20020618 TZCA Dar es Salaam

hvj ib! U') VL JLcU-) I

IN THE COURT OF APPEAL OF TANZANIA

AT MWANZA (COHAM: LUBaA UGAKINGIPALJOA.,And MROSO,J.A) CIVIL APPEAL NO. 2 OF 2002

BETWEEN MANJU SALUN MSAMBYA . APPELLANT AND TEE ATPORNEY GENERAL .......,, 1ST RESPONDENT KIFU GULANHUSSEIN KIFU 2ND RESPONDENT (Appeal from the decision of the High Court of Tanzania at Tabora) (Luke dated the 10th day of August, 2001 in Miscellaneous Civil Cause No. 6 of 2000

JUDGMENT OF THE COURT In the October, 2000 parliamentary elections, the appellant, Manju Saluni Msarnbya, contested and lost on a Chama Cha Mapinduzi (CCII) ticket in the Kigoma South Constituency. His opponents were the second respondent, Kifu Gulaxnhucsein Kifu of NCCR-Mageuzi who was the successful candidate, and Mustapha Wambali of Chadema who was a distant third. The appellant (hereinafter tith e petiti oneru) was dissatisfied with the results of the election and petitioned the High Court to avoid the same. He cited the Attorney General as first respondent in compliance with the law. The petition was based on a single ground, namely, that during the election campaign the second respondent and/,r his agents made statements with intent to exploit tribal differences in relation to the petitioner. The petitioner is a member of the Wabembe tribe while the second respondent 1 -2- is a Muha. At the conclusion of the trial, the High Court (Lukelelwa, J.) held that the petitioner had failed to establish his case and dismisse.d the petition with costs. From that decision the petitioner brought this appeal listing three grounds in these terms: That the learned High Court Judge erred in 1a.: and fact in engaging himself in conjecture on the evidence adduced and talcing what appeared to be judicial notice of matters which are not subject to judicial notice That the learned High Court

Judge erred in law and in fact in applying double standards to the witnesses who testified for the appellant as opposed to those who testified for the second respondent. That the learned High Court Judge misdirected himself on the issue of misdemeanour of wit- nesses for the petitioner. Had he properly directed himself, he should have found the petitioners witnesses to have been more credible than those who testified for the second respondent. The petitioner was the incumbent Member of Parliament for the Kigoma South Constituency for two consecutive terms before the year 2000 elections. He called a total of 21 witnesses, including him- self, eighteen of whom came to say that at divers campaign meetings the second respondent and his campaign manager, 2W2 Faida Masudi, made statements exhorting the Viaha to liberate themselves from the Wabembe, a minority tribe of Congolese origin and of indecent habits, by voting for the second respondent, their tribesman. The second .aj'3 L -3- respondent in turn called 17 witnesses, including himself, all of whom denied that any such statements were ever made. They claimed e•o respondent spoke on roads, schools, dispensaries and development generally. In a marathon 122-page judgment, the trial judge held that all the petitionert a witnesses were unwerthy of credit and all the second respondent's witnesses were credible. That alone was

sufficient to arouse curiosity and prompt an appeal. It may he useful before proceeding further to revert to sme principles pertaining to election petitions and to the jurisdiction of this Court in matters of evidence. First, the burden of proof placed on a petitioner is a heavy one: he is required to prove his allegations to the satisfaction of the court, which has been interpreted ar proof beyond reasonable doubt. In the case of M Uf /9677 E,A. 2+0, Georges, CJ said, citing Bater v. Bater /9959 2 All TB 458 : the standard of proof is one which ±nvolvds rronf to the satisfaction of the court?. In my view, these words in fact mean the same thing as satisfying the court ... one dannot be satisfied where one is in doubt. Where a reasonable doubt exists, then it is impossible to say that one is satisfied This decision and these words have been affirmed over the years and remain the standard expression of the law as it presently stands. The reason for putting the standard that high is not far to seek. liv election is the exercise of a constitutional right and the • fulfilment of an obligation by the citizenry. It is perhaps the • only occasion when the people are enabled directly to participate in ru niri' the affairs of their country. The courts, therefore, have a duty to. respect the people's conscience and not to interfere in their choice except in the most compelling

circumstances. As put in Seddy v. Sultan /970' 3 S.C.R. 4 52, In a democracy the purity and sanctity of elections, the sacrosanct and sacred nature of the electoral process, must be preserved and maintained. And the valuable verdict of the people at the polls must be given respect and candour and should not be disregarded or set at naught on vague, indefinite, frivolous or fanciful allegations or evidence which is of a shaky or prevai- eating character. The trial judge Was aware of this standard and the philosophy behind it for he similarly cited Peddy's case and correctly added: °An election petition must be construed more strictly than a plaint in a civil suit,' It is contended, however, that he misdirected himself when he set about to apply the standard to the evidence in the petition andthis Court is invited to re-evaluatethe evidence and reach its own conclusions. This leads us to another principle. As a general rule an appellate court will be slow to interfere in findings of fact by the trial court. The trial court having seen and heard the witnesses is the heat judg of questiona of fact, In Peters v, Sunday Post /958/ E.A. 424, the Court of Appeal for East Africa said: It is a strong thing for an appellate court to differ from the finding, on a question of fact, of the judge who tried the case, '/5 -5- and who has had the advantage of seeing and hearinE, the witnesses.

