The Origin of Kinyarwanda Function Words
Alexandre Kimenyi

The study of Kinyarwanda function words: complementizers, subordinators, conjunctions, prepositions, clitics and affixes shows that they come from content words. Complementizers, subordinators and conjunctions look exactly like regular verbs in their finite or non-conjugated forms. Prepositons have the same structure as prepositional phrases. Conditionals and concessives have the same form as intermediate categories namely demonstratives or pronouns. This clearly indicates that these words are a linguistic innovation in Kinyarwanda grammar. These findings also clarify the phenomenon of the genesis of signs and structures and their dynamism. More specifically, they support Peirce's theory of semiotics which states that most of the signs and structures are not arbitrary but start at icons and later on become symbols because of semiosis. They also support the theory of grammaticization which shows this phenomenon not to be unique to Kinyarwanda because a crosslinguistic study of function words (Givón, Heine, Trugott&Hopper, etc.) shows indeed that most of the World's language function words are not primitive categories but derive from nouns and verbs which refer to the most basic concepts and events.

This essay consists of four parts. The first part discusses function words which derive from verbs, namely regular verb forms and irregular ones. The second part is about function words which come from nouns and intermediate categories such as adjectives, pronouns and demontratives, the third category is about clitics and the last is about affixes which also come from independent words.

I. Deverbal function words

Deverbal function words include conjunctions namely coordinators and subordinators but also some adverbs. According to their degree of opacity or transparency, they can be classified into three categories namely (a) infinitive verbs, (b) finite verb verbs, and (c) opaque or irregular verb forms.
These distinct phonetic forms as will be shown later, correspond to the different stages of sign genesis and dynamism: transparency or iconicity, polysemy or category hibridization, phonetic change and desemanticization.

I.1. Conjunctions From Infinitive Verbs:

Conjunctions from infinitive verbs look exactly like other infinitives. The infinitive verb consists of three morphemes: the prefix ku- which is also realized as gu- before voiceless consonants because of Dalh's law, the verb stem and the imperfect aspect suffix -a. Conjunctions are given below from (1) through (6) on the left and their infinitive verb counterparts are on the right:

1. kubá 'that' ---- kubá 'to be'
a. Ku-bá bá-taa-j-e by-aa-m-bába-j-e.
to-be they-not-come it-past-me-suffer-caus-asp
'That they didn't come upset me.'
b. A-ra-shaak-a ku-bá umwáarimú.
s/he-pres-want-asp to-be teacher
'He wants to be a teacher.'
2. kubéera 'because' ---- kubéera 'to be for, to give favors'
a. Nti-y-iig-á ku-béer-a akazi.
not-s/he-study-asp to-be-for work
'He doesn't study because of work.'
b. A-ra-shaak-a ku-béer-a umugoré.
s/he-pres-want-asp to-give-favors woman
'He want to favor the woman.'
c. Nt-aa-kor-á ku-béer-a kó a-naní-w-e.
not-s/he-work-asp to-favor that s/he-tire-pass-asp
'He doesn't work because he is tired.'
3. kugeza 'until/to' ---- kugeza 'to arrive with'
a. A-ra-sóbaanur-a ku-gez-a ahó mw-uumví-ye.
s/he-pres-explain-asp to-arrive where/when you-understand-asp
'He is explaining until you understand.'
b. A-shobor-a ku-gez-a iki gití aho.
s/he-can-asp to-arrive-asp this tree there
'He can bring this piece of wood there.'
4. kuva 'since/from' ---- kuva 'to come from/leave'
a. Ku-v-a ahó a-bon-é-ye akazi ntí-tu-ki-mu-bon-a.
to-come where/when s/he-see-appl-asp work not-we-anymore-him/her-see-asp
'Since he found a job we don't see him.'
b. A-raa-z-a ku-v-a inó bí-tiinz-e.
s/he-pres-come-asp to-leave-asp here it-get late-asp
'He is going to leave here late.'
5. kugira ngo 'for' ---- kugira ngo 'to think'
a. A-raa-j-e ku-gir-a ngo mu-bon-án-e.
s/he-pres-come-asp to-cause-asp that you-see-rec-subj
'He is coming so that you can meet.'
b. A-kuund-a ku-gir-a ngo sii-n-kor-á.
s/he-like-asp to-cause-asp that not-I-work-asp
'He likes to think that I don't work.'

6. ahagána 'towards'
a. Azaaza ahagána saa taánu ya nímugórooba
s/he-fut-come towards hour five evening
'S/he will come around five in the afternoon'.
b. Ba-a-huur-i-ye a-ha-gán-a i Butare. they-pst-meet-ben-asp it-there-go in
the direction of-asp at Butare
'They met towards Butare'
For all these examples, it is not clear yet whether we are dealing with two distinct categories: namely infinitive verbs and conjunctions or polyfunctional words: one single lexical entry with different
categories and functions.

1.2. Conjunctions from finite verbs:

Most of the conjunctions derived from finite verbs occur with the second person singular pronoun u- in the present indicative or the conditional form:

6. urétse 'except' urétse /u-rek-ye/ 'if you leave aside'

_Kor-a akazi k-aa-we u-b-é u-réts-e icyo gitabo.
do work of-you you-be-subj you-leave-subj that book
'Do your work, leave that book for a while'.
_Bóose ba-ára-hámaga-ye u-réts-e umugoré.
all they-past-call-asp you-leave-subj woman
'Everybody called except the woman'.

urétse kó + finite verb 'except'
you-leave-asp
usíibye 'except'
usíibye /u-siib-ye/
'if you stay/are absent'
ndetsé 'also/too'
ndetsé /n-rek-ye/
'that I left aside'
Ndakéeka kó _n-dets-é kugeenda uyu muunsi.
I-think that I-stop-asp to-go this day
'I think that I stop going today'
_Abáana ba-a-guu-ye ndetsé ba-ra-ná-komerek-a.
children they-pst-fall-asp also they-t-also-wound-asp
'The children fell down and they also were wounded'.
_U-síib-ye muu náama, abaándi ba-a-raakar-a.
you-be absent-asp in meeting others they-cond-be angry-asp
'If you missed the meeting, everyone else would be angry'.
_Abáana ba-a-ríi-ye u-síib-ye babiri.
children they-past-eat-asp you-be absent-asp two
'The children ate except two.
usíibye kó + finite verb 'except that'
you-be absent-asp
_U-síib-ye kó bi-táa-b-á byiizá!
you-be absent-asp that it-neg-be-asp good
'It would not be good if you were absent'.
_Ni ahaantu heezá u-síib-ye kó ha-koonj-á cyaane.
be place good you-be absent-asp that it-be cold-asp very
'It is a nice place except that it is very cold'.
ukúuye hó kó 'if you remove that'
ukúuye hó kó 'except that'
uváanye hó kó 'if you remove that'
uváanye hó kó 'except that'
_Avuze kó u-kúu-ye hó/u-váan-ye hó abo baantu kó by-aamer-á náabí
s/he said that you-remove-asp those people that it-become-asp bad
'S/he just said if you removed those people things will be very bad'
_Ba-kuund-a uwo mwáana u-kúu-ye hó kó/u-váan-ye hó kó ba-ta-mu-fásh-a.
they-like-asp that child you-remove-asp that they-not-him-help-asp
'They like that child except that they don't help him'.

umenya 'perhaps'
umenya /u-meny-a/ 'you know'

_Mbeesé wowé u-meny-a ahó ba-tuu-yé?
Q you you-know-asp where they-live-asp
'Do you know where they live'?
_Abagoré ba-a-ri bá-raaka-ye u-meny-a bá-ta-záa-garuk-a.
women they-past-be they-be angry-asp you-know-asp they-neg-fut-return-asp
'The women were angry maybe they will not come back'
umenya ngo 'apparently' you-know-asp
_Ukó m-by-uúmv-a, u-meny-a ngo abagoré bá-ta-záa-garuk-a.
as I-it-hear-asp you-know-asp that women they-not-fut-return-asp
'What I hear is that apparently the women will not come back'.

ubaanza 'perhaps'
ubaanza /u-baanz-a/ 'you start'
ubaanza ngo 'apparently'
you-start-asp

