1Most of the recently published works on witchcraft in the legal arena in Africa have dealt with focused on the state’s criminalization of witchcraft beliefs, practices and accusations or with its related efforts to prove and prosecute the inherently invisible crimes of witchcraft within the bounds of a squarely evidentiary-based legal system1. Yet, despite its fruitful treatment of the intersections of supernatural situations and state legal systems, such scholarship elides another central space in which notions of “customs” and “crimes” collide–murder cases in which witchcraft is posed not as the means but instead as the motive for the murder at hand2.
2In such cases, the alleged “witch” is not on trial for witchcraft or for another crime committed through witchcraft, but is instead the decedent in the murder being tried. The defendant, in turn, is on trial not for a murder committed through witchcraft but rather for a murder motivated by the witchcraft turned against him or her by the decedent. And, in cases in which witchcraft is the motive rather than means of a murder, witchcraft operates variously as a claims-making strategy of the defense and as a means for the courts to consider further standards of “reasonableness” in local contexts.
3Defense claims about the decedent’s witchcraft work variously to mitigate the culpability of the accused, generally pointing to competing institutions of “local” justice or suggesting the defendant’s diminished capacity. Court counter-claims about the decedent’s witchcraft offer a space for the courts to assess both the legal fitness of various trial participants and the authority of the law-as-written in the context of witchcraft-related murder cases. Yet at the same time, defense claims and the courts’ counter-claims about the witchcraft of the decedent each operate on a broader epistemological level, calling into question not only essential categories like “victim” and even “murder”, but also key legal concepts such as “provocation” and “malice aforethought”.
4This article traces the legal genealogies of witchcraft claims and counter-claims within the legal arena of colonial Kenya. Attending briefly to the broader politics of knowledge production in the empire, it addresses how witchcraft emerged as a central colonial concern. This article also incorporates anti-witchcraft legislation and relevant sections of the Kenya Penal Code with reports on witch-killing cases contained in the digests of the Supreme Court of Kenya and of the High Court of Appeal for Eastern Africa in order to foreground how witchcraft claims and counter-claims made in East Africa’s highest courts contributed to the refinement of legal meanings of witchcraft and to the elaboration of legal concepts central to the prosecution of capital crimes. Finally, it points to significant continuities in colonial and post-colonial witch-killings and anti-witchcraft legislation in Kenya.
5The broad array and regular appearance of witch-killing cases in the digests of East Africa’s highest courts begs the question, “How did witchcraft enter the colonial record and become an object of official concern”? (Amin 1995). Witchcraft emerged as a locus of official disquiet via the systemized and professionalized production of anthro-administrative knowledge throughout the colonial era3. A primary goal of the archived and archiving colonial state was to render subject peoples more easily governable by making their unfamiliar cultures, customs, and institutions known quantities. In turn, a key component of an administrator’s role was to make “knowable” the formerly “unknown” people under his control; to collect, organize, report, and circulate anthro-administrative data4. Administrators produced articles and monographs based on anthro-administrative data and these texts generally evidenced a deep interest with important aspects of local cosmologies–in particular, witchcraft beliefs and practices5.
6Anthro-administrative knowledge was circulated through topical, imperial networks of varying scale throughout the colonial era through “circuits of empire” (Cooper & Stoler 1989).
7Such knowledge was passed from administrators on-the-ground to authorities in other branches of the colonial government (like the courts) situated in the capitol and on to Colonial Office officials in the metropole. Colonial Office authorities archived anthro-administrative knowledge in a complex classification system that enabled it to be accessed and (re)circulated by officials throughout the empire. Further, anthro-administrative knowledge also passed through tertiary circuits such as the academy, organs of civil society, and the press.
8The colonial bureaucratic reportage that flowed through imperial networks of knowledge was permeated with anthro-administrative discourse about witchcraft, and such discourse figured portrayed witchcraft as a central challenge to state authority from a variety of angles. In this discourse, “witchcraft” primarily worked as a descriptor and as an analytic. It described states of supernatural insecurity in the colonies and patterns of local mentalities, and it analyzed how these factors impeded administration. Indeed most colonial officials in Kenya (and in the metropole) would likely have registered an annual report’s contention that “witchcraft” was “the most serious handicap to Administration” in the Ukambani region of the colony as neither hyperbolic nor unexpected6.
9In the legal arena more specifically, anthro-administrative knowledge about witchcraft functioned in a number of key ways. First, it was the initial avenue through which witchcraft entered the courts and became an object of official juridical concern. Administrators on-the-ground were responsible for bringing to the attention of the high courts the dead and damaged bodies of “witches” which so often appeared in the bush and which the state could not ignore. These bodies, and the circumstances surrounding them, brought into high relief how the colonial monopoly on the exercise of spectacular violence was something on an imperial fantasy.
10In turn, anthro-administrative understandings about witchcraft–together with related colonial concerns about witch-killings and the broader colonial vulnerabilities they exposed–formed the backdrop to the development of the anti-witchcraft ordinances7. Knowledge about witchcraft generated by administrators on-the-ground contributed strongly to ways in which the courts were able to consider witchcraft, both as a crime in and of itself, and in relation to murder. And, such knowledge offered colonial justices avenues of “culturally reasonable conjecture” in dealing with witchcraft and helped to shape the broader knowledges brought to bear on the cases of witch-killing like those detailed below (Stoler 1992: 153).
