Even if the phonetics does not corroborate the
identification of three tones, let us suppose to deal with three. Well, we will
demonstrate that from a phonological standpoint Masa is a two tones language.
Let us follow the lexical filing by Melis (1999).
2.1.1.
MONOSYLLABLES OF CV, CVV AND CVC TYPE
2.1.1.1. the high register
It is
opposable to the two other registers independently of the nature of the initial
consonant of word:
H ¹ M
|
t½êr |
Ptilostomus afer L. |
|
t½ër |
coasts
|
|
qâm |
Feretia apodantera Del. |
|
qäm |
joke |
|
sâ |
masa game |
|
sä |
human |
|
pââ |
tobacco
|
|
pää |
turtle sp. |
H ¹ L
|
bâk |
Gyps bengalensis Gmel. |
|
bàk |
animal skin |
|
*âl |
calms |
|
*àl |
pottery |
|
lî* |
fishing basket |
|
l&* |
salty |
|
dç |
a few |
|
dô |
who already gave birth |
2.1.1.2. the medium register
It is
opposable to the two other registers, but the opposition to the low tone is
possible only on a reduced number of initial consonants: we will see their
inventory in the following chapter treating tonological problems:
M ¹ L
|
mäl |
Hyparrhenia rufa Stapf |
|
màl |
wood of support in the granary |
|
nàk |
2.nd pers. fem. sing. |
|
näk |
bird sp. |
M ¹ H
|
fül |
genius, spirit |
|
fûl |
bush |
|
höt |
Acorypha glaucopsis Walker |
|
hôt |
fast |
2.1.1.3 the low register
It is
opposable to the two other registers, with restrictions view before:
L ¹ H
|
dél |
vagina |
|
dèl |
neck |
|
gâm |
Kaya senegalensis A.Juss |
|
gàm |
Auchenoglanis occidentalis Pell. |
L ¹ M
|
m&l |
wound |
|
mïl |
genet |
2.2. THE
POLYSYLLABLES
Given the
weak percentage of trisyllables, it is practically impossible to find some
minimal pair oppositions in this category. We will limit ourselves therefore to
disyllables, to demonstrate the relevance of the three registers realized.
2.2.1 Isotonic words:
HH ¹ MM
|
t½îgâl |
fish sp. |
|
t½ïgäl |
wisdom tooth |
HH ¹ LL
|
dûwêy |
cold |
|
dùwèy |
pot |
MM ¹ LL
not attested opposition
2.2.2. Heterotonic words
One will
notice that a certain number of oppositions is impossible on the second
syllable; here we simply signal them and we will explain the reason of it in
the chapter treating tonological problems.
HH ¹ HM
|
t½égé |
francolin |
|
t½égë |
involuntarily |
HH ¹ HL impossible (HL not attested)
HH ¹ MH
|
kîlér |
|
|
kïlér |
not thick |
HH ¹ LH
|
dûwéy |
cold |
|
dùwéy |
dream |
MH ¹ MM
|
köléy |
pipe |
|
kölëy |
speed |
MH ¹ ML (ML
impossible)
LH ¹ LL
(LH = LM)
LH ¹ LM
(impossible opposition)
MM ¹ MH
|
kölëy |
speed |
|
köléy |
pipe |
MM ¹ ML (ML impossible)
MM ¹ HM
|
kûlür |
|
|
külür |
race of the hedgehog |
MM ¹ LM (not attested)
HM ¹ HH cf.
supra
HM ¹ HL (HL impossible)
LM ¹ LH (LM = LH)
LM ¹ LL
|
bçbö |
disease of the low stomach |
|
bçbç |
locust sp. |
cf.
oppositions previously given concerning the high and medium registers.
2.3. SUMMING
UP
Whereas in monosyllables the three registers can oppose each others,
even if with some difficulties concerning the L/M opposition, in polysyllables
one could have noted that a certain number of oppositions were either
impossible or not attested in the minimal pairs permitting to show their
relevance. So we give the list of diagrams attested below:
2.3.1
monosyllables
|
H M L |
fúl sä bàk |
“bush” “human” “animal skin” |
|
2.3.2.
disyllables
|
L L |
b&r&m |
“tobacco bag” |
|
L M =
L H |
b&rï |
“dry tornado” |
|
M M |
kölëy |
“speed” |
|
M H |
kïlér |
“not thick” |
|
H H |
kîlêr |
“Ficus sp.” |
|
H M |
bîrëm |
“Acacia ataxacantha” |
2.3.3.