An appellate court has, indeed, jurisdiction to review the evidence in order to determine whether the conclusion orisinally reached upon that evidence should stand. But this is a jurisdiction which should be exercised with caution: it is not enough that the appellate court might itself have come to a different conclusion. Bearing this in mind, it follows that there must be demonstrable justification in order for an appellate court to interfere in the findings of the trial court. It will interfere where there is no evidence or sufficient evidence to support a finding, where there is misdirection or non-direction on the evidence, or generally where the trial court has proceeded on wrong principles. The three grounds of appeal, which were in fact argued together, revolve around these factors and necessitate a general approach to the-appeal b2orc Retting to the particulars, if need be. It is proposedto deal with the first ground separately, then the second and third grounds together, our initial preoccupatin being to establish whether this is a fit case for this Court to interfere in the findings of the trial court. The first ground contends that the learned trial judge engaged himself in conjecture and appeared to take judicial notice of matters otherwise not subject -to judicial notice Dr. Masumbuko Laswai who appeared for the petitioner in this Court and at the trial pointed to

several instances where the trial judge claimed that the petitioner's witnesses were recruited and To 6 tutord i nrc evidence, but without evidence to support the claim. We carefully examined the judgment of the leariled judge and ct blihcd at he e the claim r erenco to p '2, jm /3 , p 4, ' Pb15, Pw6, PW8, 0W1 1-f, PW15 and PW20, whereas PWI3 and PW21 were said to have concocted their evidence s which is not different. Conjecture is an opinion formed on alight or defective evidence or none, in other words, an opinion without proof. The question, therefore, is whether there was evidence in this case upon which to find that the listed wereeses ore recruited and tutored or concocted -their evideace The judge had a set of reasons which were Consistently cited to justify his finding. These were: not remembering the date, of the campaign meeting; not reporting the tribal campaign to anyone; if a witness reported to the petitioner, then, not being named by the petitioner in his evidence as a source of information; testifying on tribal utterances only; and, absence of corroboration. Other reasons, but not equa3P.y frequent were delayed reporting or inability to remember the date of reporting. We will examine the most recurrent reasons here but reserve the issue of corroboration for consideration under the second and third grounds. First is the question of dates. It was requiring too

much of the witnesses to remomber dates or times months after the campaign meetings and much as they did not expect that a petition would be filod or that they would be called to testify. In fact in the entire triel dates tore guesswork: neither side put in evidence the co-ordinnted campaign programme which was supposed to have been worked out by the Returning Officer under s. 51 (4) of the Elections Act, 1985 as amended by Act No, 6 of 1992. Rather surprisingly, the judge subsequently conceded the mutuality -7-- in the absence of reliable evidence on dates, and appeared not to take it against the petitioner's witnesses, when he said: It has not been clearly established on which dated the alleged campaigns were made0 There was no master programme approved by the Aeturning Officer0 I therefore hold that the dates of the election campaigns were not specifically proved on either side in this case. However, and with due respect, the sincerity of this statement has to he doubted since it conic after the petitioner's witnesses had been condemned and after the petition had been dismissed. We will come to this aEnect at the end of the judgment0 In any case, it is not the law that witnesses in election petitions should be able to recall the dates of campaign meetings unless the controversy is on the fact of the meeting. Our view, therefore, is that inability to remember dates is not

necessarily evidence of tutoring. The same applies with not reporting tribal campaigns to anyone0 The electoral law as presently constituted does not make provision for such reports. A campaign speech exploiting tribal differences is not an offence under the law, save where it amounted to inciting the commission of an offence. Ironically, though, where the witnesses testified to have reported the campaigns to the petitioner, they were still considered hired and tutored because the petitioner did not name them. A petitioner in an election petition merely introduces his case; he states the grounds upon which his case is based and calls his witnesses to prove those grounds. He is not required to name his sources of information in advance and to SP -8- specify the nature of their evidence, for to do so would be engaging in hearsay0 Finally, it was seen as indicative-of tutoring the fact that some t'itnesses testified on the tribal aspect only0 Some witnesses said that they arrived at the meetings, caught the tribal stuff, and left in disgust0 The trial judge found this rather un- usual. His view was that these witnesses had been tutored on the tribal aspect only or that they were anxious to minimise cross- examination0 Dr0 Lamwai' e answer to this was that the only issue at the trial was tribal campaigns the witnesses were called to testify on that issue and were so led by learned