11The law then provided a primary avenue through which colonial regimes throughout Africa sought simultaneously to discipline witchcraft and to deny its efficacy and/or existence. In Kenya (as elsewhere throughout the continent), colonial administrations developed a series of anti-witchcraft ordinances which criminalized “black magic” and under some readings, “white magic” as well. Yet, for all of their discursive wrangling, these ordinances actually failed to offer any clear definitions of witchcraft per se.
12In the majority of cases, the evidentiary demands and the official incredulity enshrined in colonial anti-witchcraft ordinances rendered legal recourse against the activities of alleged witches impractical if not wholly impossible. Rather than curbing witchcraft-related disorder, anti-witchcraft legislation instead regularly functioned as an indirect stimulus to violence against alleged witches when the purported victims of witchcraft who were unable to find satisfaction in colonial courts took matters of “justice” into their own hands by killing alleged witches through acts of individual or vigilante violence. The following section traces the genealogy of the Witchcraft Ordinances in Kenya, highlighting how colonial legal languages aimed to (re)formulate witchcraft according to colonial prerogatives.
13In Kenya, the first anti-witchcraft legislation was debated, revised and passed by the Legislative Council in 19098. The initial goal of the Witchcraft Ordinance (1909) was “to make provision for the punishment of a person practicing or making use of so called witchcraft”. In the second reading of the bill, two members of the Legislative Council proposed without success excising the words “supernatural power” from the second section of the bill. In the same meeting, the Crown Advocate successfully moved for the insertion of language protecting African functionaries from prosecution under the ordinance. The new language read,
“No proceedings for an offence under this Ordinance shall be taken against a Chief, Sub-Chief, Headman, or elder, on account of anything done by such Chief, Sub-Chief, Headman or Elder in exercise of his authority as such, except with the previous sanction of the Governor”9.
14The bill was passed on its third reading in the summer of 1909. As Richard Waller (2003: 245) has succinctly summarized, “the 1909 legislation established [. . .] three criminal offences; to claim to be a witch or to ‘pretend to exercise or use any kind of supernatural power, witchcraft, sorcery or enchantment [. . .] for the purposes of gain’ (section 2); to advise others on how to use witchcraft or to supply them with the ‘pretended means of witchcraft’ (section 3); and to use such advice or means to ‘injure any person or property’ (section 4)” (Waller 2003)10.
15Punishment varied from terms of imprisonment of between one year and ten years.
16The evidentiary demands of the 1909 ordinance proved unwieldy and the 1918 revision of the ordinance repealed section two which stipulated that the offence of witchcraft entailed “gain” and substituted instead the following language:
“Any person who hold himself out to be a witch doctor able to cause fear, annoyance, or injury to another in mind, person, or property or who pretends to exercise any kind of supernatural power, witchcraft, sorcery or enchantment calculated to cause such fear, annoyance or injury shall be guilty of an offence and shall be liable to imprisonment of either description for a term not exceeding one year [. . .].”
17Further, the 1918 ordinance struck from section four of the original ordinance the words “to injure any person or property” and substituted the language “to cause fear, annoyance or injury in mind, person, or property to any person”11. The new language thus shifted the heart of the witchcraft offense from material gain to pretense to supernatural power and the production of psychological and/or physical harm of through such pretense.
18The core remained the same in the 1925 Ordinance, but fresh language entailed provisions shaping the legality and illegality of witchcraft accusations and the roles of functionaries and British authorities in dealing with them. First, the new language made accusing a person of “being a witch or with [. . .] practicing witchcraft” a crime punishable by fine or imprisonment unless the accusation “was made to a district commissioner, a police officer, an official headman or any other person in authority”. Further, the new terms of the ordinance made a headman’s failure to report the “practice or pretended practice of witchcraft by any person” to the district commissioner a crime of omission, punishable by fine or imprisonment. Of course, the ordinance also forbid headmen to in any way allow witchcraft or any act that could be considered to counter the provisions of the Witchcraft Ordinance. The ordinance’s new language also rendered employment or solicitation of another “to name or indicate by the use of any non-natural means any person as the perpetrator of any alleged crime or other act complained” a crime punishable by fine or imprisonment12.
19In sum, additions to the 1925 Ordinance carried two basic aims. First was to expand the involvement of functionaries like chiefs and headmen in combating witchcraft-related crimes by assigning them new roles like hearing witchcraft accusations and reporting witchcraft activities in their locations while at the same time criminalizing a willful or indifferent neglect to do so13. The additional language also placed more emphasis on the criminality of witchcraft accusations, effectively rendering the activities of diviners as crimes and again blurring the divide between witchcraft for harm and witchcraft for healing.
