alez, forthcoming).

2.9 The Muisca language 93

(77) ze-mnypqua(-o)

1S.SG-hear/understand-PA

‘I have heard/understood.’

(Gonz´alez de P´erez 1987: 98)

(78) um-gui boza um-caquy-oa

2P.SG-wife with 2S.SG-ﬁght.PA-IR

‘Did you ﬁght with your wife?’

(Gonz´alez de P´erez 1987: 356)

(79) ze-gu-za [ze-guquy ‘I said’]

1S.SG-say.PA-NE

‘I did not say.’

(Gonz´alez de P´erez 1987: 115)

Interrogation and negation are indicated by adding the sufﬁxes -ua [wa]/-o

and -za,

respectively, to a verb form. The future tense has a special negative ending -zi-nga (not

*-nga-za). The auxiliary verb gue/gu



has as its interrogative and negative counterparts

ua/o and nza, respectively; the corresponding future forms are nnua and nzinga.Alonger

and more explicit negation of the verb ‘to be’ is ma-gue-za.Itincludes a negative preﬁx

ma-oflimited productivity.

(80) hycha-n i-na-zi-nga

I-TO 1S.SG-go-NE-F

‘I will not go.’

(Gonz´alez de P´erez 1987: 335)

(81) chie nnua

we be.F.IR

‘Shall we be?’

(Gonz´alez de P´erez 1987: 78)

(82) chie nzinga

we be.NE.F

‘We shall not be.’

(Gonz´alez de P´erez 1987: 79)

Muisca has three nominalisations or relative verbs which have been interpreted tradi-

tionally as participles (agentives) of the present, future and past. With all three, the b-/m-

preﬁx is lost. The present participle can be formed by adding a sufﬁx -sca,very often

accompanied by a vowel change in the root (83). This sufﬁx -sca can only be added to

roots that have the phonological characteristics allowing the presence of the imperfec-

tive marker -squa (see above). The other roots, namely, those that take the imperfective

marker -suca, are treated differently. They receive either a sufﬁx -suca (transitives) (84),

or -uca (intransitives) (85).

The distribution of -ua in relation to -o is not entirely clear; it may have been a case of free

variation (Ostler and Gonz´alez forthcoming). The marker -o is reminiscent of the interrogative

marker -o in Ika.