Trisyllables
|
H H H |
t½îgîdém |
“with small mouthfuls” |
|
H H M |
t½ôlôgö |
“tisserin bird” |
|
H M M |
tâbärä |
“belt of pearls” |
|
H M H |
mâ$älâ* |
“crab” |
|
M M M |
kütürü |
“leprosy” |
|
M M H |
tï*ïrîk |
“basket” |
|
M H H |
kökûlô |
“scavenger” |
|
L L L |
bàÀàrà |
“palm leaf ” |
|
L L H |
bàgàsî |
“pearls set” |
|
L M H |
gùÀübî |
“nasse” |
|
L M M |
*&lärä |
“crevasse” |
The
diagrams L H H and L H M are not attested.
3. Tonological
problems
3.1. In the
tonal system: 2 or 3 tones
Treating distinctive units we saw that it was
possible to demonstrate, in non verbal items, the relevance of three tonals
registers - high, medium, low; however the possible oppositions, as we
mentioned, are entirely determined by the nature of the initial consonant.
Indeed on the first syllable if the high
register can mark terms having for initial any consonants of the system, it can
not oppose to the medium register except that in words having for initial,
either unvoiced occlusives and fricatives, or the glottals /x, q, /, or the nasals and the sonorants, or
semi-consonants /j/ and /w/, and to the low register except that with words
having for initial, either the voiced occlusives and fricatives, or some nasals
and the sonorants, or the glottal //.
Moreover, the medium register can oppose to the
low register on the first syllable only in words having for initial, either
nasals, or sonorants, or the glottal //.
Thus this behaviour divides the inventory of the initial consonants
(syllabic onset) of non verbal (nominal) terms in three sets that we will call
in the following of our exposition: first set, second set, third set.
FIRST SET (non depressor consonants): unvoiced
obstruents (occlusives and fricatives) or the injectives /x, q, /.
SECOND SET (depressor consonants): voiced
obstruents (occlusives and fricatives).
THIRD SET (neutral consonants): nasals, or
sonorants, or the glottal // or semi-consonants.
In the second syllable it also exists, as one
saw explaining the relevance of each of the three registers, a certain number
of incompatibilities independent of the nature of the initial consonant of the
internal or final syllable, this one, one saw it, being inevitably a voiced
consonant in internal position and an unvoiced one in final position.
If now we examine the different registers that can mark verbs we are
going to see that one finds also a set of constraints dependent of the nature
of the initial consonant. The interference between nature of preceding
consonant and tone is a common trait with Musey (Shryock, 1995), and Zime (Jungraithmayr, 1978).
3.2. THE TONAL
BEHAVIOuR IN THE VERBAL PARADIGM
In a first time, the nature of the initial
consonant will determine two classes of verbs:
- the first one regrouping verbs with consonantal initial belonging to
the first set as well as a certain number of verbs having as initial some
consonants belonging to the third set;
- the other one regrouping verbs with
consonantal initial belonging to the second set as well as a certain number of
verbs having as initial some consonants belonging to the third set.
Then on these bases, aspectual distinctions
will be marked. Indeed the Masa involves three aspects:
- neutral aspect (Ø)
- unaccomplished aspect (unac.),
- accomplished aspect (ac.),
and when the verb is followed by a pause a
vowel marked by a tone is added inevitably to the root.
3.2.1. The
tone of the neutral aspect
Two heights are attested in the verbal paradigm
of the neutral aspect: medium for the first class, low for the second class:
|
1 |
t-ä*ü
//eat(Ø)-you// |
“eat” |
|
|
f-ä*
pütü //find(Ø)-you /bovine// |
“find a bovine” |
|
|
|
|
|
2 |
d-à*ù //speak(Ø)-you// |
“speak” |
|
|
v-à*
pütü //take(Ø)-you/bovine// |
“take a bovine” |
3.2.2. The
tone of the accomplished
One finds
again in the accomplished the two classes of verbs, one marked by a medium
register, the other by a high register:
|
1 |
àn tïé // I
/ eat(ac.)// |
“I ate” |
|
|
àn fï pütü // I / find(ac.)/ bovine.// |
“I found a bovine” |
|
|
|
|
|
2 |
àn dîï // I / say(ac.) // |
“I say” |
|
|
àn vî pütü // I / take(ac.) / bovine // |
“I took a bovine” |
3.2.3 The tone
of the unaccomplished
The
opposition between the two classes of verbs is shown by a high register for the
first class and by a low register for the second class.
|
1 |
àn tîë // I / eat(unac.)// |
“I eat” |
|
|
àn fî pütü // I / find(unac.)/ bovine.// |
“I find a bovine” |
|
|
|
|
|
2 |
àn d&ë // I / say(unac.)// |
“I say” |
|
|
àn v& pütü
// I / take(unac.)/ bovine // |
“I take a bovine” |
In summary we have the following diagrams, the register in brackets
indicating the tonal register of the final vowel before pause:
|
|
class
1 |
class
2 |
|
neutral |
M
(M) |
L
(L) |
|
accomplished |
M
(H) |
H
(M) |
|
unaccomplished |
H
(M) |
L
(M) |
As we will see thereafter, these diagrams apply to all Masa verbs,
whatever the syllabic structure of their root is: CV, CVC, CVCV, CVCVC. The
trisyllabic verbs in CVCVCVC always undergo a reduction in CVCCVC entailed by
the fall of the second vowel and the tonal diagram is then the one of the
disyllabic verbs.