20Overall, none of the laws in the series of anti-witchcraft ordinances offered a clear definition of witchcraft per se. And each subsequent law was ineffective in diminishing witchcraft because as elements of an evidentiary-based legal system, the ordinances required tangible evidence to prove the perpetration of a “crime” which was inherently invisible. The inability of the alleged victims of witchcraft to effectively seek recourse under the Witchcraft Ordinances often precipitated the sorts of witch-killing cases analyzed below.
21The case reports of the Court of Appeal for Eastern Africa and the High Courts of Kenya, Tanganyika, and Uganda indicate that crimes related to witchcraft were regularly tried by these bodies14. In the bulk of these cases, defendants were alleged to have killed because they believed that the deceased had in some way “bewitched” them or someone close to them. These cases contributed to the refinement of the legal meanings of witchcraft and to the elaboration of legal concepts central to the prosecution of capital crimes15.
22The first opinion on a witchcraft-related case to be recorded in the digests of the East Africa Court of Appeal was Rex versus Karoga was Kithengi and 53 Others (1913) in which the members of a Kikuyu “kiama”, a type of pre-colonial governing body or council of elders co-opted into the colonial administrative apparatus, appealed from a murder conviction for having burned to death two alleged witchdoctors with the sanction and participation of their chief. During the High Court trial, solicitors for the kiama members raised a three-pronged defense: (1) members of the kiama had not been instructed as to the limits of their jurisdiction under the Native Tribunal Rules, 1911, (2) they acted on the advice of their chief, and (3) in doing so, they were justified in exercising their “ancient customary jurisdiction to sentence witchdoctors to death”. Unsurprisingly, the courts rejected this defense on the grounds that under colonial law, the kiama no longer had the authority to exercise customary forms of justice16.
23The appeals court located the primary responsibility for the killings in the chief’s neglect to instruct the councilors that they were no longer to practice certain forms of “customary” justice and found the kiama members guilty of the “abatement of culpable homicide”. In this case, witchcraft is tangled up with the larger issue of the limits of the power imbued or accorded by the colonial state to “traditional” authorities and how such authorities should and could come to know the limits or expansions of their powers17.
24Four cases recorded in the digests aim to refine witchcraft as laid out by the 1909 Witchcraft Ordinance and the 1925 Witchcraft Ordinance. Rex versus Mtuiniwara (1914) overturned the conviction of a witchdoctor on the grounds that in order for a conviction to stand, sections of the ordinance stipulating “holds himself out as a witch doctor” and “for the purposes of gain” had to be read and proved in tandem18. Working on a similar principle, the 1915 decision in Rex versus Joke overturned a conviction under the Witchcraft Ordinance on the grounds that while the appellant had accepted payment for administering “medicine” he neither “held himself out to be a witchdoctor” nor demanded payment. These two cases likely contributed to the removal “gain” as a constituent element of the offence of witchcraft in the 1918 revisions of the ordinance discussed above.
25Next, the 1916 decision in Rex versus Matolo overturned a conviction under the Witchcraft Ordinance on the grounds that though the appellant did hang up a calabash of “witchcraft medicine” on his property, an “intent to injure” could not be proved19. Finally, the decision in Shangati ole Sauroi (1929) held that the phrase “means and process” in Section 4 of the 1925 Witchcraft Ordinance could be taken to include spoken words, thereby rendering witchcraft an act as much of speech as of substances20.
26In addition to the “kiama” case, another case addressed the alleged witchcraft of the deceased as a defense to murder. In Rex versus Kimonirr and Five Others (1916), the principal appellant had accused the deceased, Chesang, of killing her husband and son by sorcery. Chesang, as local custom was alleged to dictate, hanged himself on the instructions of his family. The decision of the court turned on local custom, as the Chief Justice explained first that the death sentences could not be expunged because “local practice” had not been followed to the letter, but that the sentences could be reduced to “abatement of suicide” because findings proved that among the tribe in question, killing a witchdoctor was not “morally wrong”21.
27As evidenced in the reports summarized above, a range of topics surrounded both cases of witchcraft and murder cases in which witchcraft was raised as a defense. Up to the early 1930s, colonial High Court and East Africa Court of Appeal opinions in witchcraft-related criminal cases, especially murders, aimed to refine the fuzzy definitions of witchcraft laid out in anti-witchcraft legislation and considered witchcraft in relation to the broader issues of local practice or “custom”.
28At the conclusion of the 1932 case, Rex versus Kumwaka s/o Mulumbi and 69 others, 60 Wakamba men were sentenced to death in the High Court of Kenya for killing a neighbor woman whom they alleged to have been a witch. The most high-profile witch-killing case of the colonial era, Rex versus Kumwaka, refined and reinforced precedents in regard to key legal concepts such as “malice aforethought” and “grave and sudden provocation”. More generally, the case redirected the focus of legal reckoning in witch-killing cases from custom to commutation.
29A summary section from the Law Reports of Kenya succinctly narrates the events of Rex versus Kumwaka as assembled by the courts.
30The first accused (Kumwaka) summoned the rest of the accused and brought them to the vicinity of the hut in which was his wife, the woman believed to have been bewitched. Next, the witch, the deceased, was seized and brought to the sick woman’s hut and ordered to remove the spell. The accused alleged that she had removed half the spell during the night. Early in the morning, the witch was detected running away. All accused ran after her and beat her with the thin sticks referred to above. As a result of the beating the witch was killed22.