counsel. There is merit in the answer0 In fact a proper reading of the evidence does not suggest that the witnesses heard nothing else, rather, their evidence was a distillation of the tribal aspect0 Intereetingl;. though, even where witnesses testified to have attended the meetings from beginning to end, it made no impression on the learned judge0 He said that was, to use his words, a certain semblance of a trend of the petitioner' a witnesses inproving their evidence during examination in chief with the progress of the caso0 The petitioner'r burden was really made heavy0 We are unable to find, in the light of the foregoing, evidence upon uhtch t. cruld have been found that the petitioner's witnesses \-!ere recruited and tutored or that they concocted their evidence0 The reasons relied upon to make the finding were speculative, tenuous and untenable in law 0 Indeed, in yet another surprise the judge subsequently puryrirted to reeile from his finding and to concede that there was no evidence but only the suspicion of tutoring0 He did so apparently because that charge could have been levelled at the second respondent's side0 He said: 00 -9- .Lfl most of the evidence adduced by the petitioner there is a suspicion of concert between the witnesses. A similar situation is inflected in the evidence of the second rerpondent and his witnesses. In fact in their written submissions the

learned counsels (skc) raise accusations and counter- accusations of tutoring their witnesses. My finding is that there was only a suspicion of that tutoring which has not been proved to the required standard by either side. As on the previous occasion 9 this purported change of heart was extremely insincere since it similarly came after the petitioner 9 s witnesses had been condemned and after the petition had been dismissed. The second limb of the first ground of arpenl is that the trial judge appeared to take judicial notice of matters which are not subject to judicial notice. An examination of the judgment truly shows that it is littered with assumat ions and presumptions at every turn, To cite all the instances would be an unnecessary exercise but it will suffice to cite two passages which are sufficiently representative. After considering the evidence from one campaign meeting 7 the judge said: The totality of the defence evidence suggest that under the normal conduct of human beinrs 9 in an election campaign attended by Watongwe 9 Waha, Wahembe, kafipe, Wngoma and Wabwari, a candidate in his normal senses cannot exhort only one tribe of Waha to vote for him. It 9 s impossible that T1,41 /he second respondent/ would have acted so recklessly in broad daylight. C 10 Again in the course of considering evidence from another campaign niecting, he said Our hunion experience,

the Tanzanian enco, shows that people had never voted on the basis of tribalism. So was the experience of Kigoma South Constituency. In the absence of an established tribalistic tradition, manifested by tribal animosity, can a c&ndidate puhiicl utter those words? Theme arosumptions and inferences characterise the entire judgment; the learned judge considered it out of the question for a candidate to campaign on the basic of tribe in a sulti-tribal setting and, to him, Tanzanians had never voted on that basis. Given this position ; the whoLe case must have appeared strange to him. Mr. Nabe±e Marando who appeared for the second respondent argued that a. 59 of the Evidence Act, 19677 which sets out matters which courts 'shall tales judicial notice of, does not purport to be exhaustive, but courts may from time to tine add thereto. He gave some reported. examples where a court mipht be prepared to infer the existence of certain facts, end referred to PJAP'S LAW OF EVIDENCE, 13th Edn.. p. 59, where it in stated that the list of facts mentioned in the equivalent a. 57 of the Indian Evidence Act, 1872, is not exhnustive Counsel for the petitioner submitted, however, that the Evidence Ace leaves no room for extending the list of matters to be judicially noticed and courts have to 'confine themselves to the four corners of s.

59. The behaviour of audiences at rallies, the reaction of some coimmunitics to tribal exhortation, etc., are matters that would require proof-. - 11 - We incline to Mr. Maarandovs submission that s. 59 does not purport to be exhaustive. The settion sets out matters whica the court shall take judicial notice of, but it does not prohibit the court from taking judicial notice of other matters which they are so directed by an other Act. These are termed presumptions of law and may be either conclusive, e.g. that a child under seven years of age is incapable of committing an offence, or rebuttabJ.o e.g. that a person not heard of for five years by those who would naturally have heard from hinhad he been alive, is dead. Section 122 of the Evidence Act also permits a court to infer the existence of any fact which it thinks likely to have happened regard being had to the common course of natural events, human conduct and public and private business, in relation to the facts of the particular case. We are of the view that the section was intended to embrace a number of facts that are not mentioned in s.

59. There are certainly mattsrs so notorious or well-known to all or recur or happen so regularly in the ordinary course of nature or business that production of evidence in proof of them becomes unnecessary, In Comwonwealth ftc v. Peninsular Etc, 1923 App Cas. 191, 210, cited in SAPKAR (supra) at P. 595, Lord Sumner said: To require that a judge should affect a cloistered aloofness from facts that every other man in court is fully aware of, and should insist on having proof on oath of what as a man of the world he knows already better than any witness can tell his, is a rule /whicV' may easily become pedantic and futile. ,/12 i2 In the light of these observations it faIls to ask whether the presumptions manifested in the passages cited from the judgment, namely, that there can be no tribal campaigns in•a nulti-trihal setting and that Tanzanians had never voted on the basis of tribe, are in the nature of human conduct, notorious, well-known to all, recurring regularly or facts which every other man in court would be aware of. The answer is certainly negative. In the first place, it would have been unusual on the part of the legislature to abuse its wisdom and enact a law prohibiting tribal campaigns if such campaigns were out of the question. It is precisely becaue the contrary is the case that the electoral law has consistently carried the prohibition. Moreover, there is

