Early twentieth-century South Africa was a composite society—“part settler state and part African colony … includ[ing] diverse recently conquered African polities as well as a divided white population.”Footnote ¹ Mining industrialization and British imperialism, particularly after the discovery of substantial gold deposits and the founding of Johannesburg in 1886, put pressure on southern African peoples and states to function as an integrated labor market, and on their leaders to submit to an overarching political authority. These developmental and administrative rationalizing forces were given greater scope in the years following the South African War of 1899 to 1902, especially in the defeated Boer republics of the interior. Renamed the Transvaal and Orange River Colonies, these territories were initially under the direct rule of British High Commissioner Alfred Milner. They took the lead in a process of state-building that continued well beyond their political amalgamation with the coastal colonies of the Cape and Natal to form the Union of South Africa in 1910.Footnote ² It has been argued that this institutional reconstruction left South Africa with “a modern civil service, with controls and an information-gathering capacity sophisticated enough to … make the competence, helpfulness, and honesty of individual state officials relatively less crucial.”Footnote ³

The post-1902 civil service undoubtedly did evince many characteristics of the ideal type of the modern “bureaucratic state,” staffed by technically competent salaried officials committed to carrying out their duties dispassionately, without “regard for persons.”Footnote ⁴ Many Transvaal-based English-speaking officials took pride in a professionalism that they believed distinguished them from their reputedly corrupt and capricious Boer predecessors in Paul Kruger's pre-1900 Zuid Afrikaansche Republiek.Footnote ⁵ But how far could the officials of the new state, with its systematic “information-gathering” ambitions, truly succeed in isolating themselves from the often patriarchal and personalized idioms and expectations of authority and rule that endured among the incompletely colonized peoples of this composite society? These idioms and expectations—encompassing a sense of the respect and implicit obedience due to male elders and patrons by women, juniors, clients, and other dependents—were bound up with the privileges and performances of “chieftaincy,” an increasingly reified but still very real political and distributive institution for many black South Africans in the early twentieth century.Footnote ⁶ European power in South Africa depended upon African intermediaries, including “traditional” rulers (or “chiefs” and “headmen”), messengers, clerks, translators, and police, whose employment afforded them some scope to shape the social order. Could the “white” state limit that scope without engaging these intermediaries on their own terms and in their own idiom—without, that is, its representatives becoming “chiefs” to the “chiefs” and their subjects?Footnote ⁷ And if it could not, what were the consequences for the standing of the state's officials as “modern” bureaucrats?

Colonial administrators everywhere confronted variations on this dilemma of having to depend on “indigenous systems of communication” and local intermediaries whose “voices” they “were forced to register … in ideology and heed … in practice even if they despised and misrepresented them.”Footnote ⁸ Although this article's main focus is on African police rather than “customary rulers,” studies of “indirect rule—rule through indigenous authority structures and norms”Footnote ⁹—help to specify what was at stake. Scholars have offered very different theorizations of indirect rule, not all of them reducible to the singularities of specific colonial contexts. Yet many would recognize Karen Fields's depiction, drawn from British Central Africa, of indirect rule as “doubly articulated. One articulation made the African masses subject to customary rulers; the other made customary rulers subject to the Crown's representatives.” Not all scholars, however, accept Fields's further claims that, given the fiscal and other constraints that colonial governments experienced, the first articulation was the “more fundamental”; that “real power issued from the ruled” for whom the procedural legitimations of modern states were less relevant; and that consequently the colonial state, thus inflected by African idioms and concerns, was an “archaic” “throwback,” a less than fully “‘modern’ specimen.”Footnote ¹⁰ Crawford Young's emphasis on the role of exploitation and violence in the construction of the Belgian Congo leads him to conclude that “real power” moved in the opposite direction. He identifies “Bula Mutari”—a nickname meaning “crusher of rocks,” given by Congolese to Henry Morton Stanley when he worked for the Belgian King Leopold—as a “metaphor captur[ing] well the crushing, relentless force of the emerging colonial state in Africa.” From this perspective, “allies … in … existing African political structures” were “intermediaries of [colonial] hegemony,” “imprisoned by the imperatives of reproducing the system they served.”Footnote ¹¹