31Similar to defense claims made in the “kiama” case, the defense in Rex versus Kumwaka asserted that the men had done nothing wrong in killing the alleged witch but, instead, had been carrying out king’ole, the Kamba institution of justice directed against social malefactors like recidivist witches and thieves. The justices of the Supreme Court of Kenya were, however, unconvinced and handed down death sentences on the 70 accused. Writing for the court, Chief Justice Jacob Barth explained that “the fact of death in his judgment was sufficiently proved, and the issue to be decided was whether or not the facts proved amounted to murder”23. The opinion that “the facts of the case amounted to murder” rather than manslaughter hinged on the principle of malice aforethought, “knowledge that the act or omission will probably cause the death of or grievous harm to some person”24. Applying the principle to Rex versus Kumwaka, Chief Justice Barth wrote: “It seems obvious to me that if 70 men and lads combine to beat a woman although with thin sticks, they must have the knowledge that their actions would probably cause death or at least grievous harm to the woman”25.
32The Wakamba men appealed their case, but the High Court of Appeal for Eastern Africa struck down the appeal. Issues of precedent, procedure, and the law as written framed Chief Justice of Uganda Sir Charles Griffin’s opinion on Rex versus Kumwaka. The opinion also addressed the appellants’ contentions that beliefs in witchcraft constituted mitigating circumstances in capital crimes. “The principal grounds of appeal” proffered by the defense were that Jacob Barth, Chief Justice of Kenya, ought to have accepted the evidence of accused No. 1 that there was no mens rea, that there was no evidence of combination or conspiracy to administer corporal punishment, nor of concerted action as to the cause of death; and that homicide was excusable, or alternatively, that it did not amount to murder, but to manslaughter, having regard to Article 7 of the Kenya Colony Order-in-Council, 1921.
33The defense also argued that amongst the Kamba appellants, “There was genuine and real belief in powers of the witch and fear as to the result of her spell”. Countering the appellants’ claims, the three justices of the High Court of Appeal for Eastern Africa concurred that “the evidence justified the finding that the deceased died as a result of the beating” and that “the Crown was not obliged to prove intention to kill”. Furthermore, the Justices closed down the appellants’ last argument adding, “native belief in witchcraft does not justify deviation from law by inflicting punishment on suspected witch”26. Authoring the decision, Griffin elaborated on the points addressed above. Griffin and the other justices supported Barth’s opinion that the beating administered by Kumwaka and his group had led to the alleged witch’s death and pointed out that the 70 Wakamba had never suggested that they were innocent of beating her. Griffin explained, “On perusing the evidence we entertain no doubt that she died, and died as a result of the beatings administered. The accused themselves all admit the beating and killing and none has sought to place blame on the rest or to exculpate himself”27.
34Griffin then turned to the argument that “there was no intention to kill the deceased”. He asserted that “it was no part of the case for the Crown that there was any intention to kill”, nor was Chief Justice Barth’s verdict “found[ed] on intention to kill”. He then cited the relevant sections of the Kenya Penal Code on which Barth’s decision had been based. “Section 186 of the Penal Code” Griffin wrote, “defines murder as follows: ‘Any person who of malice aforethought causes the death of another person by an unlawful act or omission is guilty of murder’”. He added that the legal concept of “malice aforethought” also informed Barth’s decision. The Penal Code, Griffin explained, explains malice aforethought as follows:
“Malice aforethought shall be deemed to be established by evidence proving–(b) knowledge that the act or omission causing death will probably be the cause of death or grievous harm to some person [. . .] although such knowledge is accompanied by indifference whether death or grievous bodily harm is caused or not, or by a wish that it not be caused”28.
35The members of the court, Griffin maintained, agreed entirely with Barth’s decision that the appellants knew that beating Mwaiki would “probably cause death or at least grievous harm”29.
36Griffin and the other justices also closed down the section of the appeal based on arguments of self-defense premised on the case Regina versus Rose. In order to explain the Justices’ reasoning, Griffin first cited the head-note of the case:
“Under circumstances which might have induced the belief that a man was cutting the throat of this wife, their son shot and killed his father. On the trial of the son for murder: – Held: that if the accused had reasonable grounds for believing and honestly believed that his act was necessary for the defense of his mother, the homicide was excusable.”
37Griffin explained that “the important words are ‘that the act was necessary’” and therefore he and the other Justices were “unable to appreciate the relevance of Reg. v. Rose to the facts” of Rex versus Kumwaka, “where the killing was in no way necessary”. In the concluding the decision, Griffin argued,
“It is also widely known, and as appears from the evidence in this case that Government does not tolerate the killing of witches. The plea has been frequently put forward in murder cases that the deceased had bewitched or threatened to bewitch the accused, and that plea has been consistently rejected except in cases where the accused has been put in such fear of immediate danger to his own life that the defense of grave and sudden provocation has been held proved. For Courts to adopt any other attitude to such cases, would be to encourage the belief that an aggrieved party may take the law into his own hands, and no belief could well have been more mischievous or fraught with greater danger to public peace and tranquillity”30.