evidence thet such camua±gn.e do takt place. In the case of Lutter Nelson v. A.G. and - Ibrabim Msabaha, Civil Appeal No, 24 of 1999 9 it was found as a fact that the second respondent and his agents conducted a campaign on tribal lines in the 1995 parliamentary elections in the Kibaha Constituoncy. There was no suggestion that the Kibaha Constituency is home to a single tribe. Minally, the learned judge does not appear to have fully directed himself on the evidence jhen he asserted that there was no tribalistic tradition manifested in triba. animosity. It came from no other than the second respondents witne RW11 Yassin Meshid Shimye, that the local government election in his • ward for the year 2000 had to be postponed twice due to a controvere: over the Tabembe immigration status. In fact there was evidence Of tribal friction in a number of villages after the second respondenth speeches but the judge scoffed at it as rubbish and bizarre. Thur., therefore, we agree with learned counsel for the second respondent that the categories of matters which may be judicially noticed - 13 - is not exhausted by s. 59, we would require empirical evidence to prove that tribal campaigns are not possible in a multi-tribal netting or that Tanzanians never vote on the basis of tribe. The trial judge misdirected himself fundamentally in purporting to take judicial notice of those matters. In

sum, we uphold the first ground of appeal. The second ground of appeal is to the effect that the trial ( judge applied double standards in the assessment of the petitioner's witnesses on one.side and the second respondent's witnesses on the other. This is interrelated with the third ground which says that the learned judge misdirected himself on the issue of the demeanour of witnesses. Beginning with demeanour, we find the petitioners witnesses being branded rude, partisan, evasive, cautious, uneasy, ccnfused, draniat ic, uncouth, Swahili, unimpressive, uncultivated, obstinate, and one youfig mn was branded a hooligan and addict of narcotic drugs. None of the second respondent's witnesses earned such epithets but they were impressive, robust, eloquent, firm, composed, forceful, sincere, sober, straight forward and excellent. 'i5 find it difficult to accept this as deriving from a rational and objective assesOment of the witnesses. Common sense alone would dictate that good and bad elements are t,.be found in any camp. I We did not see and hear the witnesses testify but we would need convincing that all the petitioner's witnesses were bad and all the second reapo,idents witnesses were good. It is clear to us that different standards were, applied. 4 1: 'U - ik - There are more facts to fortify this vieo In the course of recordiny evidence, tn: trial judge appended remarks on

the I demeanour of witnesses, including witnesses for the second resondent In the course of judgment, however, he made no mention of the remarks on the letter witnesses There were remarks that RW2 and R14 were evasive and reearks that RW3, RW4, Pw8 1 RW9 1 R110 and RW1 1+ were taking long thinking ut answers, but none of this appeared in the judgment. ,R142 could not remember the names of some villagers who hosted him and the secoid respondent during the campaign, while PW3, R115 7 P118 and PW13 could not remember some important dates, but none of this was held against them Three witnesses, 1RW9 RW11 and RW12, excelled in exaggeration by claiming that they took note of everyone at the meetings but no adverse inference was drawn Apart from aspects of demeanour, there were other material factors touching on the credit f the second respondents witnesses which were glossed ove' Y:W2as conr,icted for stealjng by agent and sentenced tn six months imprisonment in 1998 but the judge constituted himself an appellate tribunal and came down hard - n primary courts which he accused of jailing a lot of innocent people on civil wrongs, 'and declared: °Courts cannot .rignore.to take notice of this nituation. . In fact RW2 did not say which court convicted him TW11 was remanded for over one year being suspected of committing murder in the - purported execution of sungusungu duties, but it

received no ,ent,ion Aport from PW2, other witnesses came forward to say that they campoigned for.the second respondent, notably RW3, RW7, R110, RW11 T1412 1 RW13 and RW17, They were, of course, free to do so, and possibly so obliged if they were members or sympathisers of the ~ opposition camp, but the point is that such witnesses could not he - 15 - expected to abandon their rnan.and to testify that he campaigned on the basis of tribe. It was necessary for the judge to revisit their role and to warn himself accordingly but he did not do so. Finally, mention should be made of the judg&s application to the betitioners witnesses racist stereotypes taken from cdloiial literature about the African mind and his little sense of veracity, The stereotypes were not applied to any of the second respondent's witnesses as if they cane of a different race. In the light of the foregoing it is evident that the trial judge did not take proper advantage of having seen and'neard the witnesses in the assessment of witnesses but was guided by considerations unknown to law. It is equally evident that double standards were applied in this exercise, denying it the impartiality expected of a .udtcial tribunal. We profose to go fuither and demonstrate that the way he went about the whole exercise of evaluating the evidence and his percertion of the law relating to election petitions may have combined