_Iyó u-jy-a gutwáara imódoká u-baanz-a kwaatsa motéeri.
when you-go-asp drive car you-start-asp warm engine
'When you are going to drive a car you start the engine first'
_Imódoká y-a-anz-e kwaaka u-baanz-a motéeri y-ará-pfúu-ye.
car it-refuse-asp light you-start-asp engine it-past-die-asp
'The car has refused to start maybe the engine is dead'.
_Mariyá ntí-y-aaj-e kw'iishuúri u-baanz-a ngo a-rwáa-ye.
Mary not-she-come-asp at school you-start-asp ngo she-be sick-asp

ubooshye (kó) 'as if'
?uboóshye /u-ba-oshy-e/ 'mislead them'
you-them-mislead-asp

_U-ra-kúbit-a umwáana gútyo u-booshy-e kó a-rí igikóokó!
you-pres-hit-asp child like that you-V-asp that he-be animan
'You are hitting the child like that as if he were an animal'!

níi (m) bá 'if'
níimbá /ni-n-ba/ 'if I am'
if-I-be
_N-zaa-gu-h-a amafaraanga meénshi níi-m-b-á umukiré.
I-fut-you-give-asp money many if-I-be-ap rich
'I will give you a lot of money if I become rich'.
Sii-n-zí níi-m-b-á bá-az-a ku-uz-a.
Nit-I-know if-I-be-asp they-come-asp to-come-asp
'I don't know if they are coming'.

máze 'then'
máze /n-mar-ye/ 'if I finish'
I-finish-asp
_U-saanz-e maz-é ibiryó.
you-find-asp I-finish-asp food
'You arrive at the moment when I just finished the food'.
_Mu-baanz-é mw-iig-é maz-é mu-ryaam-é.
you-do first-imp. you-study-imp. I-finish-asp you-go to bed-imp.
'Study first and then go to bed'.

kabone 'even if'
kabone /ka-bon-e/ 'may you see'
hortative-see-subj
_Ka-bon-e ibyáago. 'May you be in trouble'
hort-see-imp misfortune
_U-túm-i-ye uríiya mugoré ntí-y-aaz-a ka-bon-é n'íiyó w-aa-mw-eér-er-a amafaraanga.
you-send-appl-asp that woman neg-she-come-asp hort-see-imp. and if you-cond-her-accept- appl-asp money
If you invited that woman, she would not come even if you promised her money'.
agatsíinda 'in fact'
agatsíinda /a-ka-tsíind-a/ 'and then s/he wins'
s/he-tense-win-asp
_N-zi kó umugabo y-a-búuran-ye, a-ga-tsíind-a
I-know that man he-past-plead-asp he-cons-win-asp
'I know that the man pleaded and won'
_Umwáana y-ará-tsiinz-w-e, a-ga-tsíind-a nti-y-iig-á
child he-past-win-pass-asp he-cons-win-asp not-he-study-asp
'The child failed school, as a matter of fact he doesn't study'
byóongeye 'if they add' 'moreover'
_U-ra-bón-a ibyo by-óonge-ye ikí?
you-see-asp that it-add-asp what
'In your opinion, what is the importance of that?'
_Y-ará-mu-tút-se by-óonge-ye kaándi a-ra-mú-kubit-a.
S/he-past-him/her- hit it-add-asp also s/he-pres-him/her-hit-asp
S/he hit him/her, moreover s/he insulted him/eher.
It is also interesting to note that in the case of the word 'maybe', the three subject singular pronouns n- 'I', u- 'you' and a- 's/he' are found with the verb stem -end- 'take'.

7. néenda 'maybe'
néenda /n-end-a/ 'if I take'

wéenda 'maybe'
wéenda /u-end-a/ 'if you take'
you-take-asp
yéenda 'maybe'
yéenda /a-end-a/ 'if he takes'

This linguistic variation is natural. It will either result in doublets or triplets or the selection of only one single form and thus the elimination of others.
Like function words from infinitives, finite verb-derived conjunctions are also formally speaking ambiguous as to their nature. Are they indeed separate lexical items or finite verbs with a different function. All these questions will be answered in section 5.

1.3. Conjunctions in Irregular Verb Forms:

This category consists of conjunctions which manifest regular verb stems and aspect markers but with irregular prefixes as shown in (8).
8. cyákorá 'but' /cya-kor-a/
pref-work-asp
gukora 'to work'
cyáangwá 'or' /cya-ang-w-a/
pref-hate-pass-asp
kwáangwa 'to be hated'
keretse 'except' /ke-rek-ye/
pref-let-asp
kureka 'to leave aside'
keretse kó + finite verb 'except that'
/ke-rek-ye/ from kureka 'to leave aside'
pref-let-asp
níiburá 'at least'
/ni-bur-a/ from kubúra 'to miss/lack'
pref-lack-asp
byíiburá 'at least'
/byi-bur-a/ from kubúra 'to miss/lack'
pref-lack-asp
yíibá 'if'
/yi-ba/ from kubá 'to be'
pref-be
nyámará 'however'
/nya-mar-a/ from kumara 'to finish'
pref-finish-asp
aríko 'but'
?/ari ko/? ari ko 'it is that'
aháarí 'maybe'
?/aho ari/? aho ari 'where he is'
booshye 'as if' ---also ubooshye
/u-ba-oh-ye/
you-them-mislead-asp

It is also very important to note that the irregular verb derived conjunctions appear in a variety of phonetic forms: the stem remains the same but the prefix which corresponds to the subject marker can change either to first person, second person or third person , the suffix which shows aspect can appear with either the imperfective aspect marker -a or the perfective aspect -ye, high tone or low tone or short or long vowel in the prefix position as seen in the examples below without changing the meaning or the function.

9.-búr- 'miss/fail'--->'at least': byíiburá, byáaburá, byíibuzé, cyáaburá, cyáabuzé, cyíiburá, cyíibuzé, náabuzé, níiburá, níibuzé, yáaburá, yáabuzé, yíibuzé.
-áang 'hate/refuse'--->'or'-: cyáangwá, byáangwá, cyáangá, cyáangwé, náangwá, yáangwá.
-kór 'do/work'--->'but': cyákorá, cyókorá, cyókozé, icyákorá, icyóbikorá, icyóbikozé, icyókorá, icyókozé
-rek- 'let/leave'--->'except': kereka, keretse
-bá- 'be/live'---> 'if at least': níibá, yíibá, yáabá, naamba, naaba, niiba
-mar- 'finish/be useful'--->'but': nyámará, yámará
-aang-'refuse/hate': naangwa, 'at least'
-rek- 'leave/let'---> 'also': ndetse, ndetsé, naandetsé
-aang-'hate/refuse'--->'moreover': nkaanswe, nankaanswe, nankaanswé
-eend-'take'--->'perhaps/maybe': yéenda, byéenda, wéenda

This linguistic phonetic variation is the one which is responsible for deiconization or grammaticization. As we will see with other examples, words which create function words are first of all polysemous, that is single lexical items with multiple meanings or functions or hybrids but later on, either desemanticization, the loss of meaning, or phonetic change take place, resulting in deiconization.

It is important to note that the conjunction ngo 'that' which is used in "subjective" reports behaves also like a true verb in some cases as seen in (10):
10. ngw'iinó /ngo inó/ 'come here'
ngoo tugeendé /ngoo tu-geend-é/ 'let's go'
ng'iikí igitabo /ngo iki igitabo/ 'Here is the book.'
Cognates of this ngo are found in other Bantu languages such as Swahili in which it is realized as njo 'come'.
The verb -ti 'to say' which in other Bantu languages is used as a complementizer is restricted in Kinyarwanda to the reporting of direct speech. It is never found in the infinitive form. It takes the prefix of the nominal subject.
11.
Ba-ra-vúz-e ba-ti: "Nti-tu-záa-garuk-a". They just said: "We will not come back'
they-pres-say-asp they-say: not-we-fut-return-asp
Mw-aára-vúz-e mu-ti: "Abáana ba-ra-rwáa-ye". You said: "The children are sick"
you-past-say-asp you-say: children they-pres-be sick-asp
Another function word which might have a verbal origin is the conjunction na 'and'.It is used as a free morpheme to coordinate nouns and verbs, or as a verbal prefix to coordinate verb phrases as seen in the example which follow.
12.
Ha-ri hotéeri na baáre 'there is a restaurant and a bar'
there-be hotel and bar
ba-ra-andik-a ba-ka-na-som-a 'they write and read'
they-pres-write-asp they-cons-and-read-asp
tu-ra-ríriimb-a tu-ka-ná-byin-a 'we sing and dance'
we-pres-sing-asp we-cons-and-dance-asp
This same na is probably the same as the reciprocal suffix
Yohaáni na Máriyá ba-ra-kúund-an-a 'John and Mary love/like each other'
John and Mary they-pres-like/love-rec-asp
In Swahili again, this na has a double function. It acts as a preposition with multiple functions or as a verb which denotes possession 'have'.
Although the conjunctions that have been presented so far look like other verbs, they are indeed conjunctions because of their frozen status: they cannot change pronouns, tenses and moods. They cannot incorporate object pronouns like regular verbs, either.
All conjunctions come from verbs, the only exception is the kaándi 'also' which comes from the indefinite adjective -ndi 'other': akaándi kaantu/akaantu káandi 'another thing'; abaándi baantu/abaantu báandi 'other people'; ubuúndi 'otherwise', etc.