38The High Court of Appeal for Eastern Africa’s dismissal of the appeal in the case of Rex versus Kumwaka was thus constructed from a range of angles. With reference to precedent and law as written, the Chief Justices’ opinion illustrated that beliefs in witchcraft did not legally constitute mitigating circumstances in the case of capital crimes. The opinion asserted that evidence of “malice aforethought” in the killing of witches was sufficient ground to consider their killings murder rather than manslaughter. Also, according to the opinion, assaults on or killings of witches could not be considered acts of self-defense. Finally, the opinion took a strong position against vigilante justice. These positions became key reference points in subsequent cases of witch-killing discussed below.
39From the 1930s to the eve of the Mau Mau period, witch-murder cases continued to be regularly tried in the colonial High Courts and the East Africa Court of Appeal and were increasingly recorded in the digests of these bodies31. Unlike the opinions on pre-Kumwaka cases discussed above which did not reference each other, those opinions dating from the late 1930s onward were highly referential and concerned primarily with the larger legal issue of “grave and sudden provocation”. This section briefly summarizes the witchcraft-related murder cases heard in the East Africa Court of Appeal and recorded in its available digests from the late 1930s to the early 1950s32. It follows the development and refinement of the concept of “grave and sudden provocation” vis-à-vis defenses claiming the witchcraft of the deceased and traces the “conversations” about these issues within the opinions. These “conversations” in turn contributed to the imperial network of knowledge about witchcraft, custom, British justice, and native mentality. Overall, the digest cases ultimately turn on the same question: When, if ever, is the witchcraft of the deceased sufficient to commute a capital sentence?
40The witchcraft-related murder cases recorded in the East Africa Court of Appeal digests between 1939 and 1941 were appeals from death sentences for murders in which the appellants claimed that they had killed the deceased after coming to believe that the deceased had somehow practiced witchcraft against them or members of their families. The appeals were made on the grounds that the witchcraft of the deceased had constituted “grave and sudden provocation”, a legally mitigating condition capable of reducing sentences of murder to ones of manslaughter or even lesser charges. In the 1939 case, Rex versus Kimutai arap Mursoi, the appellant argued that he had killed the deceased because he believed the deceased was a wizard and had laid a spell on the appellant’s child33. In the 1940 case, Rex versus Mawalawa bin Nyangweza, the appellant admitted he had killed the deceased because “witch doctors” had told the appellant that the deceased had bewitched the appellant’s brother and the appellant then decided to kill the deceased for bewitching his whole family34. Similarly, in the 1941 case, Rex versus Sitakimatata s/o Kimwage, the appellant argued that he had been told by the deceased that the latter had killed the appellant’s wife by witchcraft and would do the same to the appellant. The appellant then decided to kill the deceased and did so a few hours later35. The appellants in the 1941 case, Fabiano Kinene s/o Mukye, Seperiano Kiwanuka s/o Kintu, Albert Iseja s/o Kintu, argued the witchcraft of the deceased had driven them to kill him by inserting 20 green bananas into his anus after coming on the deceased crawling naked around their compound36.
41In deciding these cases, the courts debated the constitutive elements of “grave and sudden provocation”, and turned to the precedent established by the decision in Rex versus Kumwaka. The following paragraph from Kumwaka emerged as a veritable “go-to passage” in adjudicating witchcraft-murder appeals. The passage reads as follows:
“The plea has frequently been put forward in murder cases that the deceased had bewitched or threatened to bewitch the accused, and that plea has been consistently rejected except in cases where the accused has been put in such fear of immediate danger to his own life that the defense of grave and sudden provocation has been held proved”37.
42In the first case, Kimutai arap Mursoi, the appeals court dismissed the appeal on the grounds of the Kumwaka passage cited above, and subsequent decisions existed in dialogue with each other. For example, in Rex versus Mawalawa bin Nyangweza, the justices dismissed the appeal, citing both precedents established by Kumwaka and Kimutai arap Mursoi, while the appeal in Rex versus Sitakimatata s/o Kimwage was also dismissed, again citing Kumwaka and Kimutai arap Mursoi. In contrast, the murder convictions were reduced to manslaughter in Fabiano Kinene s/o Mukye and Others, the courts relying on Kumwaka, Kimutai arap Mursoi, and Mawalawa bin Nyangweza, to reach the decision that “grave and sudden provocation” had been held proved.
43But while straightforward in their rejection or acceptance of the appeals, the decisions in these cases also complicated “grave and sudden provocation”, parsing it and reading it in conversation with other legal conditions present in the cases such as “malice aforethought” and a “real but mistaken belief” in witchcraft. In Mawalawa, for instance, the courts were concerned with assessing the veracity and reasonableness of the appellant’s claims about the witchcraft of the deceased, but nonetheless ruled that “There is some evidence tending to support his [the appellant’s] story but even assuming it to be true the decided cases in East Africa establish that it does not amount to such legal provocation as would justify us in reducing the offence to manslaughter”38. Next, in formulating the opinion in Sitakimatata s/o Kimwege, the justice criticized the “go-to passage” from Kumwaka, attending to the subjectivity of the affective element of “fear” and raising the question of “native mentality”. The opinion explained,
“The phraseology used in this passage seems to me, with respect, not to be entirely free from obscurity. It is rather difficult to discover from the concluding phrase what standard of fear is required to establish a defense of provocation based on a belief in witchcraft, and the emotion of fear (which does not seem to me to have any place in the English doctrine of provocation) is confused with the emotion of anger, which is, I think, the natural and only product or result of provocation received”39.