3.3.
INTERPRETATION
How to interpret these constraints and these incompatibilities that one
finds both in nouns and in the verbal paradigm where they not only influence
the type of the tonal register in the neutral aspect, but also the way trough
which distinctions are expressed among the three aspects: neutral,
accomplished, unaccomplished ?
3.3.1. For
terms with initial consonants of the 1st and 2nd sets
If, on the basis of the three registers attested, one establishes the
sum of all theoretically possible realizations in the disyllables one gets:
|
uniform register |
HH |
MM |
LL |
|
ascending register |
MH |
LM |
LH |
|
descending register |
HL |
HM |
ML |
However if
we compare this inventory of possibilities with the attested realizations, we
note that:
- the
successions M L and H L are not attested;
and that
- the
succession L H is a free variant of L M.
To explain
these gaps, the only possible interpretation is to postulate that the tonal
realizations obey to two rules (Melis 1999):
1 - the
existence of two tones, the marked one, with high register (H) independent of
the nature of the initial consonant, the unmarked one, whose register depends
on the nature of the initial consonant and that is realized, as we saw, medium
(M) with consonants of the first set, and some terms of the third set, low (L)
with consonants of the second set and some terms of the third set;
2 - the
impossibility in a tonal succession inside the same word, to show a difference
bigger than one register.
Applying
these two rules to the diagram of possibilities, gives:
|
|
marked |
HH |
|
uniform |
|
MM |
|
|
unmarked |
|
|
|
|
LL |
|
|
|
|
|
|
|
MH |
|
ascending register: |
unmar./mar. |
|
|
|
|
LH that
realizes as LM in application of the 2nd rule |
|
|
|
|
|
|
|
|
|
descending register: |
mar./unmar. |
HL that
realizes as HM in application of the 2nd rule |
|
|
|
|
The impossibility, that we signalled, of the ML combination
corroborates this analysis in two tones, the marked one of high register, the
unmarked one of medium or low register according to the initial consonant,
since both these two registers correspond to unmarked tones.
3.3.2. For
terms with initial consonants of the 3rd set
If all initial consonants divide exclusively between the first and the
second set, our analysis would permit to keep phonologically only two tones,
since the medium and low register corresponding to the unmarked tone are
conditioned by the nature of the initial consonant of words.
But the behaviour of terms of the third set, regrouping the nasals, the
vibrant, the liquid and the glottal, is a problem because, in this case, the three
registers are opposable.
4. The tonal
system
While leaning on the minimal pair oppositions and on remarks that we
made on the interpretation of the tonal realizations, we will keep the
following system:
a marked tone: realized as high (H, v ¥), that doesn't undergo the
influence of the nature of the initial consonant
an unmarked tone: realized, on the first syllable, according to
the nature of the initial consonant, either as medium (M, vƒ), or as low (L, v), and on the
second syllable according to the admitted combinations.
From the
tonological standpoint Masa shows two levels, as – for example – Mbara does
(Tourneux, Seignobos & Lafarge 1986).
Verifications:
We could
give 4 types of verification.
- phonetic verification: the Masa
corpus doesn't offer 3 tones.
- resemblance with Mbara (Tourneux,
Seignobos & Lafarge 1986): 2 tones
- resemblance with Musey
(Shryock, 1995), Zime (Jungraithmayr, 1978) for consonant-tone
interference
- tonogenetic role of the
consonant-tone interference: Proto-Chadic was an atonal language (Wolff)
with voicing contrast between consonants. But this contrast becomes
amplitude contrast in the Chadic languages.
To explain
the syntagmatic distribution of phonetic tones in Masa, we propose 3 new tonal
rules:
- Two
superficial tones exist (Melis, 1999): the marked one (H register)
independent from the nature of the consonant in syllabic onset; the
unmarked one, whose register depends on the consonant-tone interference: M
register after non depressor consonant (1 set, part of the 3 set); L
register after depressor consonant (2 set, part of the 3 set).
- rule
of non adjacency of two marked tones (H) ® tonal dissimilation to the
right (with variation of only one degree of tonal register).