Mahmood Mamdani likewise sees indirect rule as a top-down system of domination, a “decentralized despotism” introduced by calculating European officials, which has endured in the postcolonial era and remains a potent obstruction to democratization. “Behind the mask of indirect rule lay the day-to-day routine—customary—violence of the colonial system.” Indirect rule, as Mamdani conceives it, was a self-contained sphere of “Native Authority”—“a world of the customary from which there was no escape,” and in which the “authority of the chief … fused in a single person all moments of power: judicial, legislative, executive and administrative.” This sealed-off rural domain was one aspect of a “Janus-faced, bifurcated” colonial order whose other, urban “face” “spoke the language of civil society and civil rights.” Mamdani designates ordinary people in the former domain “subjects,” and those in the latter “citizens,” concepts that define mutually exclusive ways in which distinct categories of person (usually specified racially in the colonial era) related to the state. For Mamdani, then, indirect rule was not crucially shaped from below by Africans in situations of colonial weakness, or “articulated” with other parts of the colonial state, thereby permitting its logic to pervade the whole and make of the colonial state a “throwback.” He describes the colonial state as “bifurcated”: “organized” from above, it had a dual rather than a hybrid nature—“two forms of power under a single hegemonic authority.”Footnote ¹²

While Fields's “articulation” model of indirect rule better captures the contradictions, contestations, and qualifications of power in the colonial order, Mamdani's deterministic binaries nonetheless illuminate how at least a section of European officialdom relied upon—and was confident that it could understand and control—the agents, practices, and institutions of “custom,” whatever their provenance. Further, although Mamdani devotes more space to the making and persistence of “Native Authority,” his observation that colonial states spoke other “languages” of power that evolved in opposition to, alongside, or in tandem with that of “custom” is valuable.Footnote ¹³ In early twentieth-century South Africa, one key mode and language of power was that of legal-rational bureaucracy. Marxist and feminist analyses identify the instrumental effectiveness and the material and gendered attributes of this mode of power in their documentation of how the state in this period intervened in and interacted with African societies to “free” the land, labor, and other resources that enabled accumulation by settlers and corporations.Footnote ¹⁴ Foucauldian analyses show how this language of power positioned the state to “name” and monitor its “others”; officials took seriously their claims for South Africa to be recognized as a “modern,” “white,” self-governing dominion equal in status to its Canadian and Australasian counterparts—it was neither a “Native state” nor a “Boer state.”Footnote ¹⁵

Marxist, feminist, and Foucauldian analyses explain why white South African administrators may have experienced even more keenly than their counterparts elsewhere the contradiction of upholding a racially exclusive monopoly of power while relying on ever more black intermediaries to provide the control and specialized knowledge of Africans, and to gather the instrumentally valuable information, that an ambitious settler project required. But to understand how officials managed this contradiction, we must look inside the post-1900 state as it was being constructed. We need to locate the protagonists who debated the issues within and across departments, identify what these protagonists and bureaucracies stood for and why they argued and acted as they did, and show how they shaped the more mature state that had solidified by the 1930s.Footnote ¹⁶ For this task a Weberian approach is useful, not to measure the South African state against some externally derived putative norm of rationality or modernity, but rather because Weber's concepts of rationalization and differentiation are particularly relevant in describing and explaining the historical evidence. Weber is also a better guide than are Marx or Foucault to the cultural as opposed to the material and discursive attributes of governance. Finally, administrators at the time justified their claims about the merits of personalized forms of rule versus impersonal bureaucracy in quasi-Weberian terms.