II. Function words from other sources

Although it is possible to find other function words without a verbal origin, it is interesting to note that they come from other categories as well such as nouns, demonstratives, pronouns and adjectives. The latter three also happen to be derived from primitive categories in Kinyarwanda (Kimenyi, 1989 and 1995)

2.1.Prepositions from Nouns

The nouns isi 'earth', ijuru 'sky', imbere 'front' and inyuma 'back' created prepositions and spatial and temporal adverbs.
13. ijuru 'sky'--->heejuru 'on the top/above'
haruguru/ruguru/ejuru 'up'
igitabo kiri heejuru y'áaméezá 'the book is on the top of the table'
ibicíiro byaagiiye heejuru 'the prices went up'
aravúgira heejuru 's/he is talking outloud'
igira haruguru 'move up'
ari haruguru y'úrugó 's/he is at the upper side of the compound-fence'
14. isí 'earth/ground'--->haasí 'on the floor/below/down'
--->muunsí 'under'
icara haasí 'sit down'
ibicíiro byaagiiye haasí 'the prices went down'
ibitabo biri muunsí y'áaméezá 'the books are under the table'
The words imbere 'ahead/in front' and inyuma 'behind', have lost their original meaning 'front' and 'back', respectively and are now used as temporal or spatial adverbs only.
15. imbere --->imbere 'in front of/inside'
mbere 'before'
haambere 'recently'
a-ri imbere y'ínzu 's/he is in front of the house'
s/he-be front of house
a-ri muu nzu imbere 's/he is inside the house'
s/he-be in house front
a-ri imbere 's/he is ahead'
s/he-be front
u-záa-hamágar-e mbere yó kuuza 'call before coming'
you-fut-call-subj before of to-come
a-heruuts-e guhámagara haambere 's/he called recently'
s/he-do last-asp to-call before
16. inyuma--->inyuma 'behind/outside'
nyuma 'later'
haanyuma 'afterwards'
a-ri inyuma y'ínzu 's/he is behind/outside the house'
s/he-be behind of house
a-ri inyuma 's/he is behind'
s/he-be behind
ni wé w-aa-j-é nyuma 'it is him/her who came last'
be him/her she-past-come-asp behind
mu-kor-é haanyuma mu-taah-é 'work and then/afterwards go home'
you-work-subj behind you-go home-subj
Note again from the examples above that doublets have been created to differentiate spatial and temporal meanings. The temporal expressions have lost the initial prefix mbere/nyuma or added the preprefix ha- as in haanyuma and haambere.
These same words are used in other Bantu languages to derive spatial prepositions and temporal adverbs as well.
17.chini ya mti 'under the tree' nchi 'ground/earth'
juu ya mlima 'on the hill' juu 'sky'
ndani ya sanduku 'inside the box' nda 'stomach'
mbele ya watu 'before the people' mbele 'front'
nyuma ya nyumba 'behind the house' nyuma 'back'
miongo mwa watu 'among the people' mwongo 'a group of ten things'
mpaka 'until' mpaka 'boundary'
Luganda:

under: wansi 'earth'
in front of: mu maso ga 'in the eyes of'
behind: emabega wa 'the shoulders of'
up/above: waggulu
To differentiate spatial expressions from temporal expressions, many languages have tendency to create doublets such as French devant/avant 'in front'/before'; d'où/donc 'from where'/'since'; derrière/arrière 'behind'/'in the back'; here/hence; this/thus; there/then; where/when; farther/further, etc.
It is usually body parts and spatial expressions which in many languages give rise to prepositions: Swahili ndani 'inside' (in the stomach) or English head, breast, back.. .which gave...ahead/abreast/aback.
These prepositions as the expressions muu nsi 'under' and heejuru show are exactly like prepositional phrases in the synchronic grammar of Kinyarwanda because they consist of a preposition and a noun. To create a separate category, these expressions have undergone phonetic changes. The synchronic preposition phrases of the head nouns are mw'iisí 'in the earth' and kw'iijuru 'on the sky'

2.2.The Demonstrative Origin of aho and ubwo

Function words derived from demonstratives behave exactly like other textual deictic expressions found in other languages. It is always the distal demonstrative as opposed to the proximal demonstrative which gives rise to textual demonstratives.
Kinyarwanda has four distal demonstratives, namely -o, -riíya, -´rya, and -áa which always follow the "subject" class marker of the head noun.
-´rya and -áa are textual demonstratives and are used only to refer to objects or events in the remote past.
18a. abo bagabo 'those men'
ibyo bitabo 'those books'
b. báriíya bagabo 'those men'
bíriíya bitabo 'those books'
c. bárya bagabo 'those men'
bírya bitabo 'those books'
d. báa bagabo 'those men'
byáa bitabo 'those books'
The spatial distal demonstrative, however, also has the textual function referring to objects or events previously mentioned in the discourse or in the far past .
19. Abo bagabo n-a-ku-bwií-ye tw-aa-gaaniiri-ye. 'I conversed with those men I told you about'.
those men I-past-you-tell-asp we-past-talk-asp
Ibyo bitabo n-a-taa-yé n-a-rí n-á-bi-guz-e amafaraanga meénshi.
those books I-past-lose-asp I-past-them-buy-asp money many
'Those books that I lost I have bought them for a lot of money'.
20. ubwo bwáato 'that canoe' (near the hearer) or (in the story)
aho haantu 'that place' (near the speaker) or (in the story)

This demonstrative when it occurs by itself just before a verb, has the meaning of 'therefore'/'then' with class 14, the prefix -bu- , and expresses uncertainty or doubt with the prefix -ha-. The latter is always found in interrogative sentences only.
21. ubwo a-gíi-ye, nti-tw-oónge-ye kubónana 'then when s/he left,we never saw e again'
then s/he-go-asp not-we-do again-asp to-see-eacher-other'
ubwo tu-zaa-bon-an-a 'we will see each other then'
then we-fut-see-rec-asp
22. aho a-zaa-garuk-a? 'is s/he coming back?'
there s/he-fut-be back-asp
aho bi-ra-shobok-a? 'is it going to be possible?'
there it-pres-be possible-asp

