Artikelaktionen
Remnants of a noun class system in Bezen (Southern-Jukunoid)
urn:nbn:de:0009-10-38342
Abstract
Harshen Bezen ya yi kama da agin sūnā na harshen da kèe Benue-Congo. Akwai ire-iren ɗoriyar farkoo a tattare da Kalmomin sūnā. Saɓanin tūnanin māsānā harasa (na harshen Bezen), tsarin gituwar bā yā nuunaa yawan (lambōbu) yankunan dōrūga (anna) duk da dai ya ragu zuwàa ajujuura huɗu. Akwai ɗaya tilo kuma akwai ukù jam’i. Arufanin ɗoriyar farko naa suunaa bai tsaayaa kawai a nūnā alamar yāwā ba, amma yā hāɗā da kirkiro dà kalmōnin sūnā dà yāwā daga filōlī da sifōfī.
Abstract
Bezen looks like a typical Benue-Congo noun class language: A rich variety of prefixes adorn its nouns. Contrary to the linguist’s expectation, the agreement system does not reflect the high number of nominal prefixes but is reduced to four classes: One singular and three plural classes. The function of the nominal prefixes is not restricted to number-marking, however. They also serve to derive a large array of nouns from verbs and adjectives.
Zusammenfassung
Bezen Nomen sehen wie typische Benue-Kongo Nomen aus, sie bestehen aus einem Präfix, das Singular oder Plural markiert und einer Wurzel. Entgegen aller Erwartungen spiegelt sich die Vielfalt der Präfixe nicht im Konkordanzsystem der Sprache wider: Es sind nur noch vier Klassen von einem ehemals wohl elaborierterem System übrig. Die Funktion der Präfixe ist nicht auf Numerusmarkierung reduziert, sie sind äußerst produktiv bei der Derivation von Nomen.
1. Introduction [1]
<1>
Bezen nouns have the typical appearance of Benue-Congo nouns, having a prefix and a root, with one set of prefixes marking singular, the other plural. The prefixes have the form CV- or V-; C can be represented by /k/ or /b/; V having the form /a/, /i/, /e/, /o/, or /u/ [2] . The root prototypically has the form CVC, but there are also roots bearing the form CV or VC. In the latter case, a consonant loss might be responsible for the syllable-structure. More complex roots, as CVCCVC, CVCVC, or CVCV are assumed to be the result of compounding or associative constructions.
<2>
Many nouns show a qualitative coherence of the nominal prefix and the root vowel as a result of assimilatory processes that may operate in both directions across morpheme boundaries. That means that in some cases, the prefix vowel can trigger a phonetic change of the root vowel and in other cases, the root vowel triggers an allophonic change of the prefix. And sometimes, the change of a prefix vowel just cannot be explained anymore.
<3>
Regressive assimilatory processes, in which the quality of the nominal root vowel influences the prefix vowel, lead to a high allomorphic variety of the latter. One possible assimilation pattern is presented in Table 1, where V represents the plural prefix. While an /a/ in the root triggers an a- as the plural prefix morpheme, the unrounded vowels /ɨ/ or /i/ lead to an ɛ- as the plural morpheme. A rounded vowel /u/ in the root triggers the prefix vowel o-. This pattern occurs also occurs with bV- and kV- prefixes. However, it is not universal, and there are many examples that do not show assimilation.
|
Table 1: |
Allomorphy of nominal prefixes |
|||
|
V |
→ |
a- / _ Ca(C), e.g. kìzá / àzá |
winnowing tray |
|
|
V |
→ |
ɛ- / _ Cɨ(C), e.g. kídɨ́r / ɛ́dɨ́r |
sausage tree (Kigelia Africana) |
|
|
ɛ- / _ Ci(C), e.g. kíʃí / ɛ́ʃí |
egg |
|||
|
V |
→ |
o- / _ Cu(C), e.g. kíkúr / ókúr |
end |
|
<4>
The nominal prefixes additionally show a variation of the prefix vowels /u/ ~ /o/ and /i/ ~ /ɛ/, which will be considered as allophonic in the following, but is not triggered by the quality of the root vowel. Prischnegg, (2008:133), who noticed the same phenomenon in Yukuben, proposes that the variation is the result of an assimilatory process, where the underlying prefix vowel /u/ or /i/ merges with a root-initial vowel /a/. However, it is difficult to provide an evidence for such an explanation in Bezen, as there are very few vowel-initial nominal roots left in this language. Shimizu (1980b) also recognizes the variant oo- of the noun class prefix u- in Yukuben, but he does not explain the variation and considers it as an allomorphic phenomenon. Shimizu’s reconstructions of Proto-Jukunoid nominal roots and the classes they belonged to (1980b, 1980c) might give a hint to the former class membership of Bezen nouns. Anyhow, in time lexemes might change their class belonging and sometimes different lexical roots are used in Bezen instead of the reconstructed PJ-roots. Thus, the number of corresponding cognates in an original class is quite small. A final explanation for the variation of these prefix vowels cannot be provided here.
<5>
Furthermore, in many lexemes an allophonic variation of the root vowel can be observed. Depending on the roundedness of the prefix vowel, the root vowel / ɨ / might change to its allophones / ʏ /, or / ʉ /, respectively, as shown in Table 2. Rounded prefix vowels trigger the realization / ʉ / of the root vowel, palatalized root-initial consonants trigger the realization / ʏ /, whereas unrounded prefix vowels do not have an effect on the root vowel. This process might also be triggered by bV - and kV - prefixes.
|
Table 2: |
Allophonic variation of nominal root vowel |
|||||
|
/ɨ/ |
→ |
[ʉ, ʏ, ɨ] |
||||
|
/ɨ/ |
→ |
[ʉ] / V[+round]C _(C) |
e.g. |
/ūtɨ́b / [ūtʉ́b] |
spear |
|
|
/ɨ/ |
→ |
[ʏ] / V[+round]Cy _(C) |
e.g. |
/ùhyɨ̄n / [ùhyʏ̄n] |
boundary |
|
|
[ʏ] / V[+round] y _(C) |
e.g. |
/ùyɨ́n / [ùyʏ́n] |
monkey |
|||
|
/ɨ/ |
→ |
[ɨ] / V[-round]C_(C) |
e.g. |
/ītɨ́b / [ītɨ́b] |
spears |
|
|
[ɨ] / V [-round]Cy_(C) |
e.g. |
/ìhyɨ̄n / [ìhyɨ̄n] |
boundaries |
|||
|
[ɨ] / V[-round] y_(C) |
e.g. |
/bèyɨ́n / [bèyɨ́n] |
monkeys |
|||
Just as with the processes that have been described before, not all nouns adhere to this assimilation patterns and there are many exceptions.