It was evident from the outset that post-1900 South African officials were highly sensitive to the contradictory imperatives to maintain a “white” state and yet also to use increasing numbers of African intermediaries. Regarding the maintenance of a white state, white power was qualified by a reliance on black police, and in 1901 this was bluntly addressed by the British representative in Swaziland (whose administration as part of a broader federated southern Africa under settler leadership remained on the agenda until the mid-twentieth century): “If Swaziland is to remain a native State like Basutoland [a British protectorate entirely surrounded by South Africa], then European police may not be essential but I think if we are to prepare it for incorporation into the Transvaal the less we have of native authority the better.”Footnote ¹⁷ With industry, land, and commerce overwhelmingly in white hands, twentieth-century South African governments endorsed the principle that whites and their property should be policed mainly by their own kind. Therefore, unlike elsewhere in colonial Africa, or in India—where, “since Indians controlled the bulk of the means of production, commerce and capital,” indigenous functionaries, police included, far outnumbered expatriate onesFootnote ¹⁸—in South Africa the majority of rank-and-file police, as well as all commissioned officers, were white. The second governmental imperative was equally apparent to officials: since black intermediaries were indispensable in the surveillance of the African majority, white power was limited without them. It would “always be found necessary,” one administrator insisted in 1904, to operate “through the medium of” black police, who were “the eyes and ears of Government and gather information regarding the Natives which it is essential for Government to know.”Footnote ¹⁹ Whites' voracity for labor and other resources intensified colonial intrusion into Africans' lives and tended to enlarge the number of black police on whom rulers relied. Thus while white police were more numerous, there was always a sizeable black minority; between 1910 and 1939 there were six to eight thousand white police and three to four thousand black police.Footnote ²⁰

Anxious about “native authority,” administrators instituted South African Police (SAP) policies on recruitment, training, deployment, promotion, dress, and the possession of firearms that enfeebled African policemen organizationally within the state and socially among African civilians. Elsewhere I have discussed how black policemen themselves viewed and experienced this precautionary regime, which, while limiting their potential to subvert white power, also impaired official knowledge about African communities and weakened the state's capacity to intervene in them. It thereby institutionalized rather than suppressed the contradictions of rule.Footnote ²¹ Here I am interested primarily in the prior processes that located the task of managing the contradictions in a differentiated police bureaucracy. That this task would fall to the police, or that policing itself, and particularly “information gathering” in black communities, would become largely the responsibility of an institution like the SAP, was not self-evident in 1902, at the beginning of the protracted era of state-building from which modern South Africa emerged.

The governmental imperative to possess specialized knowledge about Africans, without ceding authority to the black intermediaries who gathered it, came to the fore during this period of institutional formation and shaped the bureaucracies that evolved to secure the social order. Two sets of central government officials especially—Police, and Native Affairs—vied for precedence in the supervision of African intermediaries. In what follows I consider the extended debate and competition between them, and touch on cognate developments in urban black residential (“location,” later “township”) and mine policing. Claiming competence in relating as “chiefs” to Africans was a key strategy in these early-twentieth-century rivalries. Such claims testified simultaneously to a number of things: to the importance that the emerging South African bureaucracy attached to the expertise typically valued in “modern” states; to state officials' perception of the continuing relevance of patriarchal idioms in relations among the governed; and to the incorporation of these idioms into the workings of government. How far such incorporation qualified bureaucratic rationality in the ostensibly “modern” South African state, and with what implications, are this article's larger questions.

BUREAUCRATIC RIVALRY FOR CONTROL OF AFRICAN POLICE

The debate began as British officials started setting up government departments in the former Boer territories after the South African War. Godfrey Lagden, the Transvaal's Commissioner for Native Affairs, and Robert Baden-Powell, the Inspector-General of the South African Constabulary (SAC), staked competing claims to leadership in controlling African police. Lagden viewed ruling Africans as delicate and expert work that his department alone was competent to supervise. Jingoistic British rhetoric denigrating the amateurishness, venality, and brutality of their “rotten and corrupt” Boer republican predecessors' handling of “Native Affairs” only lent influence to this claim.Footnote ²² Alfred Milner, Britain's High Commissioner, had set the tone in reviling “central office men, ignorant of native life, issuing general orders to local administrators of the second-rate kind, having no personal hold of the natives.”Footnote ²³ In 1903, Milner appointed an inter-colonial South African Native Affairs Commission, “a body of experts” that Lagden chaired, tasked with considering “native policy” in subcontinental perspective after gathering “accurate information.”Footnote ²⁴

