Revisits
p. 59-83
Texte intégral
1Full of energy and enthusiasm on my return from Africa, I did not take long to yield to invitations here and there, and I accepted too many administrative duties. Unfortunately, goodwill alone was not enough. In Nigeria, the distance between Ibadan and my direct hierarchical superiors in Nairobi and Paris made me more or less the sole decision‑maker and executor of whatever concerned the life of the research centre. I was very disconcerted by the tardiness and cumbersomeness of the processes of decision‑making and execution in the university environment in which I was now going to work. I learnt from bitter experience the importance of habits acquired within work groups in operation for several decades, and about the force of emotional solidarity, which carries decisions. All these, in addition to the preparation of my courses, left me with little time for Zaar. I was, however, able to arrive at some results in phonology (on the tonal system) and in morphology (on the verbal classes).
Zaar Tones
2An initial phonological analysis of my corpus enabled me to take a decision leading to the simplification of the notation of the tones.
3In her works Schneeberg noted that, in a modulated tone, every non‑High tone is realised as Mid. Therefore, of the six theoretically possible combinations obtained by combining the three Zaar tones on a single syllable (LH, LM, MH and HL, HM, ML1), only two combinations are possible: MH and HM. This was confirmed in the lexicon where no contrast is found to occur involving the other combinations. This would therefore enable us to reduce the notation of the modulated tones to a simple contrast between a Falling Tone, which could be symbolized by a circumflex (â) and a Rising Tone symbolized by a chevron (Í)2.
4However, in the conjugation, there is a problem that appears to complicate the picture. In the Imperfective, the number contrast is marked in the conjugation pronoun by the Mid tone in the singular which contrasts with a Low tone in the plural:
3s. | ÿi‹á máÑgÉnì | he arrives |
3p. | ÿì‹á máÑgÉnì | they arrive |
5Some speakers contract this conjugation form into a single syllable and mark the conjugation contrast on the syllable. This results in a modulation, showing a contrast between two Rising tones: a Mid‑High and a Low‑High tone:
3s. | ÿaá máÑgÉnì | he arrives |
3p. | ÿàá máÑgÉnì | they arrive |
6In the whole of the Zaar language, it is the only context, in which there is a contrast between two Rising tones. Furthermore, up till now I have not been able to find more than one Falling tone. I therefore took the decision to maintain the contrast only for this conjugation set /aá/ vs /àá/ and to use /àá/ (on long vowels) or /Í/ (on short vowels) to mark the rising tone elsewhere else.
7From the point of view of the tonal system of the language, we are probably looking at the transition from a system with two modulated tones towards a system with three modulated tones, through the phonologization of the consequences of a morphological phenomenon. This evolution will be definitively integrated into the language when it spreads into other parts of the grammar.3
Zaar Verbal Classes
8I was also able to prepare a paper on the verbal classes in Zaar for a congress of Afro‑Asiatic Linguistics. This work on verbal classes saw the light of day through a comical turn of events that is worth recounting.
9Among the Chadic languages there are verbal classes whose working system, to state it simply, looks familiar to someone who is acquainted with the “groups” of French conjugation. Let us remember that in Zaar the conjugation prefixed to the verb is marked by a person morpheme combined with aspectual, temporal and modal markers. I call “conjugation sequences” the paradigms of these markers, numbering several tens of them. Following the conjugation sequences to which it belongs, a given verb will have a tonal pattern (TP) which will vary according to two criteria. First, it will vary contextually according to the conjugation sequence; second, it will vary lexically. This lexical variation, which is not predictable from the morphology or the phonology of the verb, is characteristic of Chadic languages and had made linguists group together in classes the verbs behaving in the same manner morphonologically in terms of conjugation. In the case of Zaar, it is only the tonal pattern of the verb that varies.
10Schneeberg (1971) had identified three of these classes. The first consists of polysyllabic verbs (essentially two‑syllable verbs), monosyllabic verbs dividing themselves between the other two. To clearly show the form of these different morphonological classes (having no semantic content), we will use here the third person singular of the Perfective.