2.3.The Origin of búrya/hárya/kúrya

The demonstrative -´rya refers to things in the past or textual remote demonstratives.
23. búrya bwáato 'that boat'
kúrya gukóra kwé 'that way of hers/his of working'
hárya haantu 'that place'
Without a head noun, with prefixes bu-, ku- or ku-, it becomes a function word as shown in (24) and (25). It is ambiguous, however, when prefixed with bu- because it can have a pronominal temporal reference.
24. Búrya ntí-tw-aá-bon-an-nye. 'We didn't see each other that time'
dem. not-we-past-see-rec-asp
Si-n-a-rií n-zi kó búrya w-a-rí u-rwáa-ye. 'I didn't know that you were
sick then'.
not-I-past-be I-know that dem. you-pst-be you-be sick-asp
The demonstrative -'rya with the prefix bu- is used to correct a misguided belief or to show surprise or disbelief as examples in (25)a indicate.
25a. Eesé, búrya u-nyw-a itaábi? 'Eh! you smoke!'
Q dem you-drink-asp tobacco
Búrya u-ta-garúts-e, n-aa-gir-a guté? 'In case you don't come back,
what shall I do?'
dem. you-not-come back-asp I-cond-do-asp how
b. Kúrya u-ta-garúts-e? 'And in case you don't come back?'
dem. you-neg-come back-asp
Kúrya u-fit-é amafaraanga, u-ra-yá-tagaguz-a. 'Because you have
money, you waste it'.
dem. you-have-asp money you-pres-waste-asp
c. Hárya u-zaa-garuk-a! 'Eh! is it certain that you'll come back!'
dem you-fut-come back-asp
Hárya w-iit-w-a ndé? 'What is your name again?'
dem you-call-pass-asp who
The demonstrative -riíya 'that' is used to refer to objects which are far away from both the speaker and the hearer (26), but when it occurs with the prefix bu- in front of a verb which has a different subject marker, it has an adverbial meaning as shown in (27).
26. Báriíya báana ba-ra-rwáa-ye. 'Those children over there are sick'
dem. children they-pres-be sick-asp
Bíriíya bitabo ni bibí. 'Those books over there are bad'
dem. books are bad
27. Búriíya nt-a-záa-kw-iishyur-a. 'Actually, s/he will never pay you back'
dem. neg-s/he-fut-you-pay back-asp
Uríiya mugabo gu-kór-ta kúriíya, búriíya a-fit-e i-ki-bí-mu-teer-a
dem. man to-act-asp dem. dem. he-have-asp it-it-him-cause-asp
'For that person to act like that, he has a reason to do that'

2.4.From Demonstratives to Pronouns: The Pronominal Origin of kó, yúukó, kukó, iyó and
ubwó
Pronouns in Kinyarwanda have the VCV shape. That is, they take both the preprefix and the prefix of the nouns they stand for. The last vowel, however, realized as -ó with a high tone is the stem . Examples are given in (28).
28a. Siinzí icyó ashaaká 'I don't know what s/he wants'
b. Siinzí ikiintu ashaaká 'I don't know the thing that s/he wants'
c. Siinzí ahó agiiyé 'I don't know where s/he is going'
d. Siinzí iyó agiiyé 'I don't know where s/he is going'
e. Siinzí ahaantu agiiyé 'I don't know the place s/he is going'
f. Siinzí ahó azáazira 'I don't know when s/he will be back'
g. Siinzí igihe azáazira 'I don't know the time s/he will be back'
h. Siinzí ubwó azáazira 'I don't know when s/he will be back'

The subordinators iyó 'if' and ubwó 'since' are exactly like their pronominal counterparts seen in (28)d and h from which they derive.
iyó without a head noun is thus either a pronoun referring to a noun belonging to class 9 or to both the locative expression 'where' or the temporal expression 'when' as seen in (29 )a and (29)b.
(29)a. iyó a-geend-á a-ba y-ará-ha-tú-bwii-ye 'S/he would have told us where
s/he goes'
where s/he-go-asp s/he-be s/he-past-it-us-tell-asp
b. ubuúndi iyó a-gii-yé a-ra-tú-bwiir-a. 'Usually, when s/he goes, s/he
tells us'
otherwise when s/he-go-asp s/he-pres-us-tell-asp
If the verb is in the conditional tense, however, as shown in (30), it has the meaning of 'if'.
(30). iyó a-geend-á a-ba y-ará-tu-bwíi-ye 'If s/he had left s/he would have
told us'
if s/he-go-asp s/he-be s/he-past-us-tell-asp
ubwó in the preverbal position has only the temporal meaning (31) or the adverbial meaning as seen in ( 32).
(31). Ubwó a-herúuk-a guhámagara nti-tw-aá-ri dú-ha-ri.
when s/he-aux-asp to-call not-we-past-be we-there-be
(32). Ubwó a-geend-á, tu-ra-sígar-a 'Since s/he will be leaving, we are
staying'
since s/he-go-asp we-pres-stay-asp
Ubwó akorá, atuguríze amafaraanga
'since s/he is working, s/he should loan us some money'
(33)a. Ni uuko uyu mugabo a-shaak-á kó u-bi-kor-á.
be that this man he-want-asp that you-it-do-asp
That is how this man wants you to do it'.
b. Ni uuko uyu mugabo tú-maz-e kuvugana a-ra-geend-a.
then this man we-finish-asp to-talk-each-other he-pres-go-asp
'Then this man after we finished talking to each other, he left'.
34a. Ni aaho abo baantu ba-tuu-yé.
be there these people they-live-asp
'It is there that these people live'.
b. Ni aaho reeró abo baantu bá-rek-e ba-geend-é.
then then those people them-let-subj they-go-subj
'Then let those people go'.
35a. Ng'uukó ukó by-aá-geenz-e.
presentative how how it-past-go-asp
'That's how it went'.
35b. Ng'uukó
36a. Ng'aahó ahó umwáana y-a-shyiz-é ibitabo.
presentative there there child he-past-put-asp books
'It is there that the child put the books'.
b. Ng'aahó ní-muu-z-é tu-geend-é.
then cl-you-come-subj we-go-subj
'Then let's go'!

Kinyarwanda has three complementizers namely kó, yúukó and kóngo. The last two
yúukó and kóngo are derivatives of kó: yá ukó and kó ngo, respectively. The complementizers kó and yúukó are used to report direct evidence. Whereas the embedded proposition after kóngo lacks this evidence (Givón and Kimenyi, 1974).
33. umugabo yavuze kó umugoré arwaayé. 'The man said that the woman is
sick'.
umugabo yavuze yúukó umugoré arwaayé
umugabo yavuze kó ngo umugoré arwaayé.

This kó comes from the pronoun ukó. Evidence for this assertion comes from its morphology and phonetic similarity with this pronoun. First, all pronouns end with the vowel -ó with a high tone which is actually the stem and they all take also the "subject" prefix of the head noun like demonstratives and subject pronouns. Comparative evidence shows that this pronoun comes from the demonstrative because in other Bantu languages, distal demonstratives and pronouns look the same. This high tone in Kinyarwanda might have occurred later to differentiate between the two since indeed many cases of doublets are found in this language. As (34) shows this high tone on pronouns disappears also in some environments such as after the negative marker nta. It may be related to the focus phenomenon or the clause final or non-final position. Indeed Kinyarwanda has a group of words and verbal tenses which assign a high tone depending on the position they occur in the sentence structure.
34a. Nta ko a-fit-é 'S/he doesn't have it'. (ukuguru: leg)
not it s/he-have-asp
Ntaa kó. 'There is none'/'No way'
neg it
b. Nta bwo tu-bón-an-a. 'We never see each other'.
neg it we-see-rec-asp
Ntaa bwó. 'Never'
neg it
There are times also like in the case of these complementizers when the pronoun prefix is deleted as examples in (35) indicate.
35. ni byó twaavugánye 'it is it that we talked about'
si byó twaaboonyé 'it is not that we saw'
nta ko ntaagizé 'there is anything I didn't do'
nta ho ntaagiiyé 'there is anyplace I didn't go'
ku byó twaavuzé 'on what we said'
mu byó twaareebyé 'in what we looked'
The preprefix is always deleted after the copular ni/si 'it is/it is not', the exclusive ntaa 'there is not' and the prepositions ku(rí)/mu(rí).