The contradictions are already evident in these early pronouncements and actions. A “modern bureaucracy” values “the expert, who is … indifferent in human terms, and so all the more completely ‘objective.’”Footnote ²⁵ But colonial officials had difficulty imagining an expert on “native life” as someone “having no personal hold of the natives.” Milner wrote quite explicitly that he was “afraid of the native administration getting too bureaucratic.… [T]he sort of men whom we should wish to have … [are] men of high character and a personal gift of command, like those, who, in India and elsewhere, have made British government of natives a success by their individual influence over the people among whom they lived.”Footnote ²⁶ Milner had in mind the stereotypical “old India hands” of the “Punjab school of administrators … who emphasized personalistic rule as the key to social and political order” in a society viewed as “deeply rooted in patriarchal, tribal authority,” and who were credited with having rescued the British position in the 1857 Mutiny. Perhaps Milner thought such chiefly qualities were particularly valuable when embarking on state-building in a disturbed post-conquest context, but in most parts of India itself British rule had for a century been moving progressively from dependence on “human” forms of knowledge to reliance on more “institutional” forms.Footnote ²⁷

African police had a pivotal role in both Milner and Lagden's conceptions of how to govern Africans expertly. Milner, unlike later settler politicians beholden to constituencies of poorer whites who lacked employment opportunities, was prepared to dismiss costlier white constables to pay for a requisite number of black policemen, and indeed he thought “Natives could be more utilised than they are … in Police work.”Footnote ²⁸ Lagden, as Resident Commissioner in Basutoland before the war, had favorable experience of a purely African police force, and had supplied Basotho recruits to the Bechuanaland Protectorate Police.Footnote ²⁹ For him, Baden-Powell's white constabulary were not only expensive, but more importantly, being mostly “perfect strangers to the country,” were also “useless for matters connected with Native affairs.” Without black police, Lagden opined, “the close touch that ought to be maintained between the Natives and the Government would be entirely lost.” And his own department's Native Commissioners, he emphasized, were the appropriate officials to manage African policemen “to the best advantage in the maintenance of law and order.” Lagden clearly had Baden-Powell's SAC in mind in warning “that many of the wars and troubles in the history of South Africa have arisen from the use of inexperienced and unsympathetic persons dealing with the Native population.”Footnote ³⁰

Lagden had reason to fear the encroachment of Baden-Powell's SAC, a force of over ten thousand men, initially under the command of the British Army, that became the principal policing and occupying institution in the South African interior after the war.Footnote ³¹ Colonel Sam Steele, the towering Canadian who commanded the SAC's Northern Transvaal “B” division, energetically described the force's “information-gathering” ambitions in his memoirs: “There was not,” he wrote, “a Kaffir kraal, or chief, nor a burgher before the war who was not on the rolls and his character and opinions known to us. As I said to a friend of mine who was a resident magistrate … and objected to our placing the posts ourselves at the outset, ‘Though you know the district, in three weeks’ time we will be better acquainted with it than even the Boers and Kaffirs.'”Footnote ³² Baden-Powell needed his own African police for such work, but he wanted control of the discipline, pay, and provisioning of Lagden's, too, arguing that it was not “conducive to efficient Police or Intelligence work to have different police forces working under different heads in one country.” His own officers did not want Native Affairs Department (NAD) interference, and he claimed that the “system of Native Commissioners police” had failed in Rhodesia and Zululand.Footnote ³³

Lagden, fearing that SAC control would “impair” the “usefulness” of black policemen to his Native Commissioners, held out for a separate NAD police.Footnote ³⁴ Milner, partly to redistribute the politically controversial fiscal burden of policing, but also because he accepted this argument, adjudicated in Lagden's favor and allowed the NAD to raise its own force of two hundred African policemen while leaving the SAC in command of an additional 620 Africans.Footnote ³⁵ We know that the SAC conceded reluctantly, however, because three years later, in 1905, J. S. Nicholson, Baden-Powell's successor as Inspector-General, observed that the question had “been discussed over and over again.”Footnote ³⁶ Transvaal Native Commissioners themselves sometimes complained that their African constables were “very slack” and should be subject to regular police discipline and sanctions. More than once during the Crown Colony period the Transvaal NAD solicited model regulations on “Native Police” from other South African colonies, though it apparently never promulgated its own.Footnote ³⁷ The NAD's acknowledgment of the unreliability of its police undermined its claims that expertise in “Native Affairs” was relevant to policing. In 1907, following the return of the Transvaal to settler self-government under former Boer generals Louis Botha and Jan Smuts, the principle of separate control was reconsidered.