early to fashion his attitude toward the petitioners case. First, his approach in evaluating the evidence was wrong. He would first analyse the evidence of the petitioner's witnesses from one meeting, pass judgment on it, then turn to consider the evidence on the other side. In affect the petitioner's case was decided in isolation. The issue in the case was one - whether there was a tribal campaign by the second respondent. After deciding that the petitioner's witnesses were false, meaning there was no such campaign, the judge was bound to find that the second respondent's witnesses were truthful, keference to the defence was, therefore, a matter cf routine. Mr. Marando argued that that was a matter of ./16 - 16- style and, cited Nohamedv. Republio /997 T.L.R. 138, where it \ae held that every magistrate or judge has got his or her own style of cnposing a judgment, and what vitally mattered was that thc' esgential ingredients should be there, including a critical analysis of th evidence n both sides. There is no doubt about thai, but we think it gces beyond style vhen one side of the case is consideree. and virtually sealed before the other side is considered. That, a our view, is wrong. In Okethi Okale v. Republic /965/ l.A. 555, .t wan held that it is the duty of a trial judge to look at the evide:'c a whole, and fundamentally wrong to evaluate the case for one c' de in

isolation, but that no single piece of evidence should be woigh d except in rclaticn to all the rest of the evidence. That wc's c' c,atter but the observations therein apply equally in clvi matters. Mohrnsds case (supra) speaks oi critical analysis of credence on both sides which is by no cleans the same as critical analysis of evidence in isolation. It is clear, therefore, that the judge's approach to the evidence was entirely flawed. Another feature pervasive in the assessment of the petition witnesses was the requirement for corroboration. At the conclusion of analysing each witness the judge would remark that his evidenac was not corroborated. He considered corroboration a legal necessity and viewed the absence of evidence from election officers to be fatel. He had this to say while concluding his judgment: Apart from the reasons I have adduced on each issue following evaluation of evidence, a common characteristic running thc'ouc'nout the twelve sub-issues is that the petitioner's evidence and that of his ./17 C Ii. - 17 - witnesses remain uncorroborated. In the absence of express or implied admission of the tribal campaigns by the secnd respo- ndent or his agents, no reasonable tribunal can found a convict ion on uncorroborated evidence of the petitioner and his witnesses only. The failure of the petitiner to have the evidence of the Teturning Officer, or one or more of

the officials of the eturn- i5 Officer, has dealt a fatal hlw to the entire case of the petitioner. Further on he said: If those tribal campaigns really existed, he could not have failed to report the matter to the iReturning Officer or one of the officials of the Returning Officer. The petitioner's failure to report the alleged incidents is a clear indication that there were no such tribal campaigns. It is just an afterthought, so that he can again try his chance in a byelection. The judge then made critical comments about the petitioner's omission to call particularly one Masudi Juma, a Ward Education Co-ordinator, to whom PW16 claimed to have reported a tribal campaign at Sunuka village, and proceeded. to support his argument with reference to artha Wejja V. A.G.& Three Others /19827 T.L.R. 35, 57, where this Court said: We find no reports by the Supervisory Delegates concerning tribalism at the Ukonga Najumba Sita and Garezani meetings. The testimony of the witnesses for the petitioner regarding those meetings, there- fore, remain uncorrohorated. - 18 - The judge consented: That is exactly what happened in the entire case of the retitionar, and promptly dismissed the petition It is evident, therefore, that corrbration weighed decisively in his judgment and Martha We,ja a.s his trump card 0 huh perfect respect to the learnod judge, he was seriously • in error0 A petitioners

witnesses in an election petition do not require corroboratin as a matter of law and if their evidenbe is credible, it can be acted upon without corroboration0 It is apparent that the judge had the wrong law in mind fr Martha tiejja was decided on the basis of what the law was before 1992 Under the erstwhile one-party system an election campaign was organised on behalf of both candidates by the Folitical Committee of CCM. The organisation and conduct of the campaign was subject to the suporvision of supervisory delegates appointed for each constituency0 The supervisory delegates were entitled to attend every campaign meeting and were charged with the duty of reporting to the Political Cemmittse any irregularity or non-compliance in the canipaign Most of all, no election petition could have been presented on the grounds of tribalism, racialism, religion or exploitation of sex dsffercnces without the consent in writing of at least one of the supervisory delegates0 It is therefore understandable why Martha Weja was decided as it was; corroboration was then required in the -form of a written consent, but that is no more following the amendment of the Elections Act in 1992 to reflect the i'o-introduction of muiti-partism 0 Under s 51 (1) of the Act as amended election campaigns are organised by the candidate, his party or his agent0 000 /19 • . c - 19 - Apart from co—ordinating the