<6>
Corresponding SG and PL prefixes bear the same tone in most cases. Therefore, it must be concluded that the prefix tone is prescribed by the tonal value of the root. However, the variety of tonal patterns of the nouns does not allow a conclusion about how exactly the tone of the root influences the prefix tone.
<7>
Bezen lacks an elaborate agreement system, and the agreement markers that are left can often be used interchangeably. Therefore, a sharp division of noun classes is not possible anymore. Thus, the singular/plural pairings will be presented before an elaboration of the agreement system, which will be dealt with in chapter 3. The combination of several SG/PL pairs into one group is partly based on formal criteria as the assimilatory phenomena described above. Furthermore, the Bezen prefixes have been compared with Proto-Benue-Congo (De Wolf 1971) and Proto-Jukunoid (Shimizu 1980a; 1980b) reconstructions. In few cases, a semantic grouping of nouns can be observed with one SG/PL-pair, human beings for example are often denoted by nouns bearing u- ~ o- / bV- prefixes, while animals prefer i- ~ ɛ- / bV-. Liquids are exclusively found with a bV- prefix bearing a mid- or a high tone.
<8>
Six nouns combine the singular prefix u- ~ o- which bears a mid- or low-tone with a ba- plural prefix. Five of the nouns denote human beings (1) – (3) or animals (4) and (5). The lexeme ùtàr ‘garment’, which also allows the plural prefix ì-, is a semantic exception within this set of nouns. It is not clear, why lexemes (1) – (4) have an o- prefix, while (5) has an u- in the SG. Shimizu (1980b:172) reconstructs a class u- / ba- noun root *ngiT ‘person’, which seems to be a cognate of the Bezen lexeme.
<9>
Further nouns that denote human beings appear within the prefixes u- / bɛ-, as in (6) – (8).
<10>
Other nouns within this group denote inanimate objects (9), (10).
It is arguable, whether a separate group u- ~ o- / bɛ- is justified or whether these prefixes should be considered as a part of the u- ~ o- / ba- ~ bo- group. One evidence for a joined u- ~ o- / ba- ~ bo- ~ bɛ- group would be the lexeme ūpí ‘slave’, for which a class u- / ba- PJ-root *pyí has been reconstructed by Shimizu (1980b:10). Considering examples (6) – (8), one could argue that the bɛ- PL-prefix is a result of regressive assimilation to an unrounded close front- or central-vowel. However, the nouns in (1) – (3) show the same vowels in the root, but bear a ba- PL prefix, for which there is no obvious explanation. Therefore, two different groups have been established here.
<11>
In Bezen, many nouns denoting animals have the prefix ì- ~ ɛ̀- in the singular and bà- ~ bò- prefix in the plural (11) – (15). The quality of the plural prefix vowel is affected by the roundedness of the root vowel, resulting in bò-, when the root vowel is rounded (11) – (12) and in bà- when it is unrounded (13) – (15) [3] . īkáʒím / bā- ‘spider’ is an exception within this group of nouns, bearing a mid-tone prefix.
|
ìkúr / bòkúr |
crocodile |
|
|
ìsɨ́n / bòsʉ́n |
fowl |
|
|
ìhīr / bàhīr |
helmeted guinea fowl |
|
|
ìkār / bàkār |
baboon |
|
|
ìkāháŋ / bàkāháŋ |
bushfowl |
<12>
A set of nouns denoting animals have ɛ̀- / bà- as prefixes, among them ɛ̀mʌ̄n ‘goat’, ɛ̀ɛ̀m ‘hippo’ and ɛ̀ɛ̀r ‘buffalo’. A comparison with the closely related languages Yukuben and Kuteb shows that the latter two nouns have lost a root initial glide /y/ in Bezen. Regressive assimilatory processes have led to a lowering of the SG prefix ì- to ɛ̀- and a change of the root vowel in the plural, triggered by the bà- prefix (16). Why ɛ̀mʌ̄n ‘goat’ decided to take ɛ̀- as its SG prefix, is not clear.
|
Bezen |
Yukuben |
Kuteb |
||
|
ɛ̀ɛ̀m / bààm |
īyìm / bēyìm (Prischnegg 2008:142) |
ìyém / ìyém (Koops 2009:277) |
hippo |
|
|
ɛ̀ɛ̀r / bààr |
īyà / bāyà (Prischnegg 2008:141) |
ìyāg / ìyāg (Koops 2009:99) |
buffalo |
<13>
Five relational nouns bear the prefixes i- / ba- ~ bo- with mid-tones in (17) – (19), the lexemes ìrâ / bàrâ ‘friend’ and ímbār / bámbār ‘sibling’ being the only examples with a L-tone or a H-tone.
<14>
Several nouns that denote animals and humans bear the prefixes ì- ~ ī- / bɛ̀- ~ bɛ̄- (20) – (25).
|
ìgbɨ̄r / bɛ̀gbɨ̄r |
dog |
|
|
ìkyàm / bɛ̀kyàm |
horse |
|
|
(22) |
ìkyɨ̀n / bɛ̀kyɨ̀n |
guest |
|
īyī / bɛ̄yī |
in-law |
|
|
īmbyɛ̄r / bɛ̄mbyɛ̄r |
mother-in-law |
|
|
ītʃīn / bētʃīn |
parent |
Here again, it is arguable, whether it would make more sense to combine the two prefix groups i- ~ ɛ- / ba- ~ bo- and ì- ~ ī- / bɛ̀- ~ bɛ̄- to one, as both groups contain nouns denoting animates. However, considering a joined group, there would be no plausible explanation, why in some cases the plural prefix vowel is /a/, but in others /e/, as both vowels co-occur with unrounded root vowels.
<15>
Several nouns denoting abstract concepts (26) and (27), transnumerals, which can refer to singular or plural objects (28) and (29), and fluids (30), take a ba- prefix.