The initial impetus came from the problem of African tax evaders who, resisting arrest by NAD constables, claimed that the latter were not legally “authorised Police.” Native Commissioner C. A. Wheelwright, alarmed to learn he could not prosecute for this opposition, and might even be liable for damages for illegal arrest, felt the resistance could “lead to serious effects” and urged Pretoria to empower NAD messengers “to act as Police in an unquestionable way.” Wheelwright knew that he was reopening the jurisdictional issue, for he was at pains to stress the inadvisability of “alienat[ing] the present control by the Native Affairs Department.” Echoing Milner's emphasis on personalized command, he considered it “most essential in treating with Natives that the responsible Officials should thoroughly know the Native Police to be employed by them in connection with the carrying out of their duties, and therefore they should have the principal say in the engagement and employment of men in particular districts.”Footnote ³⁸ William Windham, Wheelwright's superior, agreed that the NAD police were “in a very invidious and weak position,” and hoped they might be sworn in as “Peace Officers.” Windham was aware that the new government wanted to reorganize policing in the colony, and that the NAD might lose exclusive control of its constables. He insisted that in any system of shared control his district officers should recruit and supervise Africans doing NAD duties, and that this approach was demanded by the sensitive “political and administrative” work these constables performed as NAD “intelligence agents.”Footnote ³⁹

In late 1907, the Transvaal Cabinet, following reports of a committee of magistrates appointed to recommend economies in the civil service, decided to bring all African constables under the unified police force it was planning for the entire colony.Footnote ⁴⁰ In those districts with few white settlers, where, according to the committee, Africans “still live[d] under the old tribal conditions,” the NAD's Sub-Native Commissioners (closely supervised by Law Department magistrates, who elsewhere already doubled as Native Commissioners) would continue to rule through recognized chiefs with an allotted number of “Native Police,” who would be paid and disciplined by, and receive their direct orders from the local police station commander. Resident Magistrates and police District Commandants would decide who were “the most efficient and trustworthy” black constables to retain in the first instance, although Sub-Native Commissioners subsequently were given a say in selecting new recruits locally.Footnote ⁴¹ This shift demoted the NAD's status, and devalued its claims that its special expertise and the leadership qualities of its personnel justified its exclusive control of African police to, as Lagden had put it, “the best advantage in the maintenance of law and order.”

When the Union of South Africa was formed in 1910, discussions began on amalgamating the several colonial police forces, and officials revisited many of the earlier Transvaal arguments. Cape civil servants taking up senior positions in the new Union NAD were interested and vocal participants in these discussions. In 1912, the NAD independently controlled 67 white and 379 African police and detectives in the Transkeian Territories. This was an area reserved for African occupation where the former Cape administration's ethos of “benevolent paternalism” was well entrenched.Footnote ⁴² Transkeian Chief Magistrate Arthur Stanford objected to policing “by a department without a specialized knowledge of Native conditions or the particular requirements of these Territories.” African police, he held, were specially chosen and trained to perform politically sensitive “missions” for magistrates, and therefore “the Magistrate—not a police commissioner or inspector—must be their ‘chief.’”Footnote ⁴³ Here too, then, specialization and the “division of labour in administration” were thought to depend on less rather than more depersonalization.Footnote ⁴⁴ Secretary for Native Affairs Edward Dower (a missionary's son who had worked under Stanford's older brother Walter in the Cape NAD) agreed that even if there was to be only one police force for the Union his department should jointly direct those sections serving “in the Native Territories.”Footnote ⁴⁵ Colonel T. G. Truter, who before Union had been a Resident Magistrate in the Transvaal, and was now charged with organizing and heading the new South African Police (SAP), not surprisingly declared in favor of the Transvaal system of police precedence.Footnote ⁴⁶

Truter's SAP was not, however, immediately given control of African police “in the Native Territories.” In 1912, Jan Smuts, who had overseen the earlier Transvaal civil service reforms and was in charge of administrative planning in the early Union Cabinets of Prime Minister Louis Botha, assigned all existing police in “Native Territories” to the South African Mounted Riflemen (SAMR) under the Department of Defence.Footnote ⁴⁷ Smuts reportedly was “most anxious” to alleviate NAD concerns about these changes, particularly in the Transkei, and he instructed senior SAMR officers to cooperate with the Chief Magistrate.Footnote ⁴⁸ However, although the NAD still had to pay for the “Native Police” assigned to its field officers, the principle of ultimate institutional control over African policemen was provisionally settled in favor of the SAMR and SAP.