candidatee caTnaign programmes. in order to avoid clashes, the Returning Officer and his offic.ie ... have nothing to do with ;the orapisation and supervision .ef -cainuaj•• meetings and do not have to attot such meetings.- - This. should have. registcred.in the mind of the trial judge since the seeondresponden -t-s -meetings at Mwalizega, Igalula, I4ganza and Sunuka, -• mention. rnly.four, were uponed and presumably presided over by village o.ff±ulals f opposition parties. There is no provisi.nin the Act for reporting any irregularity, misconduct or n.n—eompliancc in a campaign meeting................ t.:, the Returning Officer or any .fficer. And if such reports were. - made the evidence thereof would be hearsay. Hence neither. the Returning Officer n.r his officials or Masudi Juma, .. whose absence........ caused so much fuss, was a necessary or useful witness. We are satisfied in the end that the second and third grounds also succeed. And in view r5f the derisin 4bn the first groundi we consider it suffisiently demonstrated that there were fundamental. misdirections in the approach t^ the evidence f the petitioner, and his witnesses, much as wrong principles peaded the entire exercise.. This is therefore a proper case where an appellate court would be justified to revisit the evidence and come to its own conclusions,.. ... and we propso to do so. . . The only substantive

issue before the High Court was •, exploitation of tribal differences by the second respondent and - his agents in their speeches. With that sort of issue, the law is without flexibility. A single speech, if proved to the .... satisfaction of the court, is sufficient ta avoid the eledtion' there being no necessity to show that the result of the election ./2o 20 was affected thereby. In Luttef Ne1dn, (supra) the Court. stated: - When, contrary to e. 108 (2)' (a) of the Act,• .... an election campaign is nulcted an the lines of tribe, race, religii or sex, the law is jealusly qualitative, not alumsily quantitative, in its- nullification test. It does not count the numberf'contravem- tins,'. for, ne co-traventiAn has the same ultimate result as ten. eontraventions. When there is 3uc1pr.4f tho'question of . ... the contraventih affecting the resWlt f the election does not arise. In the instant case, the pOtitioner brought evidcneQ from twelve .o' the campaign meetings ad.drossod by th6 s4c-ond respondent. . With the above in mind, however,, it is WnOsnary to review the evidene'Q.. from all the meetings but a purpsive selection will d0-4 It s'' 'C-- proposed to take up three meetina's withá view to illustrating the essence of the utterances complained Af, the involvement of an agent' and the use 'f double etaxidards We start with the meeting at Igalula village in .Igalul'a

'1ard 'held on or about th Septethber, 2000 The petitioner cafled two - witnesses, 'PW6 Athunani Yapun the Igalula village .CCM Chairman, and PW7 Maulidi Juma. In the evidence of Pw6 1 the 'second respondent ,,,poke to the following efrect. Waha wenzangu, mp'! Waha wenzangu tuelewe kwarnba mwaka huu nd mwaka 'wtu 'k*jLk.mbAa. " Tumetawaliwa na kabila la Wo],embe kwa muds mrefu. Kabila la Wabembetunaji.ni. watu wa J(ngô siyo wa Tan7ania. Ni ajahu Waha ./21 - 21 - gr tuiio wengi katika jimbo hili tuendelee kutawaliwana kabila chache Muha popote alipo.ahakiJcishe .anipigie mimi Muha .mwenzie. Angalierii wanakuja .watu huku. Yuko mtu aliyepandilcizwa ma Msambya. Huyu - m1nranakuja na Kaburou, naye ml Mustapha - Jambali. Mtu huyu. amepandizwa kwa lengn la kuja kunivurugia kura zangu. . Kabourou mwangalieni. Kabourou na Msambya wte ni - damu ye Kongo, wote ni wanyakunawe. .. PW6 reported the campaign to the petitioner. On the other hand PW'/ testified the second respendent stating that he decided t--centest because one Ndeka .a Muha, had been defeated in the CCM preliminary - nominations, adding: Tunietawaliwa na wageni kwa muds mrefu .... . - umefika wakati we kujikomboa kutoka kwa . wageni kutoka Kong. ,.. .i wakati muri • kuni chagua mimi kama Muha, .Muha mwenz enu, •,. ... . .. kwa kuwa sisi Waha ndie ,ralio wengi katika jimbo hili. . . • The witness continued that the secnd

respondent claimed.. that the petitioner had sabotaged a .schol project because it was to benefit the .*Jaha and promised that if elected his priority would be to repatriate the Wabembe backto Onge---The -witness was horrified because he had an Mbembe wife and similarly reported the incident - .to the petitioner when he cane campaigning at the viflage. The second respondent and his campaign manager (2W2) denied that such statements were ever made. They called RW8 Masudi Moshi who repeated the denials and stated that the second respondent. spoke on roads and educatin. - ./22 e - 22 - 0 In his evaluation of the evidence, the trial judge declared PWE unworthy of belief becauoe he could not remember the date of the meetiur and had not noted it in a diary; did not report the incident to anyone in authority, like the ward Executive Officer, but the petitioner; being CCM village chairman, did not ask the village secretary to convene a meeting of the defence and security committee to discuss the issue; and because he was not named by the petitioner as informant. The judge concluded that PW6 was recruited. Turning to PW7, he said he, too, was not worth of credit because the petitioner did not name hire as a source of information. He also clained that there were what he called inconsistencies between PJ7 and ?w6 and listed the omissions by Pw6 about the school project and about