<16>
Other abstract nouns (31) and (32) transnumerals (33) and (34) and fluids (35) and (36) take the prefixes bi- ~ bɛ-. Nouns denoting fluids appear exclusively with mid- and high-tone prefixes.
|
bɛ̄ŋmàm |
laziness |
|
|
bɛ̄ɛ̄n |
whistling |
|
|
bìkpōŋ |
forehead |
|
|
bīmám |
bushy end of a tail |
|
|
bɛ̄sɨ̄m |
corn beer |
|
|
bímí |
water |
<17>
Two ethnonyms are found with the L-tone prefix bì- (37) and (38).
<18>
Further mass nouns (39) and (40) and abstracts (41) and (42) bear the prefix bu- ~ bo-.
One could consider a joint grouping of the five prefixes ba-, bi- ~ bɛ- and bu- ~ bo- as they appear with nouns bearing the same semantic content. Even though there is a tendency of rounded root vowels triggering rounded prefix vowels, not all nouns adhere to this assimilatory pattern, as in (30) – (33) and (41). For this reason, the three groups are kept apart.
<19>
Nouns with a mid- or high tone SG prefix u- take almost exclusively the PL prefix i-(43) – (47).
|
ūkúŋ / īkúŋ |
edge |
|
|
(44) |
ūsɨ̄n / īsɨ̄n |
hair |
|
ūzɨ̀ / īzɨ̀ |
broom |
|
|
(46) |
úsàn / ísàn |
farm |
|
úyāk / íyāk |
stirring stick |
<20>
A part of the nouns has the mid- or high-tone variants o- / e- as number marking prefixes (48) – (51).
|
ōkūn / ēkūn |
firewood |
|
|
ōtʃī / ētʃī |
tree |
|
|
ɔ́ɔ̄ŋk / ɛ́ɛ̄ŋk |
flute |
|
|
ómɨ̄n / émɨ̄n |
raw one |
Furthermore, three nouns take the L-tone prefixes ù- / ì- (52) – (54). The lexeme ùhyūn ‘boundary’ (53) allows bɛ̀- as alternative PL-prefix and ùtàr ‘garment’ (54) accepts the alternative bà-.
A remarkable number of nouns denoting elongated objects is found within this group of nouns, as in the examples (45), (47) – (50) and (53).
<21>
The prefixes i- ~ ɛ- occur with mid and high tones in transnumerals (55) and (56) mass nouns (57) and (58) and nouns that denote abstract concepts (59) and (60).
<22>
The SG prefix ki- appears with a low, mid or high tone and combines with a tonally fitting PL prefix a- ~ o- ~ ɛ-. The vowel of the plural prefix is dependent on the quality of the nominal root vowel. While an /a/ in the root leads to a- as plural prefix (61) and (62), a rounded root vowel /u/ prescribes the plural prefix o- (63) and (64) and any other unrounded vowel as /ɨ/ or /i/ leads to an ɛ- plural prefix (65) and (66).
|
kīgār / āgār |
forest |
|
|
kìbàr / àbàr |
bag |
|
|
kíhùr / óhùr |
hole |
|
|
kíkúr / ókúr |
bundle |
|
|
kìhyɨ̄ŋ / ɛ̀hyɨ̄ŋ |
drum |
|
|
kīʃī / ɛ̄ʃī |
head |
<23>
Even though the majority of the nouns adhere to this assimilation pattern, there are also exceptions: Some nouns combine a rounded root vowel with an unrounded plural prefix (67) and (68) or the other way around (69). kízɨ́n allows the alternative plural form ɛ́zɨ́n in all three meanings.
<24>
A set of transnumerals and abstract nouns appears with an a- prefix, bearing either a L, M, or H-tone (70) – (75). Two lexemes that denote human beings show an ā- prefix without a fitting plural form: āyà ‘mother’, and ābà ‘father’.
<25>
There are three nouns denoting inanimates that bear the prefix kà- ~ kā- in the SG and kù- ~ kò- ~ kō- in the PL (76) – (78). The examples are too few to allow a conclusion regarding the direction of assimilation in (77) and (78), but the long vowels in both nouns indicate a consonant loss.
|
(76) |
kàkāŋ / kòkāŋ |
cap |
|
(77) |
kàátàk / kùútùk |
calabash |
|
(78) |
kāār / kɔ̄ɔ̄r |
canoe |
<26>
A small number of nouns bears the prefix ka- ~ kɛ- in the singular and an a- prefix in the plural. The quality of the SG prefix vowel is again dependent on the root vowel of the noun: While /a/ as root vowel triggers ka- as SG-prefix (79) and (80) any other root vowel as /u/, /ɛ/, /ɨ/ or /i/ will trigger kɛ- (81) – (84).
<27>
Several transnumerals (85) – (88) and a noun that denotes an abstract concept (88) bear the prefix ka-.
<28>
A set of nouns appears with the SG prefix ku- ~ ko- and the PL prefix a- (89) – (94).
|
(89) |
kōtɨ̄ / ātɨ̄ |
bow |
|
(90) |
kōkʊ́n / ākʊ́n |
cup |
|
(91) |
kùmán / àmán |
grasshopper |
|
(92) |
kùgbān / àgbān |
lizard |
|
kókùŋ / ákùŋ |
sugarcane |
|
|
kóbɨ̄ / ábɨ̄ |
palm frond |
<29>
In two cases the plural prefix a- seems to be assimilated to the rounded root vowel and has the shape o- (95) and (96).
<30>
Three nouns show the number-prefix pairing kū- / ī- (97) – (99).
<32>
Lexemes with bi- as SG prefix might take bu- ~ bo- in the PL (100) – (105).
<33>
In three cases the pairing bī- ~ bí- / ī- ~ í- can be observed, for example in (106) – (108).
<34>
Two nouns appear with bū- as the singular prefix and bī- ~ bɛ̄- in the plural (109) and (110).
<35>
The singular prefix bū- combines with the plural prefix ī- in būlāk ‘palm oil tree’.