Polysyllabic verbs (TP Mid‑High): ndará, to be good, beautiful; jisÉÑ, to know; rŒtsÉ, to decorate.
High‑tone monosyllabic verbs: súú, to like, to want; ®áá, to pull; ¨ÉÉr, to open; ›ÉÉr, to stay put.
Mid‑tone monosyllabic verbs: ßa¨, to break; ÿim, to call; lŒk, to cut off; mbii, to take.
11But while transcribing a text with Sunday, my main informant, I stumbled on the verb “to rest” which, in the Narrative, had the TP Falling‑Mid: ndáàlaa. This made me to revise the systematic study I had done of all the verbs in my lexicon and of their forms according to their conjugation sequences. The tonal patterns obtained in the Narrative led me to come to the following conclusions.
12The verbs of a High‑tone syllable remain unchanged compared to the forms highlighted by Schneeberg4:
syllables | Perfective | TP | Narrative | TP | meaning |
1 | ®áá | H | ®áá | H | to pull |
13This class of verbs with the TP H constitutes a totally separate class, which is not large. The phenomena of variation are found in the other verbs, mono‑ and disyllabic together, which may therefore be designated as ‑H.
One‑syllable verbs: The class with Mid tone in the Perfective appears either with a Falling (ÿáà, to put; ßôk , to peel) or a Low TP: dùùm, to ruminate; vàà, to be hot; ›òp, to stab; bwàà, to choose.
Two‑syllable verbs, which have the TP Mid‑High in the Perfective fall into two classes in the Narrative: (i) the class with the TP Low‑High: ©Èlár, to halt; ©ì‹Ét, to shake; ©ÈrndÉ, to snore; vÈrtÉ, to fear; (ii) the class with the TP Falling‑High: ßûmtsÉ, to equalize; ndwáàtsÉ, to grow old; máàra, to slap; túùrá, to push.
14So, in the Narrative we have five different tonal classes: three classes of verbs with two syllables and two classes of verbs with one syllable.
15A quick survey of the classes thus constituted led me to conclude that all verbs with an initial low tone began with a voiced consonant. I put this against a remark by Schneeberg with regard to verbal nouns, concerning the phenomenon of the conditioning environment of the low tone by “non-implosive voiced obstruants” (from now on +V), and of the high tone by the other segments (‑V) (Schneeberg 1974: 139)5. Therefore, the two new classes were phonological variants of the two classes with the TP ‑H (that is to say Mid and Mid‑High) discovered by Schneeberg. In the following table, the five tonal patterns of the Narrative will be compared with those of the Perfective, for which only three tonal patterns are found to exist.
syllables | characteristics | Perfective | TP | Narrative | TP | meaning |
1 | unchanged | ®áá | H | ®áá | H | to pull |
1 | +V | bwaa | M | bwàà | L | to choose |
1 | ‑V | ÿaa | M | ÿáà | F | to put |
2 | +V | ©i‹Ét | H | ©ì‹Ét | LH | to shake |
2 | ‑V | ßumtsÉ | MH | ßûmtsÉ | FH | to equalize |
16But what about the verb Ñgáàlaa, to rest, which had put a bug in my ear and whose tonal pattern was always so unique? Another fact set it aside from the rest of the words in the language: its final vowel was abnormally long for a word of two syllables. Only some badly digested loan‑words e.g. àmmáá, but, borrowed from Hausa, had this type of final vowel. But this had nothing to do with borrowing. For a long time the mystery was total until the evidence flashed before my eyes: I had to deal not with one word but two: Ñgaa, to take, and laa, work. “to rest” is therefore expressed in Zaar as “to take work”6. Once again, the structure of the working language (in this case English, French, and even Hausa) had prevented me from seeing the uniqueness of the described language. But the irony of the situation is that this persistent error had led me to be more accurate in my description of the verbal classes.