36.uko ni kó kubóko a-rwaa-yé 'that one is the one arm that he is suffering
from'
that is it arm s/he-be sick-asp
ni kó bi-mez-é 'that is how it is'
be it it-be-asp
All these examples clearly show that the complementizer kó with its derivatives yúukó and kóngo is the same as the ukó which is realized as ukó or kó in some environments.
This pronoun kó stands for the expression 'how':
37a. sii-n-zí ukó a-mez-e. 'I don't know how s/he is doing'.
not-I-know how s/he-be-asp
b. ni kó kugeenda 'that is the reason of leaving'
be it to-go
ni kó a-záa-bi-geenz-a 'that is how s/he will do it' kó kó 'really'
be how s/he-fut-it-do-asp
In the sentence initial position kó acts as a question marker:
38a. kó abagabo ba-a-gii-yé ba-tá-tu-bwíi-ye? 'How come the men left
without telling us'?
Q men they-past-go-asp they-not-us-tell-asp
b. kó mw-aa-byuuts-é kare? 'How come you got up early'?
Q you-past-get up-asp early
It is the same kó from which the conjunction kukó which is the same as kubéera ko 'because' and kókó 'really/indeed/right' are derived.
39a. nt-aa-kor-á kukó a-rwaa-yé
neg-s/he-work-asp because s/he-be sick-asp
b. a-ra-naniw-e kubéera kó y-a-koz-é cyaane 'S/he is tired because s/he
worked a lot'
s/he-pres-be tired-asp because s/he-past-work-asp much
c. kókó ibyo ni byó 'Indeed that is true'
really that is it
y-ará-mu-túts-e kókó. 'S/he insulted him/her indeed'.
s/he-past-him/her-insult-asp really

It is not only pronominals and demonstratives which gave rise to textual deixis but nominal temporal deixis as well as examples in (39) indicate.
40. reeró 'then' <---*leo 'today' (Swahili)
nóone 'today'--->nóone 'then' --->nóone reeró 'then/therefore'-->na
nóone 'again'
ejó 'yesterday/tomorrow'--->ejó+verb 'by any chance/accidently'
ható 'soon' --->ható+verb 'consequently'

III. The Verbal and Nominal Origin of Clitics and Affixes

Givón, Heine, etc. have shown that in many languages, clitics and suffixes come from nouns and verbs. The most known example is that of the future tense marker in French where the tense is marked by the infinitive marker plus the conjugation form of the auxiliary verb in the present tense.

41. chanter+avoir
chanter+ai ai 'I have'
chanter+as as 'You have'
chanter+a a ' has'
chanter+ons (av)ons 'we have'
chanter+ez (av)ez 'you have'
chanter+ont ont 'they have'

3.1.The Verbal Origin of the Proclitics ni- and nti-

In Kinyarwanda also, the clitic ní found in front of some temporal expressions, and some verb forms such as the conditional, the emphatic imperative, comes from the copular ni 'to be'
42. nínjoro 'in the evening' ---- ijoro 'night'
nímugórooba 'in the evening'----umugórooba 'evening'
nímuúnsi 'in the evening' ---- umuúnsi 'day'
This temporal prefix clitic is the same as the locative suffix in Swahili: mezani 'on the table'; sokoni 'in the market'; nyumbani 'at home'; etc.
43. níbakorá 'if they work' --- bakorá 'that they work'
nímuvugá 'if you talk' ----muvugá 'that you say'
nítugeendá 'if we go' --- tugeendá 'that we go'
nibakoré 'they should work' --- bakoré 'they should work'
nimuvugé 'talk' ---- muvugé 'talk'
nitugeendé 'let's go'---tugeendé 'let's go'

The preprefix negative marker nti- probably has a verbal origin as well because for the first person singular, this nti- is substituted by si- which is similar to the negative copular si 'it is not'.
44. a. ntibákore 'they should not work' --- bakoré 'they should work'
ntimúvuge 'you should not talk' --- muvugé 'you should talk'
ntitúgeende 'we should not go' --- tugeendé 'we should go'
siínkore 'I should not work' -- nkoré 'I should work
siíngeende 'I should not go' --- ngeendé 'I should go'
b. ntibakorá 'they don't work' -- bakora 'they work'
ntimuvugá 'you don't talk' -- muvuga 'you talk'
ntitugeendá 'we don't go' -- tugeenda 'we go'
siinkorá 'I don't work' -- nkora 'I work'
siingeendá 'I don't go' -- ngeenda 'I go'

3.2.Locative post clitics: From promotion to incorporation and lexicalization.

Locative clitics -hó, -mó, and -yó which have been lexicalized in many cases, becoming part and parcel of the verb stem, come from the locative prepositions ku(rí), mu(rí) and
i, respectively. This started first as a result of promotion of locative noun phrases to the direct object status, second as pronoun incorporation and finally, lexicalization. Promotion to object position is shown first. When this happens, the locative becomes either a clitic to the verb or to the promoted locative phrase as shown in the following examples.
45.gutáanga urugero ku mukoóbwa 'to take the girl as an example'
gutáangahó umukoóbwa urugero
gutáanga umukoóbwa hó urugero
gukúura kabiri murí kané 'to substract two from four'
gukúura kané mó kabiri
gukúuramó kané kabiri
kuvaana umuuntu ku butégetsi 'to remove somebody from power'
kuvaana ubutégetsi hó umuuntu
kuvaanahó ubutégetsi umuuntu
kuvaana ubutégetsi ku muuntu 'to remove power from somebody'
kuvaana umuuntu hó ubutégetsi
kuvaanahó umuuntu ubutégetsi

In Kirundi and some dialects of Kinyarwanda, the locative clitic derived from ku is realized as -kó. In Standard Kinyarwanda, it is found only in a handful of lexical items.
46kugeendera kó 'to leave for good'
Y-a-gíi-ye ku-ryáam-a a-geend-er-a kó. 'When s/he went to bed, s/he went for good'.
s/he-past-go-asp to-sleep-asp s/he-go-appl-asp kó
guhéera kó 'to start right away'
Ni u-gér-a yó u-heer-e kó u-m-pámagar-a. 'When you get there, give me a call '
when you-arrive-asp there you-start-subj. kó you-me-call-asp immediately'.

Locative cliticization on the verb creates ambiguity, however, because not only do -hó/mó and -yó signal locative promotion to object status but pronoun incorporation as well as the following examples show.
47.Abagabo b-iica-ye kuu ntébe. 'The men are sitting on the chair'.
men they-sit-asp on chair
Abagabo ba-y-iica-ye hó. 'The men are sitting on it'
men they-it-sit-asp on
Abagabo b-iica-ye hó. 'The men are sitting on it'
Abáana ba-ryaam-ye mu gitaánda. 'The children are sleeping in the bed'
children they-sleep-asp in bed
Abáana ba-ki-ryaam-ye mó. 'The children are sleeping in it'
children they-it-sleep-asp in
Abáana ba-ryaam-ye mó. 'The children are sleeping in it'
children they-sleep-asp in
Abagoré ba-gii-ye kw'iisokó. 'The women are going to the market'
women they-go-asp to market
Abagoré ba-gii-ye yó. 'The women are going there'
women they-go-asp there

Some of these verbs which have been lexicalized with the locative clitic are the following:
48.kwíirirwahó 'to spend the day okay' as in the greetings wiiriweho?
'good afternoon'.
kuramukahó 'the spend the night okay' as in the greetings waramutsehó?
'good morning'
kubáhó 'to live'
kurushahó 'to increase'
kwiigirayó 'to push' gukúrikirahó 'to follow'
kubónerahó 'to seize the opportunity'
kuvúgirahó 'to tease'
gushírahó 'to finish'
kumarahó 'to finish' trans.
gukómahó 'to put in trouble'
kwíitahó 'to care for'
gukúuramó 'to substract'
guhítamó 'to choose'
kuvaanahó 'to remove'
gushyírahó 'to name'
gusígahó 'to stop doing something'
kugumahó 'to stay in a position'
gusúbiramó 'to repeat'
gusúbirahó 'to go back'
kurwaanahó 'to help'
kuramukahó 'to survive the night'
kwíirirwahó 'to survive the day'
kurekerahó 'to let stay in a position'

3.3.The origin of the postsuffix -ga

The suffix -ga the imperfective aspect marker which in Kinyarwanda shows repetivitive or progressive events as seen in (49), comes from the word erega as examples in (50) indicate.
49. Abagoré ba-a-kór-aga mu gitóondo. 'The women were working in the morning. Abagoré ba-a-kóz-e mu gitoóndo. 'The women worked in the morning'
Abáana ba-ára-sóm-aga ba-fit-e imyáaka iné. 'The children used to read when they were four'.
children they-past-read-asp
Abáana ba-ára-som-ye bá-fit-e imyáaka iné. 'The children read when they were four'.
childrden they-past-read-asp
erega aba bagoré ba-kuund-a abo báana! 'Actually, these women like those children'.
actually these women they-like-asp those children
aba bagoré ga ba-kuund-a abo báana!
these women actually they-like-asp those children
aba bagoré ba-kuund-a ga abo báana!
these women they-like-asp actually those children
aba bagoré ba-kuund-a abo báana ga!
these women they-like-asp those children actually
*ga aba bagoré bakuunda abo báana
aba bagoré baákuundaga abo báana!
aba bagoré baákuunda abo báana! (Kirundi)

The cliticization of ga in Kinyarwanda might be a new development because Kirundi, a sister dialect of Kinyarwanda doesn't have it.
-ga behaves phonologically speaking as not part of the verb form also because the vowel of the part of the verb which precedes it doesn't lengthen if it is preceded by a the glide -w-.
50a.b-aára-kúbit-w-aga/*b-aára-kúbit-w-aaga.
they-pst-beat-pass-aga
They used to be beaten'
b. Abakoóbwa b-aá-suur-w-aga/*b-aá-suur-w-aaga n'ábahúungu.
girls they-pst-visit-pass-aga by boys
'The girls used to be visited by the boys'.
This clitic developed probably like the question marker mbeesé/eesé/sé whose linguistic variation is dictated by its position in the sentence structure.