44The justice’s ultimate conclusion was that while the element of “fear” was ambiguous, the circumstances of the case did not achieve the standard of “suddenness” demanded by the Penal Code to prove “grave and sudden provocation”.
45The decision in the case of Fabiano Kinene s/o Kinene and Others entailed the most complex analysis by the courts and came to be cited in almost all subsequent witchcraft-related murder cases. While addressing “grave and sudden provocation”, the decision also focused on questions of “malice aforethought”, on the veracity and reasonableness of the appellants’ witchcraft beliefs, and on “native mentality” more generally. First, the appeals court concurred with the original High Court of Uganda judgment that the statements of Fabiano and his cohort demonstrated that they had killed the deceased with “malice aforethought”. Second, the court drew on the opinions of assessors and on general attitudes about “native mentality” and witchcraft in deciding that Fabiano and the other appellants did hold a “real but mistaken belief” in the witchcraft of the deceased. The decision explained,
“With their strong suspicions of his past history they would need very little to convince them and the sensitiveness of the African mind in this respect is shown by the evidence of the Muruka chief Fenekansi that ‘if in the night I saw a man naked crawling in my compound I would think he was a witch doctor actually practising witchcraft’”40.
46Yet drawing on the opinions in Kumwaka, Kimutai arap Mursoi, and Mawalawa bin Nyangweza, the courts concluded that the appellants’ belief in witchcraft while introducing the possibility of a defense of “grave and sudden provocation”, did not alone constitute sufficient grounds to prove “grave and sudden provocation”. Like the decision in Sitakimatata s/o Kimwege, the Fabiano decision critiqued the notion of “fear” and added that witchcraft could also be regarded as inducement to the sort of anger which in turn constituted an element of the “heat of passion” phrasing in the Uganda Penal Code section dealing with “provocation”. The justice explained the principle generally and in regard to the circumstances of Fabiano. He wrote,
“In our opinion the principle in those cases [Kumwaka, etc.] is stated somewhat too narrowly and perhaps not altogether accurately, in that the words ‘in the heat of passion’ used in s. 198 of the Penal Code (Uganda) are more properly referable to the emotion of anger than to that of fear. We think that if the facts proved establish that the victim was performing in the actual presence of the accused some act which the accused did genuinely believe, and which an ordinary person of the community to which the accused belongs would genuinely believe, to be an act of witchcraft against him or another person under his immediate care (which would be a criminal offence under the Criminal Law (witchcraft) Ordinance of Uganda and similar legislation in other East African territories), he might be angered to such an extent as to be deprived of the power of self-control and induced to assault the person doing the act of witchcraft. And if this be the case a defense of grave and sudden provocation is open to him. It must always be a question of fact as to whether he is in all the circumstances of the particular case acting in the great of passion caused by grave and sudden provocation and of course on such an issue he must be given the benefit of any reasonable doubt. We think it not unreasonable to say that in the present case the accused persons, when they seized the deceased in the compound and proceeded to kill him, may have been so acting”41.
47The above passage thus weaves together two important ideas. First, the opinion argues that more than one affective state, “anger” as well as “fear”, is sufficient to induce the immediate and overwhelming passion which is an ineluctable constituent element of “provocation”. Second, the opinion attends to “native mentality” vis-à-vis the question of witchcraft, identifying a standard of reasonableness which takes in local mindsets and mores, but also designates witchcraft as it is extrapolated in colonial law. The opinion then went on to reference directly Sitakimatata s/o Kimwago, reiterating the decision in that case that both “gravity” and “suddenness” needed to be held proved in order for a defense or an appeal on the grounds of “grave and sudden provocation” in a witch-murder case to be successful.
48Ultimately, the appeal court reduced the murder conviction to one of manslaughter. The justice explained how the elements of “grave and sudden provocation” were present in the case. The “highly suspicious actions of the deceased”, crawling naked around the compound at night, could be reasonably considered fear and/or anger-inducing behavior according to local standards, and also an offense according to colonial anti-witchcraft legislation. Thus, the behavior of the deceased constituted an “immediate provocative act” in the eyes of the courts. Also, the appeals court’s decision emphasized the issue of the time between the “provocative act” and the killing of the deceased, noting how Fabiano and his cohort had almost instantly attacked the deceased upon finding him crawling naked in their space. From the court’s perspective then, the killing had been an immediate response to “grave and sudden provocation”. Nonetheless, the decision concluded by reiterating that despite the decision in Fabiano, the justices “in no way mean[t] to suggest that we believe witchcraft per se will constitute a circumstance of excuse of mitigation for killing a person believed to be a witch or wizard when there is no immediate provocative act”42.