<36>
Loanwords in Bezen originate from Hausa (111) – (115) and Jukun (116). These nouns alert attention because of their mostly missing noun prefix in the singular. Bezen speakers integrate these nouns into their nominal number system by adding a bò- ~ bō- prefix in the plural. àlɛ́mṍ ‘orange’ in (114) has a Bezen atypical SG-prefix à-, which is not dropped in the PL. The Bezen might have borrowed the lexeme from Hausa or Jukun, together with the additional phonological material. The prefix ū- of ūtábā ‘tobacco’ in (115) could be of Bezen origin, as there is an array of Bezen lexemes bearing this prefix in the SG. It is the only loanword bearing the plural prefix bɛ̄-.
|
górò / bōgórò |
colanut |
|
|
(112) |
kúlɛ̄ / bōkúlɛ̄ |
cat |
|
tásā / bòtásā |
iron pot |
|
|
àlɛ́mṍ / bòàlɛ́mṍ |
orange |
|
|
ūtábā / bɛ̄tábā |
tobacco |
|
|
lóŋ / bòlóŋ |
trousers |
Neologisms are treated just as loanwords: a bò- prefix is added in the PL. bàkwàr is a newly created lexeme to designate a short type of bananas, which grow in the area of the Bakuri people in South-western Cameroon.
<37>
Bezen differentiates four different agreement classes: one singular and three plural classes. In the singular, all nouns trigger the agreement prefix u- ~ o-, bearing a mid- or high tone. In the plural, three different agreement markers, i- ~ ɛ-, ba- ~ bɛ- ~
bo-, and a- with the same tonal values as in the SG are found [4] (Table 3). The quality of the vowels is dependent on the target’s root vowel and will be discussed in the following.
|
Table 3: |
Agreement morphemes |
||
|
sg |
for all nouns |
u- ~ o- |
|
|
pli |
for non-humans and inanimates |
i- ~ ɛ- |
|
|
plii |
for humans and inanimates |
ba- ~ be- ~ bo- |
|
|
pliii |
for āllɨ̄ ‘days’ |
a- |
|
<38>
In the SG all nouns trigger the agreement marker u- ~ o- in adjectives, numerals, and demonstratives (117) – (120). Agreement is furthermore found with interrogative adjectives. As that is more elaborated and deviates from agreement with the other three targets, it will be explained in detail in the according chapter.
|
(119) |
ímbar |
ú-nānɨ́ |
|
sg.sibling |
sg-this |
|
|
this sibling |
||
|
(120) |
bīdáŋ |
ú-nânɨ́ |
|
sg.chair |
sg-that |
|
|
that chair |
||
<39>
Considering the plural, the picture becomes more blurred. Whereas nouns denoting human beings prefer a bV- prefix (121) and (122), all other nouns vary in their usage of the two plural agreement prefixes (123) – (125). Assuming semantic agreement, non-humans would not allow bV- agreement, but they sometimes do, as in (124). Formally motivated agreement would mean that nouns that have a bV- plural prefix trigger bV- agreement, which is also not always the case (125). That means that agreement in Bezen has to be learned together with the noun. However, the speakers allow for variation and often disagree among each other upon the “right” agreement marker. A learner of Bezen would be on the safe side to use bV- agreement with humans, and nouns that carry a bV- plural prefix and i- with all other nouns.
|
bākɨ̄b |
bá-kyɨ́r |
||
|
pl.women |
plii-small |
||
|
small women |
|||
|
bámbār |
bɛ̄-wúŋ |
||
|
pl.siblings |
plii-other |
||
|
other siblings |
|||
|
ísàn |
ī-lákàr |
ī-tār |
|
|
pl.farm |
pli-big |
pli-three |
|
|
three big farms |
|||
|
īlɨ̄k |
bɛ́-yī |
||
|
pl.ropes |
plii-new |
||
|
new ropes |
|||
|
bɔ̀hɔ̀r |
í-nânɨ́nɨ́ |
||
|
pl.hook |
pli-those (not visible) |
||
|
those hooks |
|||
<40>
Bezen has a relatively small set of adjectives that refer to the color, age, or quality of objects or people. The most prominent ones are lákàr ‘big’, màn ‘red’, rɨ̀ ‘good’ and bì ‘bad’. The agreement marker ú- ~ ó- is the same for all singular head nouns: In (126), lákàr ‘big’ refers to a human being and in (127) bì ‘bad’ to an inanimate entity.
<41>
In the plural, three different agreement morphemes can be observed: bV-, i- ~ ɛ- and a-. Whereas the agreement marker a- only occurs with the pluralic noun āllɨ̄ ‘days’, the other two morphemes are evenly distributed with the overall number of Bezen nouns. However, no clear-cut distinction of classes can be made, as one noun might take different agreement morphemes with different adjectives, or take one agreement morpheme for adjectives and another for demonstratives. Certain is that nouns denoting human beings take the agreement marker bV- with all adjectives (128) – (130).
|
bāɲī |
bá-bì |
|
|
pl.person |
plii-bad |
|
|
bad people |
||
|
(129) |
bɛ̄dɨ̄ŋ |
bɛ̄-wúŋ |
|
pl.chiefs |
plii-other |
|
|
other chiefs |
||
|
bámbār |
bá-rɛ̀ |
|
|
pl.siblings |
plii-good |
|
|
good siblings |
||
<42>
Non-humans take the i- or the bV- prefix in the plural. The adjective wúŋ ‘other’ plays an exceptional role, as many nouns might take i- as agreement morpheme with all adjectives but wúŋ, which would take bV- instead. These nouns do not necessarily have a bV- plural prefix or are animated, as in (131) – (134).
|
(131) |
íɲí |
ɛ́-kyɨ́r |
|
pl.mouths |
pli-small |
|
|
small mouths |
||
|
(132) |
íɲí |
bɛ̄-wúŋ |
|
pl.mouths |
plii-other |
|
|
other mouths |
||
|
(133) |
ɛ̄kūn |
ī-lákàr |
|
pl.firewood |
pli-big |
|
|
big firewood |
||
|
(134) |
ɛ̄kūn |
bɛ̄-wúŋ |
|
pl.firewood |
plii-other |
|
|
other firewood |
However, there are also nouns that strictly take the agreement marker i-. These nouns tend to have a kV- SG-prefix (135) – (138).
|
ɛ́ɲín |
ɛ́-kyɨ́r |
|
|
pl.scars |
pli-small |
|
|
small scars |
||
|
(136) |
ɛ́ɲín |
ī-wúŋ |
|
pl.scars |
pli-other |
|
|
other scars |
||
|
(137) |
ākún |
ī-lákàr |
|
pl.cups |
pli-big |
|
|
big cups |
||
|
ākún |
ī-wúŋ |
|
|
pl.cups |
pli-other |
|
|
other cups |
||
The noun bɔ̀hɔ̀r ‘hooks’ (SG bɛ̀hɛ̀r) takes ì- agreement with all adjectives in the plural, disregarding its bV- SG and PL-prefixes.