51.mbeesé aba bagoré ba-kuund-a abo báana? 'Do these women like those children'?
Q these women they-like-asp those children
eesé aba bagoré ba-kuund-a abo báana?
Q these women they-like-asp those children
aba bagoré sé ba-kuund-a abo báana?
these women Q they-like-asp those children
aba bagoré ba-kuund-a sé abo báana?
these women they-like-asp Q those children
aba bagoré ba-kuund-a abo báana sé?
these women they-like-asp those children Q

IV. The Origin of Affixes

In the next section, it will be shown that nominal prefixes and suffixes come from other nouns and that verbal suffixes and prefixes (tense-aspect-modality markers) come from full verbs also.

4.1.Nominal Affixes

4.1.1.The kinship term origin of onomastic prefixes

Kinyarwanda has prefixes which are added to any kind of word: noun, verb, adjective, adverb, etc. to create names. These morphemes are nya-, rwaa-, nyiri-, sée-, sa- for masculine names and nyirá-, mukáa-, and káa-, for female names (Kimenyi, 1979).
Among these we find nyirá- and nyirí- as well as sée- with its phonetic variations shée- and sa- which are the same as nyina and sé respectively and mean 'mother' and 'father'.
There are also regular words which have this onomastic prefix. Most of these refer , however, to types of animals, birds, insects, plants or crops.
52. nyirá-barazáana
nyira-muhari
nyirá-mivuúmbi 'fosse temporale'
nyirá-maniíngiri 'kind of bug'
nyira-mugeyo 'kind of rat'
nyirá-musaámbi 'crested crane'
nyirá-huúku 'domestic cat'
nyirá-bahánya 'bad luck'
nyir-ícyúubahiro 'the honorable'
nyir-ígira 'the causer'
nyir-íngoma 'his majesty'
nyir-úbutuúngaane 'his holiness'
nyir-ízina 'the person in question'
nyir-úbwiíte 'the concerned individual'
nyir-úbutaka 'the landlord'
sa-kábaka 'vulture'
sa-músure 'type of bird'
sa-rúhara 'bald eagle'
isá-bukurú 'anniversary'
sa-búduúgu 'diluted beer'
sa-búkené 'poverty'
sa-cyaánwa 'beardy'
sa-byíinyo 'name of a Volcano'
sa-byoombí 'not to have a place to be, to be in the limbo'
sa-gíhaánga 'bid head'
sa-gíhoobé 'who plays a musical instrument'
igishée-gabo 'virago'
sée-naaká 'so and so'
sée-muteneengeri 'kind of bird'
sée-ntuuzá 'so and so'
sée-rwaakira 'whirlind'
sée-rumoóndo 'kind of monkey'
sée-rutoroogo 'kind of pidgeon'
see-bukwé 'his/her father-in-law'
shée-buja 'his/her boss'
sogó-kuru 'my grandfather'
see-kúru 'his/her grandfather'
nyirá-bukwé 'his/her mother-in-law'
nyokó-bukwé 'your mother-in-law'
maa-bukwé 'my mother-in-law'
nyirá-kuru 'his/her grandmother'
nyogó-kuru 'my grandmother'
nyogó-seenge 'your father's sister'
nyirá-seenge 'his/her father's sister'
maa-séenge 'my father's sister'
nyokó-rume 'your mother's brother'
nyirá-rume 'his/her mother's brother'
maa-rúme 'my mother's brother'
nyira-buja 'his/her female boss'
nyoko-buja 'your female boss'
maa-buja 'my female boss'

4.2.The origin of nominal suffixes from natural gender

The feminine suffix -kazi which is added to all nouns to make them feminine is the proto-Bantu mukazi 'woman/wife' which is found in neighboring Bantu languages such as Runyankore, Mashi, Kilega meaning wife. In Kinyarwanda, it is no longer used and has been replaced by umugoré.
53. umunyarwaanda 'Rwandan' umunyarwaandakazi 'female Rwandan'
umuzuúngu 'white' umuzuúngukazi 'female white'
umwaámi 'king' umwaámikazi 'queen'
umwaárimu 'teacher' umwaárimukazi 'female teacher'
inkokó 'chicken' inkokókazi 'female chicken'
The proto-Bantu murume 'man' is no longer used in Kinyarwanda either. It has been replaced by umugabo. It is used, however, in a handful of lexical items as masculine suffix as seen in the following examples.
54. nyokórume 'mother's brother'
isekúrume 'he-goat'
impweerume 'male dog'
The reason there are not that many words that use it as a suffix is because of the unmarkedness of the masculine gender. In many languages, the feminine is the marked form and the masculine the unmarked.
The stem of the noun umwáana 'child' is also found in a few words as diminutive suffix.
55. ikibwá 'big dog' ---->ikibwáana 'puppy'
umukazáana 'son's wife' *umukazi 'wife'
Kinyarwanda uses usually, the prefix aka- for diminutive purposes.
In Southern Bantu languages such as Zulu and Xhosa, however, this suffix use for diminutive is very productive. Heine&Hünnemeyer (1991), have also shown that the word ví 'child' in Ewe a Kwa language spoken in West Africa, is used as a nominal suffix for express diminutives.

4. 2.Verbal Affixes

Verbal prefixes have the shape -VVC-, -VVNG- or -VC- whereas the verb stem is found as either -CVC-, -CVVC-, -CVVNC-, -VVC-, or -VVNC-. There is also evidence that historically, the initial syllable always started with a consonant but that some were dropped because of consonantal weakening. The ones whose initial consonant was lost happen also to be the oldest and the most used verbs. Since it is these old and most used ones which are also grammaticized, it is understandable why they too would lack a consonant in their initial syllable. Three suffixes -am-; -at-, and -aamb- will be given as examples.

The suffix -am- comes obviously from the verb kwaama 'to keep doing something'.
56. gusútama 'to squat'
kugarama 'to lie on one's back'
gucúrama 'to be upside down'
kwaarama 'to expect'
gupfúkama 'to kneel'
kwuunama 'to bend'
gusibama 'to be erased'
kurohama 'to crash'
kwéegama 'to lean'
kwáasama 'to have the mouth open'
guhunama 'to become stupid'
kugorama 'to bend'
guheengama 'to be obliquely bent'

The suffix -at- which is found in verbs which denote the concept of 'holding' probably come from the verb -fat- 'hold', whose cognates in other Bantu language such as Swahili are -pat-.
The p sound in Kinyarwanda became h or in a handful of lexical items f.
57. guhágatira
gupfúumbat(ir)a
gutéengat(ir)a

The suffix -aamb- might come from the archaic verb kugaamba 'to say'

58.gutakaamba 'to scream for help' <---- gutaka 'to scream'
gusakaambura 'to uncover the roof'<--- gusakaara 'to cover the house roof'
kurógoombwa 'to talk nonsense'
guháraamba 'to stratch'<----guhára
kujógoomba 'to walk with weight'<----kujóga
kugúriimba 'to speak slowly'
guhógoomba 'to speak loud and hard'
kugágaamba 'to swollen'<---- umugagá 'rich'
guhúruumba 'to look for food'<----- guhúra
gucokoombera 'to be exhausted'<----gucoka 'to get tired'
kwáagaamba 'to swollen'<---- kwáaga 'to widen'
kugáraamba 'to become tough'
This morpheme is probably from the Kinyarwanda archaic verb -gaamb- 'to say' which is found in folktales and proverbs. The stem is found in neighboring languages such as Luganda kugamba and similar cognages with the initial consonant lost such as the Swahili
kwamba. Derived words such as amagaambo/ijambo 'words/word' are found in Kinyarwanda.