17Pushed by the logic of the phonology of the verbal classes, I ended up going over again the work I had done on eight verbs I had isolated for posing problems of interpretation of their tonal forms. It had to do with verbs beginning with a consonant ‑V, and ending with a short vowel (•u, to beat; nda, to enter; fu, to say; ¨Œ, to leave; ¨ja, to drink; su, to return; ta, to climb; tu, to meet, to find). In the Perfective, I heard a mid tone for all the verbs. It had nothing therefore to do with verbs with a lexical high tone. Consequently they all belonged to the lexical and morphonological class ‑H, ‑V. What form do they take in the third person singular and plural of the Narrative? The shortness of the vowel made the perception of the tone dicey, even after making my informants to whistle as slowly as possible. Most of the time, however, I heard not a mid tone but a high, sometimes with a rippling that I was tempted to interpret as a falling tone. So, high or falling tone? This was a pertinent question because it was possible that the falling modulation on a light syllable is systematically realized as high, in which case certain high tones could, in different contexts, be hidden falling tones.
18Could I find assistance from parallels in other parts of the language? No noun in Zaar has this syllabic structure. The only grammatical morpheme bearing the same uncertainty as to whether it is a high tone or a falling tone was the Remote Past morpheme (ta). Its tonal behaviour is identical with that of the Recent Past (naa), which in certain contexts takes the form náà, with a clearly perceptible falling tone because it is carried by a long vowel. The parallel between the two past tense morphemes strongly suggests the existence of such a falling tone over the Remote Past of ta, that is, over a monosyllable ending in a short vowel. Is this tone “simplified” under the form of a high tone or did it keep its form, which meant that I could not hear it clearly? I could not resolve this problem.
19Returning to Paris, I appealed to the kindness and expertise of Annie Rialland of the Laboratory of Acoustic Phonetics of Paris III. But the complexity of the methods suggested to me, apart from the poor technical quality of the recordings I had at my disposal, made me to suspend finding a solution to this problem, until I discovered the tool for acoustic analysis of the SIL7, called CECIL. This tool consists of a blue box which can be linked to a portable computer, and which digitalizes the recordings in combination with an analysis of the sound signal, producing a representation of the intensity as well as of the frequency of the sounds, broken down into formats and melodic curbs. Thanks to a cheap equipment that is not too difficult to carry, I saw here the possibility, of returning to the field with a portable laboratory of acoustic phonetics. This is what the URA 1024, to which I now belonged, enabled me to do in November 1993.
The November 1993 Research Project
20This research project was to cover three facets of my work: (1) the phonological investigation, which I carried out at Azare with Eski and Sunday while staying at the State Hotel, to which I had now become accustomed; (2) the co‑ordination of the compilation of the French‑Hausa dictionary which I had launched before my departure from Nigeria; (3) a video coverage of the harvest festival in Sunday’s village.
21I devoted about ten days to each of these projects. I started with Hausa, at Zaria, in a hotel where CREDU, renamed IFRA (French Institute for Research in Africa), had brought together four contributors and two editors, Ahmed Amfani and myself, in a kind of conclave to work continuously on the compilation of the dictionary. Ahmed Amfani, who came up with this idea, saw this as the only way to drag his lecturer‑colleagues away from their daily concerns and to allow them to devote themselves at last to this dictionary. The idea was good: they did in ten days what they had not succeeded in doing in a year.
22The last time I was at Sunday’s home, in the village of Tudun Wada, I had been told of the harvest festival (LŒpm Zaar, literally the Zaar moon), which was held every year in November. I therefore came armed with a tape‑recorder and a video camera to record this cultural event. Without taking an explicitly ethnographic view, I wanted to collect as much information as possible about the Zaar cultural situation, which might one day turn out to be useful, if not to me, at least to colleagues who might be interested in such documentation. On arriving at Tudun Wada, I was told that the festival had ended two weeks earlier. Seeing my disappointment, Sunday proposed to me to arrange for an orchestra to come and perform to give me at least an idea of what the festival was like. I accepted the proposal.
The Tudun Wada Festival
23As the preparations were being made, I came to the realization that to arrange for an orchestra to perform was not an easy matter. It meant that a meal and millet beer had to be served to about ten people. I therefore had to buy the food items and the ingredients for the preparation of the beer, and I also had to find the people to cook the food and brew the beer. If to arrange for an orchestra entailed so much work, why not invite several at the same time?