<44>
The variation of the agreement morphemes u- ~ o-, i- ~ ɛ- and ba- ~ bɛ- depends on the quality of the target’s root vowel. An /a/ in the root of the adjective triggers the agreement prefixes u- in the SG and ba- and i- in the PL (139), whereas a root containing the vowel /u/ leads to the prefixes u-, bɛ- and i- (140). The vowels /ɨ/ and /i/ trigger the prefixes o-, ɛ-, and ba-, respectively (141) and (142). Here, the ba- plural prefix might indicate an ancient adjective-initial vowel /a/ which results in o- and ɛ- when meeting the prefixes u- and i-. Whereas in the u-, i-, ba- set, the ba- prefix might be the result of regressive assimilation to the root vowel of the adjective.
|
ūdūŋ |
ūlákàr |
b) |
bɛ̄dɨ̄ŋ |
bālákàr |
c) |
bɔ̄hɔ̄r |
īlákàr |
||||
|
sg.chief |
sg-big |
pl.chiefs |
plii-big |
pl.hooks |
pli-big |
||||||
|
big chief |
big chiefs |
big hooks |
|||||||||
|
bīdáŋ |
ū-wúŋ |
b) |
būdáŋ |
bɛ̄-wúŋ |
c) |
bàkār |
ī-wúŋ |
||||
|
sg.chair |
sg-other |
pl.chairs |
plii-other |
pl.baboons |
pli-other |
||||||
|
other chair |
other chairs |
other baboons |
|||||||||
|
úlɨ́m |
ó-kyɨ́r |
b) |
bɛ́lɨ́m |
bá-kyɨ́r |
c) |
átɨ̄n |
ɛ́-kyɨ́r |
||||
|
sg.child |
sg-small |
pl.children |
plii-small |
pl.mortars |
pli-small |
||||||
|
small child |
small children |
small mortars |
|||||||||
|
ímbār |
óbì |
b) |
bámbār |
bá-bì |
c) |
bàkār |
ɛ́bì |
||||
|
sg.sibling |
sg-bad |
pl.siblings |
plii-bad |
pl.baboons |
pli.bad |
||||||
|
bad sibling |
bad siblings |
bad baboons |
|||||||||
<45>
The Bezen numeral system is quintesimal, that is all numerals from six to nine are expressed through compounds based on 5. The agreement morphemes are ó- in óyʊ̀nə̄ ‘one’ and i- ~ ɛ-, bV- or a- in the plural and are marked on both numerals in case of compounds. óyùnɨ̄ ‘one’ is identical in combination with all nouns (143) and (144), the numbers above take different agreement markers dependent on the head noun of the phrase (145) – (148). The vowel of bV- is prescribed by the root vowel of the numeral: it is /a/ when the root vowel is /a/ as in bátār ‘three’ (147), /ɛ/ if the root contains the unrounded vowel /i/ as in bɛ̄ɲī ‘four’, and /o/ if the root vowel is rounded as /o/ as in bōtsōŋ̀ ‘five’ (147). ɛ́ɛ̄n ‘two’ is an exception, as here a mutual interference of the prefix and the root vowel can be observed, resulting in báān with a bV- prefix (145) and (146).
|
kìbàr |
ó-yùnɨ̄ |
||
|
sg.bag |
sg-one |
||
|
one bag |
|||
|
ōlɨ̄m |
ó-yùnɨ̄ |
||
|
sg.man |
sg-one |
||
|
one man |
|||
|
àbàr |
ɛ́-ɛ̄n |
||
|
pl.bags |
pli-two |
||
|
two bags |
|||
|
bālɨ̄m |
bá-ān |
||
|
pl.men |
plii-two |
||
|
two men |
|||
|
bālɨ̄m |
bō-tsōŋ̀ |
bá-tār |
|
|
pl.men |
plii-five |
pl-oo-three |
|
|
eight men |
|||
|
ātʃáŋ |
ī-tsōŋ̀ |
ī-tār |
|
|
pl.houses |
pli-five |
pli-three |
|
|
eight houses |
|||
<46>
The numerals above ten are expressed through phrases (149) and (151) – (153) or nouns (150). However, the numeral roots never appear in their bare form: They are always accompanied by a prefix, as in (149). Here, ɛ́ɛ̄n ‘two’ does not agree with any constituent in the noun phrase so that it might be plausible to establish a basic form of the numerals ‘two’ to ‘five’ that contains the PLi-prefix. Likewise, óyùnɨ̄ ‘one’ never appears without the ó- prefix. kɛ̄kɨ̄m ‘twenty’ has a plural form ākɨ̄m which is used in the formation of numerals of forty and more. In (152), an assimilation process seems to be responsible for the agreement prefix á- in áān, as in the next example, ītār ‘three’ is bearing the prefix i- instead of an expected a- (153).
|
bālɨ̄m |
kūwūb |
ōgbū |
ɛ́-ɛ̄n |
|
|
pl.men |
sg.ten |
pass.fact [5] |
two |
|
|
twelve men |
||||
<47>
Bezen has a set of demonstrative pronouns that shows the same agreement pattern as adjectives and numerals. The demonstrative pronoun nānɨ́ denotes an object or a person that is close to the speaker. The tonally modified nânɨ́ refers to an object that is further away from the speaker, but visible, whereas nânɨ́nɨ́ indicates an object that is not visible for the speaker (154) – (156) [6] . The agreement remains constant with all three targets and is ú- in the singular and í- or bɛ́- in the plural. Example (155) shows a noun denoting a human being, bearing kì- / à- prefixes, which is rather unusual. The agreement is formally motivated, as the targets bear the í- agreement marker. These three examples illustrate how disintegrated the Bezen agreement system really is. Whereas in (154) both, animacy and formality could play a role in triggering the bV- agreement prefix, in (155), agreement is formally motivated and in (156), formality does not play a role and the noun triggers PLi agreement, disregarding its bV- plural prefix.
|
èmɨ̄n |
ú-nānɨ́ |
b) |
bàmɨ̄n |
bɛ́-nānɨ́ |
|
|
sg.goat |
sg-this |
pl.goats |
plii-these |
||
|
this goat |
these goats |
||||
|
àndàb |
í-nânɨ́ |
b) |
àndàb |
í-nânɨ́nɨ́ |
|
|
pl.young. women |
pli-those |
pl.young.women |
pli-those (non-vis) |
||
|
those young women |
those young women (non-visible) |
||||
|
būtsúk |
ú-nânɨ́ |
b) |
bɛ̄tsɨ́k |
í-nânɨ́ |
|
|
sg.banana |
sg-that |
pl.bananas |
pli-those |
||
|
that banana |
those bananas |
||||
<48>
The interrogative adjectives māŋ ‘how many?’ and rɨ̄ŋ ‘which? / who?’ show agreement which is more elaborated than agreement with adjectives, numerals and demonstratives. Furthermore, the same adjectival stem is used to derive different question adjectives, as shown in the following.