In November 1914 there was determined African resistance in the Transkei's East Griqualand districts to state initiatives to combat stock disease. This followed soon after the transfer of police control, and seemed to a former Chief Magistrate to exemplify the consequences of the usurpation of the personalized style of magisterial authority by “Officers of various Departments … responsible to some other officer stationed at Pretoria or elsewhere and unknown to the people.”Footnote ⁴⁹ His protest echoed Milner's pronouncements at the turn of the century about “central office men” and “native administration getting too bureaucratic,” but no reversal of policy ensued. When the SAMR regiments were mobilized during World War I, the SAP temporarily took over policing in their areas. The pre-1914 system was briefly reinstated at war's end, but in 1920 most of the SAMR regiments were disbanded and the SAP was installed countrywide. Until the reorganization of 1920, the NAD would not concede the likelihood of its permanent eclipse and it continued to aspire to a policing function. Its regional offices monitored African political sentiment, and reported on it to Pretoria, independently, particularly so the Johannesburg office, which, concerned about African “unrest” in the country's economic heartland, set up a small “Native Intelligence Bureau” toward the end of the war.Footnote ⁵⁰

CULTURE, LANGUAGE, AND THE RACIAL POLITICS OF POLICING

Within the SAP itself, the conclusion of its rivalry with the NAD would lead to an attenuation of the patriarchal discourse of expertise in controlling Africans as “chiefs,” and to a decline in the rank of the most credible exponents of this discourse. While the rivalry persisted, however, some very senior police officers were its prominent theorists, particularly in the Criminal Investigation Department (CID), where “Native Crime” was considered a specialist category.Footnote ⁵¹ In 1908, T. E. Mavrogordato, the Transvaal's senior detective, profiled his ideal candidate to head the “Native Police” on the Witwatersrand. This officer, Mavrogordato contended, “should possess qualifications which are rare in this country.” Besides English and Dutch, he would know “several Kaffir dialects” and be “thoroughly conversant with the ways of the different Tribes in South Africa,” since only then would “the Natives look up to him as a white chief.”Footnote ⁵² S. R. Brink, a leading officer in the Gold and Diamond Detective Department, offered a similar job description in 1912: the candidate would “not only be thoroughly conversant with the languages and customs of the natives, but also be one who has organising powers, and who would inspire the natives with fear and respect.”Footnote ⁵³ The detective who assumed this position in Johannesburg, “supervis[ing] generally the Native Crime throughout the Transvaal,” was A. J. Hoffmann, whom Truter in 1916 described, in seeming paraphrase of Mavrogordato, as “specially adapted for this work being able to speak English, Dutch and Native languages in addition to having a thorough knowledge of the ways and customs of Natives.”Footnote ⁵⁴

But Hoffmann, who, like many experienced detectives, was a familiar personality to Africans, would never acquire the seniority of rank that Brink and Mavrogordato had envisaged, nor would his staff get the resources and degree of specialized training they had thought necessary.Footnote ⁵⁵ As the need to compete with the NAD on the discursive terrain of expertise in African affairs receded in the 1920s, senior SAP managers increasingly articulated a strident rhetoric that associated whites who had knowledge of African cultures and languages with lower-class status and untrustworthiness—“a poor education” and “deficien[cy] in other qualities which go to the making of a good Policeman.”Footnote ⁵⁶ By the end of his career, Hoffmann had “been prominently identified with many extraordinary cases in which natives were the delinquents,” but this did not improve his prospects for promotion to commissioned rank.Footnote ⁵⁷ I. P. de Villiers, Truter's successor as Commissioner of Police, contemptuously referred to white policemen who spoke African languages fluently as having “the native mentality.”Footnote ⁵⁸ Such attitudes ensured that before World War II no commissioned officer ever headed the “native staff” of a city CID. One way of curtailing the danger that “native authority” might colonize the state was thus to marginalize, subordinate, and minimize the number of those whites reputed to be particularly well-informed about African life and culture; if there were to be “white chiefs” in the police bureaucracy exercising personalized authority over black intermediaries in the way Mavrogordato had proposed, they were going to be very petty chiefs indeed.