24So, little by little, a real festival was organized. Nine different groups came to perform, one after the other, from 2 o'clock until the sun went down around 6:30 P.M. Two villages came together to dance and sing in the presence of the King of Tudun Wada, who wore a white lace turban and sat on an arm‑chair under a tree.8 The singing, dancing and pouring of libations went on almost everywhere in the village until very late into the night, as each orchestra continued to play in its own corner once the recording had been done. As for me, at about 7 o'clock, after eating a little rice and roast chicken, I slumped on my bed, exhausted after being on my feet all afternoon in the heat of the sun, holding a camera in my hands and trying very hard to film in the chaos. The requirements for the production of a good video recording are hardly compatible with the tradition of dancers who normally dance clustering together in a circle around the orchestra and raising clouds of dust in the process. Three masquerades covered from head to toe with hemp fibres danced and entertained the crowd with their antics.9 From time to time, the king stood up and, brandishing a leafy branch, tried to push the crowd back in order to make space for me.
25In the end, I had a video tape of three hours. Given my lack of experience, it was of an acceptable quality. It showed all kinds of orchestras: the horn‑orchestra of the king of the neighbouring village; an orchestra of women dancing and singing to the sound of a tambourine while shaking some sort of maracas; an orchestra of tambourine and flutes. I particularly liked one group: Two singers performed accompanied by extraordinary instruments. One singer, wearing an incredible chapka made from artificial fur, was playing a guitar made of wood and metal recycled from tin cans; and the other, with a guitar of reeds in each hand, hop‑danced, at the same time singing at the top of his voice with a nasal twang. The singer with the chapka was a giant of a man, enveloped in a large black coat, his cheeks inflated and covered with a long beard. The other singer, the musicians and the dancers wore bottle‑green uniforms, their waists tied with a white band, and on their feet were moulded plastic sandals that one finds in all the local small markets. Two dancers, a father and his ten-year-old son, hopped about wildly in a frenzied tap‑dance, each of their ankles heavy with a bundle of metal bells. They dared the notables of the village to encourage them by pasting naira bills on the face of the young boy.
26Later, on a tape‑recorder, I recorded some songs which women used to sing in the past while they were milling grain on millstones with their hands. These songs are no longer known except by a handful of old women, the din of the diesel engine used by mills having replaced the polyphony of feminine voices. Moreover, the festival, the dances and the masquerades no longer exist today except in the form of folklore, severed as they are from the cultural and religious meanings they used to have. In about thirty years, the Christian religion has stripped from their memory anything that in any way reminds them of the pagan past of the Zaar people.
27Sunday, who was the organizer of the festival, saw his social status move up by a peg. As for me, I had created for myself a reputation as a patron which, if it enabled me to occupy a place in the village, engendered expectations which were bound to lead to disappointment. The village people now saw in me someone who was definitely interested in raising the status of their language as well as of their culture. So much so that I had afterwards to state clearly that I was not going to come every year to organise the same festival. It was, unfortunately, what the king expected from me. Two years later, a young lady, seeing me shopping in Tafawa Balewa market, exclaimed: the European has come back; we are going to dance!
28People in the village were nonetheless beginning to show signs of impatience at not seeing concrete results from my work. My informants had to provide answers to questions that were becoming more and more pressing: How far had that work gone, that work which was so serious, so important, and which lasted so many years? When would they be able to read in their language? My problem was to make them understand that the practical fallouts from my work, if there were going to be any, did not depend on me directly. The description which I was doing was no more than the first step towards the elaboration at a later date of an alphabet that would enable the language to be written in a simple and economical way. I had proposed to several educated Zaar men to work with me to produce an orthography. I had put them in touch with the Bible Translation Bureau at Jos so that they could work towards the production of a Bible in Zaar, a project that was so dear to their heart. But this implied devoting fifteen years of their life to work that was not remunerated, or something just as demanding, and they all said they couldn’t do it.