49The decision in Fabiano became a key referent in witchcraft-related murder cases from the early 1940s onward in much the same way that the decision in Kumwaka was central in the previous decade. Subsequent decisions in both the colony’s High Courts and the East Africa Court of Appeal also asserted the precedent enshrined in Fabiano that a proven belief in the witchcraft of the deceased alone could not reduce a conviction of murder. Referencing each other as well as Fabiano and other earlier cases, a number of these decisions also aimed to refine the element of “reasonableness” in regard to appellants’ beliefs in the witchcraft of the deceased. In the 1942 case, Rex versus Nzau wa Mukwata, the defendant argued that he had killed the deceased, his mother-in-law and a reputed witch, because he believed she had bewitched his children to death, and when confronted, the deceased threatened the defendant with death by witchcraft43. In a second 1942 case, Rex versus Kelementi Maganga s/o of Ochieng, Zadoki Omoiti s/o Okechi, the two appellants beat the deceased, reputed to be a wizard, and the second appellant also speared him, claiming to the deceased’s son in the course of the beating that he had caught the deceased practicing witchcraft by walking naked around the second appellant’s home. In the 1945 case, Rex versus Kajuna s/o of Mbake, the appellant claimed that he had killed the deceased, his father, because the appellant believed his father to have been killing the appellant’s child by supernatural means44. In the 1949 case, Rex versus Petero Wabwire s/o Malemo, the appellant claimed that he had killed his wife because he believed she was practicing witchcraft against him as a result of her having “medicine” in her possession and refusing to tell the appellant where she had acquired it45. Finally, in the 1951 case, Eria Galikuwa versus Rex, the appellant claimed that he had been threatened with death by the deceased, a witchdoctor, unless he paid the deceased 1,000 sh., and being unable to pay instead killed the deceased46.
50Referencing Fabiano, the decisions in each of these cases focus on standards of “reasonableness” pertaining to the appellants’ perceptions of the behaviors of the deceased as being “provocative” acts of witchcraft. The decision in Nzau wa Mukwata held that the circumstances of the case put in evidence by the appellant before the local magistrate seemed to show that the deceased had committed an act of witchcraft according to anti-witchcraft legislation when she threatened the appellant with death-by-witchcraft and that her speech constituted a “provocative” act from the perspective of a “reasonable” person of their community. Reiterating the precedent that a “mere belief in witchcraft will not mitigate a killing” without another factor attaching and the Colonial Office’s opinion on the primacy of “local” circumstances in dealing with witchcraft-related murder cases, the appeal court commuted the death sentence. The appeal in Kelementi Maganga and Zadoki Omoiti claimed a similarity in facts with Fabiano, and the appeal court’s decision held that both the appellants’ testimony on the deceased’s behavior–“walking naked around another’s property” and the testimony of the deceased’s sons that he “was commonly regarded as a wizard and than as such his company was shunned by neighbors” was sufficient to reduce the murder conviction to one of manslaughter.
51According to the decisions in the two cases discussed above, the appeal court justices had been willing to give the appellants the benefit of the doubt in assessing the “reasonableness” of their claims about the “provocative” witchcraft of the deceased. In the cases of Kajuna s/o Mbake and Petero Wabwire s/o Malemo, the appeals court rejected the appeals on the grounds that “grave and sudden provocation” was not proved and that the appellants’ witchcraft beliefs were not reasonable. In Petero Wabwire, the court found the appellant’s belief that the deceased was practicing witchcraft against him because she possessed substances or “medicine” whose origin she refused to disclose was not reasonable according to local standards. Similarly, in Kajuna, the justices found that the appellant’s belief that the deceased was killing the appellant’s family by “supernatural means” was “unreasonable” because the appellant could not cite an immediate, “provocative” act which inspired this belief. The decision explained,
“A mere belief founded on something metaphysical as opposed to something physical, that a person is causing the death of another by supernatural means however honest that belief may be has not so far as we are aware been regarded by this Court as a mitigating circumstance in law [. . .]”47.
52Finally, the decision in Eria Galikuwa refined and reemphasized many of the principles laid out in the cases above. Criticizing the imprecision in the decisions in Kimutai, Mawalawa, Sitakimatata, and Fabiano, the justice asserted that “anger” as well as “fear” was a necessary constituent of “provocation”. The decision also underscored that the “provocative” act of witchcraft needed to be “overt”, i.e., “physical”, “visible” or “audible” and had to constitute witchcraft according to anti-witchcraft legislation. And drawing on Petero Wabwire, the decision expanded the timeframe for “provocation”. It explained,
“The Penal Code does not say that the unlawful act or insult done to or in the presence of another person must be one entailing immediate consequences of a wrongful nature, and there might be upon occasion a wrongful act which was indicative of a future intention on the part of the doer and which therefore might be of such a nature as to come within the legal definition of provocation”48.
53The appeal court found that the conditions for “provocation” were not satisfied because the appellant had shown himself to be motivated by “fear alone” and that he acted in “despair” rather than in the “heat of passion”. Further, the decision added that the appellant’s actions were the result of deliberated intention rather than of being “suddenly deprived of his self-control”. And finally, drawing on the opinion of the assessors, the appeal court found that the appellant had not availed himself of the legal options available for dealing with the threats of the deceased and instead “chose deliberately to take the law into his own hands”49.