This approach in one respect lessened the possibility that dependence on African intermediaries might qualify white power, but it did so at the cost of moderating the state's “information-gathering capacity” and thus limiting white power and even bolstering “native authority.” The evidence is compelling that dependence on white police who could not communicate with Africans impeded colonial control, and that white officials knew it. In 1921 a magistrate in Pondoland in the Transkei attributed lax pursuit of tax defaulters to the “difficulty … that the majority of European Police cannot speak the Native language and the Native constables cannot speak English or Dutch, and it is useless sending out a mixed patrol on this account.” The police District Commandant endorsed his analysis.Footnote ⁵⁹ In 1925 a Transvaal Lowveld employers' society complained of white policemen who did not know enough of “the native language even to ask [their] way,” and of inquiries that relied entirely on “the native interpreter … who has in effect … more power than a magistrate, or judge and jury combined.”Footnote ⁶⁰ In 1927 there were forty police stations in Natal where the Post Commander was “entirely dependent on his Native Police” for interpretation.Footnote ⁶¹ Since Post Commanders were also responsible for training African police, “their lack of knowledge of the Zulu language [rendered them] unable to afford this effectually.”Footnote ⁶² A long-serving Transkei Deputy Commissioner unfavorably compared recruits in the 1920s to the pre-Union Cape Mounted Police. The former were “100 per cent Afrikaans speaking men, from the Transvaal and Free State,” and “entirely unsuited” for work “among the Natives.”Footnote ⁶³ His successor agreed that “the good Kaffir speaking European Constable” was “on the decrease,” adding that low pay attracted blacks with little English, who themselves thus needed interpreters to report to white policemen.Footnote ⁶⁴ An NAD official cited several cases in which the outcome had been influenced by African police taking advantage of white constables' ignorance.Footnote ⁶⁵ Lower-ranking white policemen themselves feared becoming “the glorified clerk of [the] Native constable.” “The only thing [a non-linguist] can do,” a sergeant complained, “is to cancel circulars and Gazettes without an interpreter.”Footnote ⁶⁶ Petty bureaucrats, it seemed, were preferable to petty chiefs.

As indicated by I. P. de Villiers's contempt for white policemen having “the native mentality,” these vulnerabilities in the state's “information-gathering capacity” were partly a result of deliberate policy; only the most desultory efforts were made in the interwar years to remedy them by attracting, training, or retaining white police equipped with the desired skills. For all the vocal concern about the status of information mediated by Africans, white recruits received no special African language training during their four to six months at the Police Training Depot in Pretoria, nor even, a 1937 commission noted, preliminary instruction (for what this would have been worth) in “the mentality of the vast non-European population of the Union.”Footnote ⁶⁷ But these vulnerabilities also arose from the creeping bureaucratization associated with state formation and reorganization that Milner and others had warned against earlier in the century as being inimical to successful “native administration.” The amalgamation of the pre-Union policing forces, the centralization of white recruits' training in Pretoria, the increased likelihood of their being sent to parts of the country unfamiliar to them, greater opportunities for promotion, and frequent transfers—all placed a greater premium on depersonalization and the interchangeability of functionaries. In the words of an officer who had been an instructor at the Training Depot in the 1930s: “What is the earthly use of teaching a policeman about black ethnicity and all of that when you're going to send him to Pofadder?”Footnote ⁶⁸