A little acoustic phonetics
29Far away from Tudun Wada, at Azare, I tied myself down to phonetics works. Thanks to CECIL, the first stage was to associate the vowels and the sonorant consonants of the utterance with the melodic curve. To the unvoiced consonants correspond an interruption of the curve produced by CECIL. Only a voice associated with a cyclic sonorant sound makes the software to calculate a frequency measured in hertz, which, in relation to time, represents the melodic curve of the utterance.
30I was quickly able to lay out the levels of the three tones in Zaar: at the beginning of the utterance, H = 120 Hz; M = 110 Hz; L = 95 Hz. These levels are relative, as a result of the downdrift phenomenon which lowers the general level of tones from the beginning to the end of the utterance, so that an H tone at the end of the utterance may be realised lower than an L tone at the beginning of the utterance.
31The falling tone (F) is marked by a significant lowering of about 30 to 35 Hz, beginning at a level higher than an H tone, that is, for an F tone at the beginning of the utterance, an attack starting at about 135 Hz.
32The rising tone (R), which is also clearly perceptible to the ear, particularly when it is whistled, is indicated on the graphs by a relatively small deviation in the form of a rise not exceeding 15 Hz. Sometimes, it shows only by a simple blockage of the downdrift movement, in the form of the maintenance of the register.
33Let us recall the problems I was trying to resolve: (1) to work on the perceptions of my ear, and to verify my transcriptions; (2) to identify the tones realised on final short vowels (H or F?). The results were partially disappointing.
34I had first to eliminate certain reading errors caused by the context. The end of the utterance is marked by a sudden fall of half a tone, and every M tone is realised there as L. It was therefore necessary to find a non‑final context.
35Schneeberg had proposed to present all the verbs conjugated in the negative form by having them followed by the negative morpheme, which she wrote as hÑ with a Mid‑High rising tone. First, this negative morpheme was systematically whistled H by my informants, without any modulation. On the other hand, it changed all H tones into F tone, something that I became aware of only very late.
36I therefore started looking for another context, which might enable me to perceive clearly what precedes it. I found it in the form of the direct object personal pronoun (M tone10), followed itself by the negative form to avoid its being lowered at the end of an utterance. Once the context had been found, the melodic curve corresponding to the short vowels turned out to be of little significance. On most of the recordings, because of the shortness, this curve came down to a slight descent of about 5 to 10 Hz maximum, which could as well be attributed to the environment of the plosive consonants. The only common characteristic between the “classic” F tone and this realisation was the very high starting point (135 to 130 Hz). Therefore, the only acoustic argument in favour of the interpretation as an F tone rather than as an H tone was linked not to the variation of the height, but to the height from which it fell. This, added to the morphological argument already mentioned (that is, the parallel between the two Past morphemes naa and ta) made me to tend towards the marking of an F tone in the context. Moreover, I eventually succeeded in getting some “good” recordings, which provided confirmation of my conclusion. Thus, in the following recordings, where the monosyllabic verb ta (to climb) appears, the contrast is clear between the high tone of ta (a Past marker) in the first recording (cf. Fig. 1):
mŒ tá ÿô‹Ñ gwààs: I climbed up the tree
and the falling tone in the second (cf. Fig. 2):
á tâ ÿô‹Ñ gwààs: he climbed up the tree
Figure 1

Figure 2

The Research Project of Christmas 1995
37In December 1995, the URA of INALCO (National Institute for Languages and Cultures Overseas), which had become UMR by merging with a part of LACITO, provided me with the funding for a second research trip to Tafawa Balewa to investigate this time the dialectical differences within the Zaar‑speaking area. I spent the whole period of the research trip in Tudun Wada, Sunday’s village, during the dry season. Sunday put me up in the two completed rooms in the new house, which he was building. I enjoyed the profound tranquillity of this stay, surrounded by the affectionate and teasing gentleness of Sunday’s family. I felt perfectly at ease, relaxed, living in rhythm with the rising and setting of the sun. I spent Christmas in the village. A goat and a few dogs were killed to make ordinary life better. This gave occasion to excessive eating of roast meat and drinking of millet beer.