<49>
For the question adjective māŋ ‘how many?’, three different categories of nouns have to be differenciated: human/non-human and non-countable. Additionally to that, the word āllɨ̄ ‘days’ has its own agreement morpheme. Nouns denoting non-humans are referred to by the prefix í- (157) and those denoting human beings by the prefix bá- (158). While in (157) àwū ímāŋ ‘of them how many?’ is a NP with an attributively used question adjective, in (158)the interrogative adjective refers anaphorically to afore mentioned human beings. Nouns denoting non-countable entities, or transnumerals, trigger the agreement marker ká- (159). This agreement marker only appears with the question adjective māŋ and does not occur with adjectives, numerals and demonstratives. The agreement morpheme a- only occurs with the lexeme āllɨ̄ ‘days’ and is consistent with all targets (160).
|
w-ōkú |
àwū |
í-māŋ? |
|
|
2sg-catch.fact |
3pl.o |
pl-how.many |
|
|
How many of them did you catch? (referring to fish) |
|||
<50>
For the question adjective rɨ̄ŋ ‘which?’, the differentiation between singular and plural becomes relevant again. In the singular, the agreement marker is ó- (161). Different from the aforementioned targets, no differentiation is made between nouns that denote human or non-human beings. All nouns in the plural are referred to by ɛ́rɨ̄ŋ (162) and (163).
|
ɛ́-rɨ̄ŋ |
ārə̄ |
ùtɨ́ŋ |
ùwù |
|
|
pli-which |
be.fact |
sg.own |
2 sg.poss |
|
|
Which ones are your own? (referring to children) |
||||
<51>
The same adjectival root is used to ask for ‘who?’, referring to several people (164) and (165). Even though agreement might have been the source of the bá- prefix, it is not indicated anymore in the interlinearization as singular and plural referents are distinguished by different roots (see below).
|
(164) |
bárɨ̄ŋ |
áwū |
ámʊ́ŋ |
kírī |
nɨ́ |
|
who |
come.sub |
here |
yesterday |
def |
|
|
Who came here yesterday (knowing that there were many)? |
|||||
<52>
Asking for one person, a different question adjective, ānɨ̀ŋ is employed (166) and (167).
<53>
It is remarkable that the same adjectival root nɨ̀ŋ is also used to create further question adjectives as kànɨ̀ŋ ‘how?’ (168) and ɛ̄nɨ̀ŋ ‘what?’ (169). The interrogative adjective ɛ̄nɨ̀ŋ ‘what’ can only be used when asking for processes, not for things. ‘what is this?’ would be ánɨ́ ārə̄ ɛ̀sɨ̀ŋ?.
The usage of one root for different question adjectives could hint to a former elaborate agreement system that later developed into different question adjectives.
<54>
Verbs do not show agreement with the nominal subject in Bezen. However, all verbs in the finite form do show a prefixed thematic vowel that is not part of the root and might have been an agreement marker in former times (see also Prischnegg 2008:178). The vocalic prefix has the form a-, ɛ-, or o-, which is mostly dependent on the quality of the verbal root. Roots containing unrounded vowels as /a/, /e/, /ə/ or /ɨ/ have either a- or ɛ- as prefix (170) and (171) [7] . Verbs containing a round vowel in the root in most cases will have an o- as prefix as V1 in (173), but sometimes also have an a- prefix as V2 in the same example. The vowel might change in order to mark 3pl, however, this is the only case and very predictable and would only show with verbs that do not bear a prefix /a/ anyway. The vowel seems to be semantically completely empty, but functions as a carrier of tones that indicate different TAM categories.
<55>
Which conclusion can be drawn from the described agreement phenomena? It is certain that the Bezen noun class system is highly disintegrated and unsystematic. It must once have been semantically motivated, but nowadays formality also plays a role in the determination of agreement morphemes. There is a tendency for head nouns denoting human beings to trigger a bV- agreement marker, while a noun denoting non-humans will ask for i- ~ ɛ- as agreement morpheme. However, there are many exceptions, as nouns denoting non-humans might also trigger the bV- agreement marker, disregarding the semantics (173). Furthermore, there are nouns denoting animates and humans, bearing a bV- plural prefix, and triggering an i- ~ ɛ- agreement marker in an adjective (174), but the bV-agreement in a numeral (175). The strangest case is where the same head noun triggers different agreement prefixes with different adjectives (176) and (177).
Thus, the overall picture is blurred and does not allow the drawing of a sharp boundary between the classes. Anyanwu reports (personal communication) that Yukuben also has a ‘crazy’ noun class system, which gives hope that the lack of systematics in the Bezen noun class system is not only due to the author’s brain capacity. The scattered occurrence of the additional agreement morphemes a- and ka- hints at a formally more elaborated system of concordance in Bezen.
<56>
The nominal prefixes described before are used to derive nouns from verbs and adjectives, by prefixing them to the verbal or adjectival root. A large number of derived nouns can be observed in the Bezen lexicon: Agentive and instrumental nouns are among them, but also action and state nouns. Furthermore nouns that denote the objects of actions or embodiments of qualities can be observed. However, the grouping of the nouns is often tentative and based on the interpretation of the author so that the categories should not be considered to be irrevocable.
<57>
Comrie & Thompson (2007:336) characterize an agentive noun as a lexeme that denotes “one which ‘verbs’”. In Bezen, the agentive noun ūʒí ‘thief’ is derived from the motion verb ʒí ‘to steal’ by prefixing ū- / bɛ̄-. Further agentive nouns are derived using a ki- prefix (178) – (180).