repatriating the Wabembe to Congo. The judge found Rw8 credible because, despite his illiteracy and unreliability on dates and times, he was endowed with natural wisdom and stood his ground on basic issues in cross-examination. The judge mentioned without comment the fadt that RW8 was the second respondent's host at the village. Consideration has already been given to some of the matters raised by the judge, line inability to remember dates, omission to report to authorities, etc, and we need not repeat ourselves. It was, additionally, going beyond the requirements of the law to demand that Pw6 should have caused a defence and security committee to be convened to discuss the tribal campaign, or should have noted the incident in his diary. Incidentally, where a witness (PW3) called such a meeting, the judge did not believe him because he did not produce minutes of it. The fact that a petitioner is required to satisfy the court beyond reasonable doubt does not ./23 23 envisage proof of irrelevancies, which is what the judge's points were. He was also unable to distinguish between contradictions, inconsistancies, discrepancies and omissions, for what he termed discrepancies between £w6 and PW7 were omissions: something said by PH7 but not said by Pw6. Omissions in evidence are normal and natural. It is not possible for witnesses to recall every detail of a speech or the exact

words used but the essentials thereof. In fact omissions better indicate veracity whereas exact replication of evidence could suggest coaching. A trial judge would, of course, be entitled to reject the evidence where witnesses are materially in conflict or inconsistent: Nanyarov.P.oleSaitabau /98?7 T.L,R. 47; omissions, however, do not constitute grounds for discrediting a witness. They are largely due to lOpse of memory and the way a witness is led in giving evidence. When the testimonies of THE and PW7 are viewed objectively, they are in fact concurrent on the • essential features. When all the factors considered by the trial judge are taken away, as they should, there is left the word of Pw6 and PW7 against the word of the second respondent, PW2 and The trial judge repeatedly reminded himself that the second respondent and TW2 had an interest to serve and their evidence had to be approached with caution. He held that RW8 supplied the necessary corroboration. We are unable to agree. RW8 was the second respondent's host at Igalula and in nia better position than the campaign manager, RW2. A host endowed with wisdom, as he was said to be, is most unlikely to give evidence unfavourable to his guest. It is curious and revealing, indeed, that the learned judge merely mentioned but 9 - 24 never commented on this relationship That was not a matter which Ii an impartial tribunal

imbued with a sense of justice could afford tc gloss over. We can find no reason on the other hand for PW6 and PW7 to create a story against the second respondent on so grave a matter. We think had the judge been objective he ought to have found, as we do, that the second respondent made utterances at Igalula with intent to exploit tribal differences Next is the meeting at Nwakizega village in Ilagala Ward from which the potitioner bad two witnesses 1 •PW11 Rmmadhani Saidi and PW2 Ally Sadiki. The orme, a resident of Ujiji, was at the village on 24th September when the aecnci respondent addressed the meetings He testified that the second respondent was first introducet by his campaign manager, 2W2, in remarks to this effect: Kama mmmvyjua tumenynywa muda mrefu, ni wakati wetu mimi Waha kujikomboa. Ninachowaomha ni kuapa kura Kifu Gulamhussein akiwa, Muloa mwenzenu Huyo Mbembe, akiwa ni T3wana Msambya, hmkuna kitu chochote alichokifanya hapa ohs maendele!,. Ni lava sababu anaona hawezi kuinnya kitu cha maendeleo kwa maana kwarnba walin wengi hapa aiyo kabila lake6 Kwa hiyo naomba møeni laura Muha mwenzenu. When it came to the second respondent's turn, he spoke about eductios in these terms: Waha weazangu ninawaombeni laura, ninawaombeni mnichague mimi lava sahabu moja. Hams mnavyojua hapa elimu ice- zorota, lava sabohu aliyezorotesha elimu 6/25 Ag C 77 - 26 - referred When he

turned to P1412 1 the judge did not have much • comment but to say that it was contra-y to human nature to campaign on the basis of tribe in a mhlti-tribal setting He then listed ten omissions between PW1I and PW12 which he again termed discrepancies They are so irrelevant that it would be tautology to sot then out The learnd judge found R12 the best in the eitnoss hox and his evidence to be corroborative of the second respondent and PW2 O He mode no mention of the fact that PWI2 was a campaigner of the second There is no need to deal with the irrelevancies already disposed of earlier in the judgment0 Additionally, the failure of PW11 to recall or even to know the asses of his street leader, sheikh, CN branch chairman and the Kigoma petrol station, was another irrelevance0 It is the same story of raising the standard of proof beyond the requirements of the law. For instance PW11 had no caune to know the none of the Pusimhi 0CM.. chaman since he was not a somber of CCi1 or any political party. Other remarks by the learned judge clearly deserve no connient0 In contrast RW12 lied that he took note of every serson at the meeting, something which the trial judge belatedly refused to believe0 Furthermore •PW2 won, the second respondent's ardent campaigner. He disclosed that he campaigned at Mupobozi, •a kitongoji of Mwakizega village He did so to reciprocate the campaign which the