54In each of these cases, the courts were restrained by the law from considering witchcraft as a defense without attaching it to another legal category of defense, most often “grave and sudden provocation”. But an attention to witchcraft enabled the courts to distinguish killings in retribution for witchcraft from killings for more mundane reasons. Overall, these cases both followed and reinforced the precedent of recommendation to the Governor’s clemency, a practice which often resulted in the reduction or squashing of capital sentences.
55A central aim of this article has been to trace how the colonial state attempted to grapple with the problem of witchcraft-related violence through the refinement of law and important legal concepts within the institution of courts. Yet at the same moment these cases illustrate the state’s attempt to assert its hegemony over the exercise of violence, they also reflect how the archival remains of non-state actors point to gaps in colonial control (Pandey 2000). An historical analysis of such cases has contemporary relevance and resonances because documentary and ethnographic evidence suggests crimes related to witchcraft are not simply a colonial, but also a current problem as well50. Rather than declining during the second half of the twentieth century, sources suggest that “witchcraft”-related violence and disorder has likely increased while the law’s efficacy in dealing with such problems has not.
56For one, the state’s historic approach to witchcraft in regard to “witchcraft” described above remains legally enshrined51. The Witchcraft Ordinance of 1925 and the Witchcraft Ordinance (revised) of 1981, the law contra “witchcraft” currently on the books, are in language and substance practically identical. And, both the colonial and post-colonial ordinances entail a “poetics of incredulity” which seeks to deny the reality of “witchcraft” while at the same moment disciplining “witchcraft” practices and beliefs52. Like the colonial-era ordinance, post-colonial anti-witchcraft legislation indirectly stimulates violence against “witches” as the alleged victims of witchcraft find themselves unable to pursue justice in the post-colonial courts.
57While legal approaches to witchcraft have remained consistent, the type and frequency of violence related to “witchcraft” practices and beliefs have not changed significantly either. The following cases–one recounted in the legal language of the courts and the other in pithy journalistic prose–typify “witchcraft”-related violence in Kenya, and indeed, across Africa generally.
58The first case summary reads,
“The accused, a member of the Mkamba tribe was charged with the murder of K., his mother-in-law, who was reputed to be a witch and whom the accused believed had caused the deaths of several of his relatives by witchcraft. Accused met K. and besought her to cease from practising witchcraft. K. answered, ‘You are always accusing me of practising witchcraft, you also will die by witchcraft.’ Whereupon the accused became infuriated and attacked and killed K.”53.
59And the second summary states,
“A man and his wife were killed in cold blood on suspicion practicing witchcraft [. . .]. Villagers forcibly removed them from their house, blindfolded them and tied them to a tree before stoning them to death [. . .] a witness [. . .] told journalists that the couple unsuccessfully pleaded for mercy, promising not to practice witchcraft any more”. He said, “They had earlier confessed to having bewitched several families in the area. They vowed not to repeat the act, but angry villagers could hear none of it [. . .]” (Otieno 2003).
60The first of these cases was heard by the Supreme Court of Kenya in the early 1940s while the second was described in a 2003 edition of the Daily Nation newspaper. But despite being separated by 60 years, the circumstances of the cases are remarkably similar. In each instance, spectacular and public violence against “witches” was produced by suspicions and admissions of “witchcraft”, the latter made in defiance or in desperation. These cases also point to the consistency with which individual or “communal” violence against “witches” has been practiced in Kenya. But at the same moment, cases such as these also highlight the absence of the state from disputes over “witchcraft”. In general, the institutions and agents of the state have become involved actively and consistently involved in “witchcraft” cases after “witchcraft”-related violence has created challenges to state’s ability to maintain law and order.
61The cases analyzed above demonstrate how witch murders offered significant spaces in which the legal meanings of witchcraft were refined. In the course of these cases, the courts parsed the somewhat ambiguous language of anti-witchcraft legislation in efforts to determine whether the term “witchcraft” officially denoted substances, acts, speech, or a combination of all three. The cases also brought to the fore the question of the extent to which intent was a necessary constituent of witchcraft.
62The opinions detailed above also point to some of the ways in which murder cases contributed to the elaboration of key legal concepts. For example, opinions handed down in such cases highlighted the importance of proving “immediacy” of a threat when raising a defense of “grave and sudden provocation”. Legal opinions in witch-murder cases also attended to the importance reading and applying concepts like “malice aforethought” and “real but mistaken belief” in tandem. Cases of witch murder show that crimes related to witchcraft regularly entered colonial records and became objects of official analyses through avenues other than simple contraventions of the Witchcraft Ordinances (Amin 1995).
63In sum, an historical analysis of witch murders in the courts and on the books points to some of the ways in which witchcraft has constituted an important space in which larger questions of power have been contested from the colonial era to the present day. Such an analysis demonstrates that “witchcraft” has existed not as an anthropological curiosity, but as a readily recognized and legally intractable source of violent disorder in Kenya.