38During the day, I walked the roads across the hills to fill my questionnaires in the surrounding villages. We set out on two motor‑cycles, with me sitting behind Sunday, the second motor‑cycle driven by Michael, the mechanic accredited by Sunday and Ali their own friend. This luxury was not a waste because each motor‑cycle developed a fault at least three times a day, and Michael, with an ingenuity that is common in Africa, repaired the machines immediately using spare parts which he had brought along for the trip.
39The people in the first village we visited were very suspicious indeed. It took us more than an hour to reassure the king and to convince him that our research had nothing to do with taxes or the secret police. Moreover, the on‑going judicial investigations into the 1991 riots justified the people’s suspicion. The reputation enjoyed by Sunday’s father, the first (but now retired) motor-cycle repairer in the region, served us as a rite of passage, and we succeeded in persuading the people to answer our questions. Better still, once our working session had ended, they suggested that they announce our impending visit to the next village. Indeed, the next day on arriving at the second village we found the council of elders already assembled. They were waiting for us, having chosen one among them to answer our questions. That is how the research was conducted from one day to another in the six villages we went to. Frequently, the working session ended with a cup of kunu (a drink made from millet flour), sometimes with a meal, which gave us the energy to take off again and face the roads leading back to Tudun Wada.
40Among the villages visited I included Jigi where the Gus language is spoken. This language had excited my interest during my first visit to Azare. Research on the languages of the region (Shimizu, 1975) mention three dialects as far as Zaar is concerned, each one named after the largest village where it is spoken: Lusa, Lere and Kal. Sunday and the inhabitants of Tudun Wada and the region extending as far as Tafawa Balewa speak the Lusa variety, by far the most widely spoken. Gus (called Sigidi, from the name of the main village where it is spoken) is mentioned as a neighbouring language different from Zaar. But Sunday does not understand the Zaar spoken in Kal, a village that is nearer to Jigi than to Tudun Wada. He needed an interpreter to be able to communicate with the inhabitants of Kal, whereas there is mutual intelligibility between the Zaar of Kal and the Gus of Jigi. An inhabitant of Jigi even told me that in the past the inhabitants of Kal called themselves Gus, and that it was only recently that they were integrated with the Zaar, as they were attracted by their greater numbers and their economic success. The big towns on which the Gus depend for their economic activities are Gindiri (a village dominated by the Piem ethnic group), and, beyond that, Jos, the capital of Plateau State. Those on which the Zaar depend are Tafawa Balewa and, beyond that, Bauchi, the capital of the state of the same name. Their history is also slightly different. If they all regard themselves as coming originally from Bornu and from the banks of Lake Chad, the Gus had their ties in the west through Zaria before settling down definitively in the north‑east corner of Plateau State. It is therefore obvious that ethnic and linguistic identity is a very complex matter, which has to take several factors into consideration.
41Linguistically, the situation is quite interesting. At first sight the basic lexical corpus is common between Gus and Zaar. As far as the grammatical structure is concerned, they are also obviously related: the same organisation of the aspecto‑modal system, the same syntactic structure, the same system of nominal classification (with nevertheless more developed plurals at Jigi), and the same system of verbal derivation. But there is an important difference: the morphological support of these systems is radically different. For example, if the conjugation system is the same (the same tense, aspectual and modal system, expressing the same semantic distinctions between three different pasts, etc.), the morphemes which signal these distinctions are different and do not seem to be etymologically related. This explains the lack of mutual intelligibility between Zaar and Gus. The tonal system of Gus seems to be simpler than that of Zaar, with fewer modulated tones and an absence of tonal verbal classes.
42And so we are once again, and in a pointed manner, faced with the problem of the genetic classification and of the identity of the languages and dialects of this region of Africa, at the cross‑roads of social factors (political, economic), historical as well as linguistic. The idea of a linguistic family reconstructed according to the genealogical tree model seems to be more and more difficult to apply here.
The July 1996 Research Project
43It was the third (and probably last) research trip of two months in summer of 1996 that enabled me to formulate the ideas on the relationship between Zaar and Gus, ideas which the simple vocabulary questionnaires completed during my first research trip in December 1995 did not lead to.