Kinyarwanda has two types of tenses, simple tenses and complex tenses. Simple tenses are marked by either a vowel -V- or a syllable of the shape -CV- in the prestem position. Complex tenses have the following shape -CVCV-, CVCVV, -VVCV- or -CVV-.
60. -CVCV-
ba-raga-kor-a 'may they work'
ba-raga-kin-a 'may they play'
-CVCVV-
ba-racyáa-kór-a 'they are still working'
ba-racyáa-kin-a 'they are still playing'
-VVCV-
b-óogá-kin-a 'may they play'
b-óogá-kor-a 'may they work'
b-aára-kín-ye 'they played'
b-aára-kóz-e 'they worked'
-CVV-
ba-záa-kin-a 'who will play'
ba-záa-kór-a 'who will work'
nti-ba-ráa-kin-a 'they have not played yet'
nti-ba-ráa-kór-a 'they have not worked yet'
Evidence for the verbal origin of these prefixes comes from the fact that they indeed behave like other regular verbs. They behave completely differently from the regular monomoraic or monosyllabic tenses. The bisyllabic tenses behave as if they consisted of a tense marker and a verb stem. Indeed these tenses change form in the negative and subordinate clauses dropping the first syllable as in

61. ba-raga-kor-a 'may they work' nti-bá-ga-kor-e 'may they not work'
ba-raga-kin-a 'may they play' nti-bá-ga-kin-e 'may they not play'
ba-racyáa-kór-a 'they are still working' nti-ba-gi-kor-a 'they don't
work any more'
ba-ta-gi-kor-a 'who don't work any
more'
ba-racyáa-kin-a 'they are still playing' nti-ba-gi-kin-a 'they don't
work anymore'
ba-ta-gi-kin-a 'who don't play
anymore'
One of these tenses, namely the future tense marker -zaa- comes indeed from the verb stem -z- of kuuza 'to come'. Evidence that indeed the future tense morpheme and the verb kuuza 'to come' are the same is its Kiswahili cognate -ja 'come'. In Kiswahili also the same verb is used as a negative marker of some tenses as shown by the following examples.
62. Bado ha-ja-maliza kazi yake
not not-come-finish work of him/her
Bado watoto ha-wa-ja-rudi
not children not-they-come-return
Bado ha-m-ja-sema
not not-you-come-say
si-ja-soma 'I have not read yet'
not-come-read
ha-wa-ja-soma 'they have not read yet'
not-they-come-read
The future tense in Swahili is realized as either -ta- or -taka-. Obviously -ta- is a short form of -taka. The longer form ressembles exactly like the regular verb -taka 'want'
63. Watoto wanataka chakula 'the children want food'
Watoto watarudi nyumbani 'the children will return home'
Watato watakaporudi nyumbani nitakwenda nao 'When the
chidreln go home, I will go with them'
Watoto watakapotaka kurudi nyumbani nitakwenda nao 'When
the children want to return home, I will go with them'.

As the following examples show, the verbs which become affixes, first become tropes, third they are decategorized, being demoted from regular verbs to auxiliary verbs, fourth form auxiliary to conjunctions or subordinators, fifth they are reduced semantically and finally when they are reduced semantically, they can become affixes.
The verb kwooya 'to stop doing something' is used before other verbs to express prohibition as in (64)
64. (ni)-b-ooy-é kugeenda 'they should not go'
cl.-they-v-subj to-go
(ni)-mw-ooy-é gukóra 'don't work'
cl-you-v-subj to-work
(ni)-y-ooy-é kuvúga 's/he should not work'
cl-s/he-v-asp to-say
This verb can appear in short form also, yielding:
65. n-ooy-é 'I should not' --->noó/neé
w-ooy-é 'You should not' --->woó/weé
y-ooy-é 's/he should not'--->yoó/yeé
tw-ooy-é 'we should not' --->twoó/tweé
mw-ooy-é 'you should not'--->mwoó/mweé
b-ooy-é 'they should not'--->boó/beé

This verb thus behaves exactly like the English auxiliary and modal verbs which also appear in reduced forms : have/has--->'ve/'s; am/are/is --->'m/'re/'s; would ---->'d, etc. Once they have been reduced, they behave exactly phonetically speaking like other proclitics to the main verb. After they have become clitics, they may end up becoming affixes to the main verb(2)
Many other Bantu languages such as Shona have many tense markers which come from synchronic regular verbs.

V. Conjunctions in a Semiotic Theory:

Function words in all languages express grammatical concepts and have only an 'algebraic' function. They are created by linguistic contexts. Unlike, other words or signs, function words don't have any reference or meaning. Referential words refer to objects that exist in the real or the imagined world. Verbs are referential because they refer to states, actions or events. It is the reason why new words are not created for function linguistic items but instead are referred to by those which already exist. Words are not chosen arbitrarily, however. A semantic relation can be seen between certain conjunctions and verbs from which they are derived. Urétse and keretse 'except' are indeed close to the meaning of urétse 'if you exclude, leave aside, except'; kugira ngo 'so that, for' is close to the meaning of the verb kugira ngo 'to cause to do that'; the meaning of byíiburá/níiburá 'at least' is close to that of níiburá 'if I miss/lack'; nuuko 'then, therefore' is close to the meaning of ni uko 'it is that way/how'... In many cases, however, there is opacity, as far as their respective meanings are concerned. A semiotic theory forces the analyst to find the lost transparency, however.
The existence of conjunctions and their relation to verbs show that linguistic signs are not arbitrary. They started as icons and indices but later became symbols first on the content level: signata (or signatum in the singular form) because the semantic relation with their verbs became obscure. On the form level, signantia (singular form signans), some conjunctions have also started undergoing changes as seen in 1.3. Those conjunctions where phonetic change is noted are probably the oldest. Others are expected to change also and the semiotic theory outlined by Peirce provides for this since the ultimate level of sign dynamism is the symbolic sign.
Although, initially all these function words look like nouns and verbs they derive from because of the syntagmatic derivation, they end up changing at the phonetic level as well because of their new meaning and function in conformity with the semiotic principle of "one meaning one form".

This paper has shown that some grammatical categories are more basic than others. Most of the function words: conjunctions, prepositions derive from basic ones namely nouns and verbs. And others from intermediate categories such as adjectives, demonstratives, pronouns, etc.
Because of their high number and transparency, these function words are new in Kinyarwanda. This might be true as well for other Bantu languages as shown by the Swahili examples. This tendency of deriving derivational words from nouns and verbs seems to be universal for all the languages that have been investigated so far (Givón, Heine, Traugott, etc.). In the iconicity view of the genesis of signs and structures and in the semiotic theory of Charles Peirce's of teleology, this is expected and predicted.

VI. Implications:

The fact that conjunctions in Kinyarwanda are derived from verbs and the fact that these verbs are still recognizable indicate that not only are they new, but also that the process of conjunction is a new development in Kinyarwanda grammar. There are two factors which support this hypothesis. What are conjunctions in other languages are translated into Kinyarwanda with verbs. For instance, 'more than,' 'less than,' or 'about' are translated with kurusha 'to surpass' and ibyeérekeye 'what is related to' as examples in (66) show:
66. a. Unó mugabo aririimba néezá kurusha umugoré.
'This man sings better than the woman'
b. Yohaáni afite amafaraanga maké kurusha Mariyá.
'John has less money than Mary.'
c. Twaagaaniriye ibyeérekeye kwíiga.
'We talked about studying/studies.'
As I have shown also in other studies (Kimenyi, 1980) Kinyarwanda is a "linkless clause" language. In many cases, sentences can consist of multiple clauses which don't have morphemes or words to join them. Consider the sentences in (67):
67. a. Ba-geend-é ba-ba-bwiír-e ba-kor-é.
they-should-go they-should-tell-them they-should work
'They should go and tell them to work.'
b. Umuuntu baa-vugán-ye ni uumugabo tw-aa-guz-é igitabo.
person they-talked-to is man we-bought book
'The person they talked to is the man from whom I bought the
book.
c. Bá-maz-e ku-ger-a aha b-aára-gáaniiri-ye ba-ra-taah-a.
'they-finished to-arrive here they-talked they-went-home
'When they got here they talked and then went home.'
Coordination and subordination are accomplished by the verbal tense-aspect-modality system or by auxiliaries.
This is not unique to Kinyarwanda, because many Bantu languages such as Swahili use the tense-aspect-modality morphology for subordination and coordination (1972).
Questions which are not answered in this paper but which are worth linguistic inquiry are first why independent words have not been created to refer to these conjunctions and secondly whether these conjunctions entered the language independently or are innovations due to external factors.
To refer to new concepts a language, as Heine (1991) pointed out, uses five available strategies namely:
(a). inventing arbitrarily new words.
(b). borrrowing from other languages or dialects.
(c). creating iconic expressions such as onomatopeias.
(d). composing or deriving new expressions from already existing
lexical and grammatical forms.
(e). extending the use of existing forms for the expression of new
concepts, commonly describing strategies including analogical
transfer, metonymy, metaphor and the like.