second respondent mounted for him in local council elections0 He minced no words on his comsitment and declared: It is true that Kifu and myself eumpaigmed for each other0 It is without question that Kifu' a cempeign was ......my eseeaign /27 The judge never warned himself about the danger of relying on such a witness, in fact never reverted to this relationship at all, The fact that PW12 was the best performer in the witness box did not affect his ability to testify falsely in favour of the second respondent for he had every reason to do so. Had he warned himself, we are unable to soy that tho learned judge e'ould necessarily have found EW1E a witness of truth. In the absence of any cause to doubt the evidence of PW11 and Phi?, we are again of the view that it ci.g;ht t'e been found that the second respondent and his agent made the statements testified to, Finally, the re was this meeting at Ilalanguru village in $ Mtegow anoti Ward from which the petitioner called ?W19 Agostino Mguha and'PW2O Philipo Ezekiel Koliinanzila. Both witnesses heard the second respondent declare thnt it was time for the Waha to liberate themselves from the hegemony of other tribes including the Wabombe. They were disgusted and left the meeting. P1, 1 19 reported the incident to the petitioner when they met at Kigoma. PW20 had had the misfortune of being unseated as S counnillor by RW13 Shigelc John

who was campaigned for by the second respondent. The denials of the second respondent were duly eupported by RW1. 71RW2 did not attend that meeting Ph11 also admitted campaigning for the second respondent at the Ilalanguru meeting. The trial judge foend PW19 a false witness bocsuse he could not remember the dote of ti:o meeting or the date he reported to the petitioner and because he was not named by, the petitioner. PW20 was similarly a false witness because he had an axe to grind against the second respondent. The judge said of him: ,/28 -28- He contested for re-election but lost to a man supported by Kifu in the campaigns. This means PW20 got an axe to gring against the second respondent. An occasion to un- seat the second respondent having occurred, 2020 could not miss that rare and golden opportunity. That is why he had decided to lie to the Court. When he turned, to 2013, however, the judge mentioned without comment the assistance he had received from the second rospondeo and never mentioned the fact that the witness cemmaignod for the second respondent. There woe mc relevance in the mattOr listed against P019 and we may nothing more. What is interesting is that when the judge Caine dc 2d20 he made good capital out of the assistance the second respondent gave to his opponent. He was no doubt entitled to take that background into account and to approach the evidence of

P020 with caution, Our quarrel is about his failure to apjji-oeeli the evidence of 2013 in similar manner. This was n man who benefited from the second reesondents assistance and reciprocated by campaigning for the second respondent in turn. His evidence had to be approached with the same caution but it was not. The double standards were the hallmark of the impugned judgment and it is amazing with what ease they were succumbed to. Speaking for ourselves 5 we think P020 was a reljahle.witness as he found corroboration in P019, while 2013 was uncorrchormted. We again hc.ve no doubt that the seoo'nd respondent made the remarks attributse tc him. ./29 m - -29- Before concludinr, mention should be made of other matters o an unusual nature already hinted neon. After reviewing the evidence the trial judge went on to dismiss the petition, but was not content to end there. He went on at some considerable length either fortify - imp' his decision or backtracking on some findings as already shown. He then dismissed the petition for the second time. As if that was not suffi ciently bizarre, he went on again with what he termed final ebservat ions ever the whole case mainly in Passing,- and lamented the absence of female witnesses to, as well as press coverage of, the campaigns in the constituency. He went on 'still wish what he now described as 'in strict obita (sic) dicta', and permitted

himself the luxury of some polemics which we need not mention. At some stage we paused and asked ourselves whether there was before the Court a judgment properly so called, but aftr considerable reflection, we decided. to eschew the greater difficult', of a negatird 'answer. As indicated earlier, we find to have been established to the required standard that the second respondent and 2W2 made speeches exploiting tribal differences. Under a. 108 (2) (a) of the Elections Act such speeches are actionable if made in relation to a candidate or the election. We similarly find to have been established that the speeches were made in relation to the petitioner who happened to belong to a tribe different from the second responc':-".:' - In enacting the provision the legislature was inspired by Tanzanian' multi-tribal character and the desire to stem divisive tendencies concomitant with that character. The provision was, 'therefore, a positive measure for the realisation of a united peaceful natio /30 50 It is also for this reason that the law is inflexible for as pointed out bef-re, a single speech made in contravention of s, 108 (2) (a) i.e.nu±:ficient to avoid an election It follows from the forugoiiag that had the trial judge properly dit'ected himself he ought to hevo found meritin the petition We anew the appeal, reverse the decision of the I-rig'n Court and declare that the election of the second respondent as Member of Parliament for the Kigoma South Constituency is null and void. The petitioner/appellant will have his costs here and below. PATTI) at DAT ES SALAAM this 18th day of June, 2002. - D. I. LUBTJVA •JTSTICE OF APPEAL

JUSTICE OF APPEAL I certify that this is a true copy of the original. ,1 (F.L.K.wMBIY

DEPUTE REGISTPAR