44I was again able to contact Ibrahim Jigi, a Gus whom I met in the village of Jigi in December 1995 and who had given me his address then. We worked together for about a week. This may appear to be a short period, but thanks to my knowledge of Zaar and the many traits it shares with Gus, I was able to figure out without too much difficulty the skeleton of the latter language.
45With regard to Zaar, I had settled the main contentious issues relating to its morphology and phonology. This long trip, made possible by the hospitality of the Martinis who welcomed my informants and myself into their large house at Jos, was devoted to the revision of the texts which I had put together at Ibadan, and to the transcription of newly recorded texts. It was a comfortable well organized trip, the informants came one after the other to work in my study. Efficient work, carried out exactly according to schedule; but how boring! Locked up in the study, with no other landscape but the showers of water of the seasonal rains which were beating down on Jos without ceasing, I missed the sun and the big spaces of the Tudun Wada savannah; I missed Azare; I missed the visits to the breweries of millet beer at ªªoi; but above all I missed the warmth and affection of Sunday’s family.
46At the end of the trip, I now have in written form the six songs of Musa Zamba which I had collected during my visit to Azare; I have five Zaar tales recorded at Bauchi while we sat around calabashes of millet beer; I also have two extracts of conversations that went on by fits and starts and which were recorded at Tudun Wada in December 1995.
And now?
47Here I am at the door‑step of a new stage in my work. After the field research, the perpetual and sudden going back and forth between the recording of the data and the intuition of their structural pattern, here now comes the time for the slow work of analysis as I sit alone in front of a piece of paper, before the computer screen, or confronting a circle of colleagues whom I still have to convince about the value of my discoveries.
48I know that the task of elaborating and writing up, the time‑consuming, sometimes tedious, but necessary task, is rendered even more difficult by being done far away from all that sustained me during my research trips: the kindness of Sunday and his family, but also the landscapes of the African savannah. I have left a part of myself at Tudun Wada, and time scarcely passes without my recalling its vegetation. Rarely have I felt the serenity which comes upon me, sitting on a tree trunk, as the dusk falls on the tall trees of this region.
49The locust tree is reassuring as it punctuates the landscape of the Zaar plain with its imposing majesty. It extends afar its protective and gentle shade, which shades off the sun just sufficiently to cool down its ferocity without keeping out its gleam. It inspires the same feeling of abundance, of continuity of life as the luxuriant oak trees which are found in the English countryside.
50The baobab is unsettling, almost frightening. It is bloated, twisted, possessed. One cannot say what torment, what suffering is responsible for inflicting on its arms these incredible, grotesque postures. It just about achieves recognition as a plant. The old baobab trees, their grey bark suffused with rose, balance between the mineral and the animal. It is certain that the ghosts of the bush have chosen to remain there. If not, how does one explain their strangeness?
51Imagine a country spared of written symbols. When one is coming from the cities, writing is slyly propagated along the routes to our countryside: bill‑boards, signposts, road signs. Our eyes are constantly directed, our attention captured, our mind solicited by these pointless signals. Oh for the purity of a countryside where the human presence is confounded with natural harmony: the ochre horizontality of the earth, the yellow colour of the straw. The locust trees underline these vast expanses with their widely extended dark‑green leaves. Above them, in the distance, the dark‑blue mountains undulate. Here and there, the vertical leap of a kapok tree. Scarcely visible under the locust trees, are the brownish spots of small houses covered with thatch. And suddenly, piercing through the red sky, comes the white flight of a troop of egrets settling down for the night on a giant kapok tree.
52Little by little the sun goes down, the last glimmers of the setting sun go off, and with the morning star, the moon takes possession of the countryside. The glow is discrete, so much so that it is no longer seen in our cities where the night gleams with so many electric lights. There in the countryside, people are surprised to see their shadows cast by the moon on the ground, the white sand on the roads marked out clearly in the night, the silhouettes of the trees showing off their majesty even more than during the day when the sun crushes them under its implacable white light. In the night, the sounds seem amplified; the limited view is relayed by the ear. One hears from afar the rumbling of a diesel engine, a spirited conversation, a drum beating. Cheered up from drinking their traditional beer, neighbours intone a few songs. There is calm and tranquillity, a feeling of the infinite under the silence of the stars.