- rule
of the propagation to the right of a tone.
Reformulation
class 1
MM: [M ] ® [ M ] (unmarked tone + rule 3)
MH: [M] ® [M] ® [MHM] ® [MH] (unmarked tone + rule 3 +
floating H + upstep)
HM: [H] ® [H] ® [HM] (marked tone + rule 3 + rule 2
& dissimilation to the right of one degree)
class 2
LL: [L ] ® [ L ] (unmarked tone + rule 3)
HM: [H] ® [H] ® [HM] (marked tone + rule 3 + rule 2
& dissimilation to the right of one degree)
LM: [L] ® [L] ® [LHL] ® [LM] (unmarked tone + rule 3 +
floating H + upstep)
As you see,
we postulate the existence
of a marked
tone (H)
of an
unmarked tone (L or M)
of an
unmarked tone + floating H (upstep)
The results of the investigation allow us to conclude that Masa
language shows only one marked tone. The other tones derived as a consequence
of the application of three tonal rules and of some restrictions relative to
the phonetic nature of the initial consonant of the word. Therefore the Masa
should be classified as a tonal accent language.
References
BARRETEAU, D., (1987)
Un essai de classification lexico-statistique des langues de la famille
tchadique parlées au Cameroun, in Langues et cultures dans le bassin du lac
Tchad, ORSTOM, Paris.
BARRETEAU, D. and
NEWMAN, P., (1978) Les Langues tchadiques, in Inventaire des études
linguistiques sur les pays d’Afrique Noire d’expression française et sur
Madagascar, Conseil International de la langue française, Paris, pp.
292-330.
CAPRILE, J.P. and
JUNGRAITHMAYR, H. (1973) Inventaire provisoire des langues ‘tchadiques’
parlées sur le territoire de la République du Tchad, «Africana
Marburgensia» 6, 2.
de Colombel, V. (1996) La
langue ouldémé. Nord-Cameroun. Paris, Documents de Linguistique Africaine.
GREENBERG,
J.H. (1950) Studies in African linguistic classification : IV. Hamito-Semitic,
Southwest. “Journal of Anthrop.” 6.
HAYWARD, R.
(1990) Omotic Language Studies, School of Oriental and African Studies,
University of London.
HEINE, B.
and NURSE, D. (eds) (2000) African Languages : An Introduction,
Cambridge, Cambridge University Press.
HOFFMAN, C.
(1971) Provisional check list of Chadic languages (Chadic Newsletter,
special issue).
Jungraithmayr, H.
(1978) The Zime dialect cluster (“Kado”, “Dari”) in Southern Chad: its
verbal aspect system. “Africa und
Übersee“ , 61: pp. 1-27.
JUNGRAITHMAYR, H.
(1981) Inventaire des langues thcadiques. in MANESSY, G. (ed.) Les langues
dans le monde ancien et moderne. 1. Les langues de l’Afrique Subsaharienne,
pp. 407-413. Paris,
CNRS.
LUKAS, J.
(1936) The linguistic sitation in the Lake Chad area en Central Africa,
“Africa” 9, 3.
LUKAS, J.
(1952) Languages of West Africa, in Handbook of African langages,
London.
MELIS, A. (1999) Description du
Masa (Tchad): Phonologie, Syntaxe et Dictionnaire Encyclopédique, PhD Thesis, Université de Tours.
NEWMAN, P.
(1977) Chadic classification and reconstructions, “Afroasiatic
Linguistics” 5, 1.
NEWMAN, P.
and MA, R. (1966) Comparative Chadic : phonology and lexicon, JAL 5, pp.
218-251.
Shryock, A.
(1995) Investigating Laryngeal Contrasts: An Acoustic Study of the
Consonants of Musey. UCLA Working Papers in Phonetics n° 89, July, pp.
1-117.
SHRYOCK, A.
(1997) The classification of the masa group of languages. in “Studies in
African Linguistics”, 26, 1, Springs. pp. 29-62.
TOURNEUX, H.
(1990) Place du Masa dans la famille Tchadique. In
MUKAROVSKY H. (ed.) Proceedings
of the Fifth International Hamito-Semitic Congress (Vienna, 1987), Vol. 1, pp. 249-260. Wein: Afro-Pub..
Tourneux, H., Seignobos, Ch. & Lafarge, F. (1986) Les Mbara et leur langue (Tchad). Paris, SELAF.
Wolff, E. (1987)
Consonant-Tone interference in Chadic and its implications for a Theory of
Tonogenesis in Afroasiatic. In
AA.VV. Langues et cultures dans le bassin du lac Tchad. Paris, ORSTOM, pp. 193-216.