The (e) strategy is universal.
In Swahili the verb kwamba 'to say' has become a complementizer and retains its full shape.
A-li-sema kwamba a-ta-rudi kesho. 'S/he said that s/he will be back
tomorrow'.
But it is also found without its infinitive prefix as a relative pronoun:

68. Hiki ni kitabu amba-cho nda-ni ya-ke ni-li-i-ona picha. 'This is the
book in which I saw the picture'.
this is book that-it stomach-in of-it I-past-it-see picture
Yohana, amba-lo jina la-ke ya pili ni Marko. 'John, whose
second name is Mark'.
John, who-it name of-him of two is Mark
Neno a-li-lo-li-sema amba-lo ha-tu-ta-li-sahau. 'The thing that he
said, and which we will not forget'.
thing he-pst-it-it-say that-it not-we-fut-forget

The verb -wa 'be' has also created many function words. It either appears in the infinitive form or in the finite form as the following examples show:

69. kuwa (ku-wa)'that'
ikiwa (i-ki-wa) 'if, whether'
ijapokuwa (i-ja-po-ku-wa) 'even so, even if, although'
isipokuwa (i-si-po-ku-wa) 'except that'
huwa (hu-wa) 'generally'

In other Bantu languages such as Chichewa, it is the same 'say' verb ku-ti which also functions as a complementizer. In Kinyarwanda as noted earlier the verb ku-ti has limited usage and is used only to report direct speech.
Turkish borrowed from Persian (Erguvanli, Personal communication) the complementizer and Swahili has borrowed from Arabic kama and kana 'like/as/'. Swahili again uses kwamba 'to say' as a complementizer and as a relative pronoun.
69. Ni ndizi nzuri kama zile za jana. 'They are good bananas like those
of yesterday'.
be bananas good like those of yesterday
Rangi yake nyekundu kama damu. 'Its color is red like blood'.
color its red like blood
Kama u-penda-vyo. 'As you like it '
as you-like-it
Urefu wake ni kama futi kumi. 'Its length is about ten feet'.
Kama nyama ha-i-patani, nunua samaki. 'Since there is no meat,
buy fish'.
if meat it-is-not-found buy fish
Si kama a-li-i-ona mwenyewe. 'It is not as if he has seen it
himself'.
not like s/he-it-saw himself
Si-ju-i kama wa-ta-ku-ja. 'I don't know whether they will come'.
not-I-know-not if they-will-to-come
A-li-sema kama ndivyo hivyo. 'He said that that is how it was'.
he-past-said that these those
There are many more such as zaidi, baada, kabla, karibu, etc.
zaidi ya kumi 'more than ten'
baada ya chakula 'after food'
kabla ya saa nane 'before 2 o' clock'
baina ya vitu hivi 'between these things'

What this study has shown is that function words such as complementizers, conjunctions, subordinators, deverbal conjuntions are the newest because they are still recognizable formally and semantically. Others are monosyllabic because they are the oldest but they too are derived from other words because they don't belong to the primary category. Metalinguistic functions come second to referential words. Thus linguistic expressions referring to concrete concepts are triggers to their respective targets which happen to be more abstract concepts. Grammaticization exists because experience precedes language and language cannot keep up with the pace of human experience. This polysemous asymmetry is natural in human language. Asymmetry, polysemy, homonymy, deiconization, grammaticization, morphologization, decategorization, lexicalization or degrammaticalization , ... are all natural processes in natural languages. Asymmetry between the real world structures, conceptual structures and linguistic structures is natural. The real world structure and experiences are infinite but conceptual structures are constrained and linguistic structures and signs are finite. Words which refer to first experiences are the ones which are used to express new experiences metaphorically, this creates polysemy. This polysemy results from hybridization. Hybridization exists when the same word occurs in different grammatical categories and when the structure can tell what their nature is. In some cases, this polysemy is avoided by the creation of doublets, change in the phonetic form. This is the first step toward deiconization and decategorization.
Decategorization occurs if the new meaning has a different grammatical category such as from verb to noun, noun to verb, noun to adjective, etc. The words that have been discussed here are also desemanticized because they only have a grammatical function and no more referential function.
All this phenomenon is referred to in semiotic theory as deiconization, and in functional grammar as grammaticalization. In functional grammar, it is mostly restricted to morphologization and to syntactization, because language is supposed to consist initially of nouns and verbs as evidenced by creolization and child language acquisition and that grammar comes later.
In Peircian semiotics, deiconization includes also the genesis of tropes, metaphors and metonymies and all this process of decategorization, morphologization, syntactization.
All this process is never ending. As Givón has pointed out, it is cyclical. This has prompted Hopper to talk about emerging grammar because being dynamic, the language doesn't have any fixed structure. Once a linguistic item has been deiconized becoming an affix or a clitic, it can undergo degrammicization again: lexicalization such as the suffix -teen as in thirteen, fourteen, fifteen, sixteen, which has now become a word on its own. as child in his teens, teenage years, a teenager. All this shows that an adequate description and explanation of language which is dynamic, has always to be done panchronically: diachronically and synchronically.

Notes:

1. Swahili spelling has started differentiating, function words from content words:
kuanza 'to start' <kwanza 'first'
kuamba 'to talk' <kwamba 'that'
2. Swahili does the opposite. Instead of reducing the auxiliary, it is the main verb which is reduced instead. The auxiliary carries all the tense-aspect-modality morphemes and the main verb loses them as the verb
kwiisha 'to finish' shows which as a tense marker indicates the completion of an action.
a-me-kw-iisha lishwa na dada yake.
a-me-kw-iisha nyonyeshwa na mama yake.
a-me-kw-iisha lizwa
wa-me-kw-iisha pewa (chakula) na mama yake

References:

Anttila, R. 1971. An Introduction to Historical and Comparative
Linguistics. New York: The MacMillan Company.
Bolinger, D. 1949. "The Sign is not arbitrary." Boletin del Instituto Caro y
Cuervo, 5:52-62. Bogota.
Givón, T. 1971. "On the origin of Bantu verb suffixes". Studies in African
Linguistics 2, 2: 145-63
Givón, T. 1980. Understanding Grammar.
Guthrie, M. 1971. Comparative Bantu. Gregg International.
Heine, Bernd. 1991. "Auxiliaries in African Languages: the Lingala case".
BLS 18: 106-119
Heine, B.; Claudi U.@F. Hünnemeyer. 1991. Grammaticization: A Conceptual
Framework. Chicago and London: University of Chicago Press.
Trugott, E.&P. Hopper. 1993.
Jakobson, R. 1962 and 1971. Selected Writing I-II. The Hague: Mouton
Kimenyi, A. 1981. "Linkless Clauses in Bantu/" 12th Annual Conference on African Linguistics, Stanford, in Leben (ed.).
Kimenyi, A. 1982. "Syntax and Semantics of Reduplication in Kinyarwanda:
a Semiotic Account." 13th Annual Conference on African Linguistics.
Québec. Published in Linguistics,
Kimenyi, A.1982. "On the Non-arbitrariness of the Linguistic Sign." California
State University, Sacramento.
Langacker, R. Conceptual Grammar.
Peirce, C. 1931-1958. Collected Writing. Cambridge, England: Harvard
University Press.
Shapiro, M. Language as Semeiotic. Bloomington: Indiana University Press.