<58>
This group of derived nouns could be considered as a subsection of agentive nouns, except that the agents here are derived from stative verbs or adjectives, resulting in nouns that mean “sb. or sth. that is like that” (181) – (183). In the singular, the nouns bear the prefix ū- ~ ú-, in the plural, the prefix is bɛ̄- ~ bɛ́- if the agent is human a) and ī- ~ í-, if it is non-human b).
|
yɨ̄ |
→ |
ūyɨ̄ |
a) |
bɛ̄yɨ̄ |
b) |
īyɨ̄ |
|
|
new |
new one |
new ones (hum) |
new ones (non-hum) |
||||
|
? |
→ |
úlɨ́m |
a) |
bɛ́lɨ́m |
b) |
ílɨ́m |
|
|
child |
children |
fresh ones (non-hum) |
|||||
|
mān |
→ |
úmān |
a) |
bámān |
b) |
ímān |
|
|
red |
red one |
red ones (hum) |
red ones (non-hum) |
||||
Three more nouns belonging semantically into this group of nouns are derived by the prefixes bì-, kù- and kɛ̀- / à- (184) – (186) [8] .
<59>
In Bezen, kV- prefixes are especially prominent in the creation of instrumental nouns (187) – (190). Three different instrumental nouns can be derived from the verb rɛ̀n ‘to sift’: ùrān / bɛ̀rɛ̄n ‘basket’ and the nouns in example (187). ùrān is the only example in the corpus that employs the prefixes ù- / bɛ̀- for the derivation of an instrumental noun.
|
rɛ̀n |
to sift/untie/thatch { |
→ |
kɛ̄rɛ̄n / ārɛ̄n |
small basket for sifting |
|
|
→ |
kìrɛ̀n |
palmwine filter |
|||
|
hāk |
grind |
→ |
káhāk / áhāk |
upper part of a grinding stone |
|
|
(189) |
gbɨ̄n |
open, uncover |
→ |
kīgbɨ̄n / ēgbɨ̄n |
key, opener |
|
bɨ́b |
break |
→ |
kìbɨ́b āhām |
s.th. that breaks stones, f.e. hammer |
<60>
The group of state nouns contains by far the largest amount of nouns which can be derived from verbs (191) – (196), and adjectives (197) – (199). The prefixes bV- (191) – (195) and kV- (196) – (199) predominate here. The resulting nouns denote abstract qualities.
|
lɨ̄b |
be heavy |
→ |
būlɨ̄b |
eaviness |
||
|
(192) |
hyɨ̄m |
be sick |
→ |
būhyʉ̄m |
illness |
|
|
(193) |
tɨ̄ŋ |
think |
→ |
bàtɨ̄ŋ |
thought |
|
|
ŋmàm |
be lazy |
→ |
bɛ̄ŋmàm |
lazyness |
||
|
tsʉ̀n |
burn |
→ |
bótsʉ̀n |
heat |
||
|
hàm |
be mad |
→ |
kɛ̀hɛ̀m |
madness |
||
|
mān |
red |
→ |
kímān |
redness |
||
|
(198) |
yʉ̄ |
new |
→ |
kīyī |
newness |
|
|
kyɨ́r |
small |
→ |
kɛ́kyɨ́r |
smallness |
||
<61>
Comrie & Thompson report from the Bantu language Si-Luyana, where nouns with a passive meaning can be derived from verbs (Comrie & Thompson 2007:341; Givón 1970:74ff.). However, Givón has doubts concerning these derived nouns and shows examples that back their verbal status (1970:77). In Bezen, the derivations are clearly nouns with the meaning the «thing/person that is ‘verbed’» (Comrie & Thompson 2007:341). In the derivation of passive nouns, kV- prefixes are prominent (202) and (203), few examples appear with the prefixes ī- / bā- and ū- / ī- (200) and (201). The noun kāsām ‘peeling’ could as well be interpreted as ‘the result of peeling’ (202).
|
wān |
marry a man |
→ |
īwān /bāwān |
husband (the one who is married) |
|
|
tɨ́b |
pin |
→ |
ūtʉ́b / ītɨ́b |
spear (the one which is pinned) |
|
|
sām |
peel |
→ |
kāsām/ āsām |
peeling (the one which is peeled) |
|
|
bùk |
die |
→ |
kìbʊ̀k kōhūn |
bird (sp.) (the one which has been died with camwood) [9] |
<62>
There is a large amount of derived nouns in Bezen and in several cases, different nouns can be derived from one verbal or adjectival root (see also example (187): The root lɨ̄b ‘be heavy’ opens two ways of derivation: in (204a), the prefix bū- creates a state noun referring to the quality of an object, while in b) bō- derives a noun that denotes a concrete embodiment of the quality. In (205a), an agentive noun is derived by the prefixes ū- / bɛ̄-, while in (205b) the prefix bū- derives a state noun bearing an abstract meaning. The verb wān ‘marry a man’ allows the derivation of two nouns (206), the abstract concept of ‘marriage’, kīwān a) and the object of a marriage, the ‘husband’ īwān b). īwū ‘wife’ c) could be considered as the agent of the marriage but it would be the only case in Bezen, where the root of the verb loses a consonant in the derivation process. Similar to the previous example, the adjective mān ‘red’ (207) allows the derivation of a noun denoting an abstract concept a) and the agents of ‘redness’, here in the human b) and non-human form b).
|
lɨ̄b |
be heavy |
→ |
a) |
būlɨ̄b |
heaviness |
b) |
bōlɨ̄b |
load |
|
|
ʒí |
steal |
→ |
a) |
ūʒí/bɛ̄ʒí |
thief |
b) |
būʒí |
theft |
|
|
wān |
marry |
→ |
a) |
kīwān |
marriage { |
b) |
īwān/bāwān |
husband |
|
|
c) |
īwū / bōwū |
wife |
|||||||
|
mān |
red |
→ |
a) |
kímān |
redness { |
b) |
úmān/bámān |
the red one (+hum) |
|
|
c) |
úmān / ímān |
the red one (hum) |
<63>
Furthermore, noun to noun derivation can be observed, even though to a smaller extent than verb to noun derivation. The abstract concept ‘chiefdom’ is derived from the noun ūdʊ̄ŋ ‘chief’ (208). The plural form of the noun īzɨ̄n ‘child’, which is a relational concept, does not mean ‘children’, but ‘young animals’. For ‘children’, the pluralic noun bɛ́lɨ́m is used, instead. The noun kīzɨ̄n ‘grandchild’, is derived from ‘child’ īzɨ̄n.
|
ūdʊ̄ŋ / bɛ̄dɨ̄ŋ |
chief |
→ |
kīdɨ̄ŋ |
chiefdom |
|
|
(209) |
īzɨ̄n |
child { |
→ |
bōzʉ̄n |
young animals |
|
→ |
kīzɨ̄n |
grandchild |
|||
<64>
Semantically, some parallels to non-derived nouns are observable: many non-derived nouns denoting abstract concepts also carry a kV- prefix (210) – (213), whereas abstract nouns carrying a bV- prefix are mostly of derived origin (191) – (195). Furthermore, many derived (187) – (190) and non-derived instruments (214) – (217) are found with kV- / V- or only kV-prefixes.