Notes de bas de page
1 The tonal patterns (TP) are symbolised by the following letters: H (High), M (Mid), L (Low), F (Falling), R (Rising).
2 This problem of tone markings goes beyond the simple question of orthography. In fact, the orthography will probably not mark all the tonal contrasts but, from the point of view of economy, only certain crucial contrasts. This concerns a classic phonological problem which touches on the tonal system of the language. The simplification which I was seeking to achieve by applying Occam’s principle aims at taking into account the economy of the language itself.
3 I thank my Director and colleague of LLACAN, Mme France Cloarec‑Heiss for her contribution towards this reflection. [However, another solution seems to emerge among the younger Zaar generations: the MH modulation is reduced to M, thus keeping the original contrast as an opposition between M and R, maintaining a simpler version of the tone system. Added 2002]
4 The forms of the Perfective and the Narrative are given here to highlight the differences between the forms cited by Schneeberg and those cited by me.
5 In the text originally published in French, I went further into my analysis:
“The polysyllabic verbs with the TP Low‑High all begin with a +V consonant. But for the other verbs, a possible conditioning is yet to be found to account for the distribution of the TP between Mid‑High and Falling‑High. The criterion is given by the weight of the first syllable. The verbs with the TP Falling‑High all have a heavy first syllable (CVC: ñâÑgás, to refuse; ßûmtsÉ, to equalise; or CVV: túùrá, to push), whereas the verbs with the TP Mid‑High have a light first syllable (CV: rŒtsÉ, to decorate; tuná, to recall; sulár, to return (tr.); mbŒtsÉ, to put out).”
This was a mistake due to the fact that I didn’t hear properly Falling tones on CV syllables (cf. infra, the passage on acoustic phonetics), and heard Mid tones instead. This led me to inventing an ad-hoc rule to account for a non-existent Mid‑High verb class in the Narrative. Instead, all initial syllables in these disyllabic verbs, both heavy and light, take a Falling tone. [note added in 2002]
6 This etymology was not volunteered by my informants, and it remains unexplained.
7 The SIL, or Summer Institute of Linguistics, also known under the name of International Society of Linguistics, is an association of evangelists whose aim is to translate the bible into all the languages of the world. In order to do this, they have produced special computer devices for describing and publishing non‑European languages, which they distribute to all interested linguists at prices defying all competition.
8 You should know that the Zaar, in contrast to the Hausa, do not wear the turban. This turban, or chieftaincy, is conferred on the villagers of the region by the Emir of Bauchi, a Hausa. In a period of violent revolt against the Hausas, to see a Zaar chief turbaned in this way is rather paradoxical.
9 The masquerades which have lost all relevance in the wake of the disappearance of their traditional practices came out only to add a folkloric colour to the event. Women were allowed to joke with them, whereas, in the past, no woman or uninitiated child could confront them except under pain of receiving an immediate violent punishment. As soon as they heard the noise of a masquerade cry, they were to rush and hide right inside the houses.
10 I discovered also later that the Mid tone is realised Lo after a verb ending in a Lo tone. This was masked by the small difference separating the Mid from the Lo tone and by the downdrift phenomenon. But this lowering had nothing to do with the problem in hand.

Le texte seul est utilisable sous licence Creative Commons - Attribution - Partage dans les Mêmes Conditions 4.0 International - CC BY-SA 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.
Universal Declaration of Human Rights: English, French, Hausa, Igbo and Yoruba
Foreword by Wole Soyinka
1993
Creative writing, writers and publishing in Northern Nigeria
Abba Aliyu Sani, Jibrin Ibrahim et Emmanuel Babatunde Omobowale
1997
Archaeological investigation of Itagunmodi potsherd pavement site, Ijesaland, Osun State, Nigeria
1991-1992 season
Babatunde Agbaje-Williams
1995
Guerre et société au royaume de Fetu
Des débuts du commerce atlantique à la constitution de la fédération fanti (Ghana, Côte de l’or, 1471-1720)
Yann Deffontaine
1993