<65>
Bezen nouns show a large variety of nominal prefixes that indicate SG and PL and which are not reflected in the agreement system. A tendency towards semantic grouping of nouns with similar prefixes can be observed. The agreement system is reduced to four classes, one SG and three PL classes. It seems as if agreement in Bezen used to be semantically motivated, differentiating between humans and non-humans. However, the system has disintegrated and nowadays formality also plays a role in the determination of agreement markers [10] . The prefixes are furthermore used to derive nouns from verbs and adjectives, with state nouns denoting abstract concepts being the most prominent outcome of derivation. The Bezen noun class system fits very well into the Southern Jukunoid noun class-pattern. The Bezen nominal prefixes are almost identical with those of Yukuben, which has retained a slightly more elaborated agreement system. This hints to a probably more complex former noun class system in Bezen that has been reduced either as part of the general tendency to decline of Benue-Congo noun class systems or as part of an overall, language-intern dismantling process.
|
Abbreviations |
|
|
C |
consonant |
|
def |
definitive marker |
|
dur |
durative |
|
fact |
factative |
|
hum / +hum |
of human gender |
|
non-human /-human |
not of human gender |
|
o |
object |
|
pl |
plural |
|
pli |
plural agreement inanimate |
|
plii |
plural agreement human and inanimate |
|
pliii |
plural agreement inanimate for ‘days’ |
|
poss |
possessive |
|
sg |
singular; singular agreement prefix |
|
sub |
subordination |
|
tr |
transnumeral, a noun that can be used as singular and plural |
|
(sp.) |
species |
|
V |
vowel |
|
V1 / V2 |
verb 1, verb 2 etc. in a multiverb construction |
References
Comrie, Bernard and Sandra A. Thompson 2007 [1985]
‘Lexical nominalization.’ In: Timothy Shopen (ed.) Language typology and syntactic description. Volume III: Grammatical categories and the lexicon. 2nd edition, pp.334-381. Cambridge: University Press
De Wolf, Paul 1971
The noun class system of Proto-Benue-Congo. The Hague: Mouton
Givón, Talmy 1970
The Si-Luyana language. A preliminary linguistic description. Institute for social research, Communication 6. Lusaka: University of Zambia
Givón, Talmy 1971
‘Some historical changes in the noun-class system of Bantu. Their possible causes and wider implications.’ In: Kim, Chin-Wu (ed.) Papers in African linguistics, pp.33-54. Edmonton: Linguistic Research
Greenberg, Joseph H. 1968
‘Some universals of grammar with particular reference to the order of meaningful elements.’ In: Greenberg, Joseph H. (ed.) Universals of language. 2nd edition, pp.73-113. Cambridge: Massachusetts Institute of Technology
Koops, Robert 2009
A grammar of Kuteb. A Jukunoid language of East-Central Nigeria. Cologne: Köppe
Prischnegg, Tamara 2008
Das Yukuben und seine Bedeutung für die Legitimierung eines Südjukunoid. Vienna: Unpublished PhD thesis
Shimizu, Kiyoshi 1980a
Comparative Jukunoid. Vol. 1. Vienna: Afro-Pub
Shimizu, Kiyoshi 1980b
Comparative Jukunoid. Vol.2, Pt.1. Vienna: Afro-Pub
Shimizu, Kiyoshi 1980c
Comparative Jukunoid. Vol.2, Pt.2. Vienna: Afro-Pub
[1] The present paper is one of the results of the “Documentation of the Bezen language” project, generously funded by the Volkswagen Foundation. The data has been collected by the author during three field works in 2011, 2012 and 2013. (2012 in collaboration with Roland Kießling). I would like to thank the members of the Bezen community who willingly share their knowledge with us, especially Arama Fidelis, Amah Robert Shita and Kilang Martin Alhaji. Gratitude also goes to Roland Kießling and two anonymous reviewers for their instructive comments on earlier versions of this article. All remaining shortcomings are the author’s.
[2] /e/, /o/ and /u/ have the free variants /ɛ/, /ɔ/ and /ʊ/, respectively. The examples in this paper represent the phonetic realization of Bezen lexemes and utterances.
[3] This rule also applies to the prefixes with a mid- or high-tone.
[4] A larger number of gender categories in plural than in the singular seems to be typologically exceptional. Greenberg (1968: 95) states in his universal nr. 37 that “a language never has more gender categories in non-singular numbers than in the singular”. This atypical situation in Bezen is a result of the decline of its noun classes.
[5] The factative is an aspect form that denotes an action that can have taken place in the past or in the present, depending on the situation. The factative is marked by a mid- or high-tone on the verbal root and a mid- or high tone on the vocalic prefix.
[6] All three demonstrative pronouns might be used ad- and pronominally.
[7] The tone of the prefix vowel depends on modal features of the verb.
[8] The root kɨ̄b ‘old’ can be used as an adjective, taking agreement morphemes, but also as a verb, bearing the vocalic prefix and tense/aspect markers.
[9] The bird has the colour of camwood; kōhūn ‘camwood’.
[10] Givón concludes that for the development of the Bantu noun class system and states “what is now largely a system of ‘grammatical’ genders, was once a system of semantic classification of the noun universe” (1971: 34).
Lizenz
Empfohlene Zitierweise ¶
Kempf V (2013). Remnants of a noun class system in Bezen. Afrikanistik online, Vol. 2014. (urn:nbn:de:0009-10-38342)
Bitte geben Sie beim Zitieren dieses Artikels die exakte URL und das Datum Ihres letzten Besuchs bei dieser Online-Adresse an.
Volltext ¶
-
Volltext als PDF
(
Größe:
538.4 kB
)
Kommentare ¶
Es liegen noch keine Kommentare vor.
Möchten Sie Stellung zu diesem Artikel nehmen oder haben Sie Ergänzungen?
Kommentar einreichen.