The compromises arrived at within the SAP to manage the contradictions between the imperatives to both sustain an effective “information-gathering capability” and minimize “native authority” thus involved casting off and devaluing the police's prior claims to expert knowledge of African societies. Recognition across state departments of the SAP's primacy in “intelligence” might confer authority on it, but would not guarantee the quality of the information collected. Following the national reorganization of policing in 1920, and the conclusion of its rivalry with the NAD for the control of African police, the SAP progressively acquired a monopoly in the production and distribution of “information.” By the mid-1930s, the Secretary for Native Affairs had to ask the Commissioner of Police to continue sending him regular political reports because he was “anxious to keep in touch” with African opinion.Footnote ⁶⁹ As this suggests, the personalized, patriarchal model of interacting with Africans was in decline even within the NAD.Footnote ⁷⁰ More NAD police were permanently assigned as their duties became increasingly clerical and routinely bureaucratic. There were over seven hundred countrywide in 1929, and 842 by 1936.Footnote ⁷¹ In addition to court work, issuing passes, collecting revenue, and communicating government policy, NAD police assisted in land allocation, health and vaccination campaigns, and supervising veterinary regulations. This specialization rendered NAD constables mostly unavailable to the SAP, yet imposed much administrative work on it. Even senior police officers were led on occasion to urge that the force divest itself altogether of the responsibility for controlling them. Officers regularly suggested that NAD constables' easier civilian working environments, and their tendency to be better educated, made them difficult to discipline and fostered discontent among the ordinary African policemen with whom they shared quarters and messed. “The Native Constables of the N.A.D.,” complained a Deputy Commissioner in 1931, “are an undisciplined body, shewing defiance and causing endless trouble. Some of them I have no doubt are members of the I.C.U. [the Industrial and Commercial Workers Union, a leading black opposition movement of the 1920s].”Footnote ⁷²

But the outcome of the arguments hard-fought earlier in the century was not so readily abandoned. The white state wanted its black intermediaries kept under the most stringent disciplinary sanctions, which as members of the SAP, subject ultimately to “semi-military control,” they were.Footnote ⁷³ Complaints about NAD constables' indiscipline invariably resulted in a tightening of the regulations governing their employment and conditions of service: insistence that they wear full uniform; attend roll calls, inspections, and parades; do their share of menial labor; and reside in barracks. The only outward mark distinguishing NAD from regular policemen was a badge they wore on their left arm bearing the letters “N.A.D.,” which was introduced in 1924.Footnote ⁷⁴ This explicit subjection of NAD police to the coercive rather than the technically expert arm of the state underscored the administration's determination to limit the subversive potential that it had long feared from its reliance upon “native authority.”

The larger story here transcends the vicissitudes of a protracted and narrowly bounded interdepartmental squabble; it is about institutional rationalization and differentiation, complicated by a subplot in which officials generally posited their ideal relations with the Africans over and through whom they ruled as conforming to a patriarchal model of personalized authority or chiefship. This model was at odds with the impersonal ethos of evidence-based argumentation that, as in other modern bureaucracies, conventionally informed and legitimated communication among early twentieth-century South African administrators themselves.Footnote ⁷⁵ Yet, the evidence of officials' invocation of patriarchal idioms complicates rather than contradicts a narrative of wider, systemic processes of bureaucratic rationalization at work in this period. This can be illustrated by briefly documenting cognate developments in questions of policing policy beyond the limited issue of who should control African police employed by the central government.

AFRICAN INTERMEDIARIES IN THE “RESERVES,” URBAN LOCATIONS AND MINES

Consider, first, hereditary chiefs and headmen, who exercised significant everyday policing functions in the African “reserve” lands, but who in doing so were subject to the same differentiating, departmental, competitive pressures as were African police. For example, they increasingly reported police matters separately to the SAP, despite for all other purposes being under the NAD—a development that the Transkeian magisterial establishment viewed as a further usurpation of its authority, detrimental to successful rule.Footnote ⁷⁶ At a Union-wide level, from 1917 the Minister of Justice could designate headmen of “rural native locations” to be “peace officers” with police powers of arrest.Footnote ⁷⁷ Regulations framed under the 1927 Native Administration Act went further still in endowing all appointed chiefs and headmen in the “reserves” with formal police powers; they specifically excluded “crime and police administration” from chiefs' and headmen's usual obligation to report to the government solely through the NAD's Native Commissioners. To secure headmen's compliance they were promised £1 a month, rising to £3 after sixteen years of service.Footnote ⁷⁸ Historians have argued that the 1927 Act elevated the NAD's profile within the state, but chiefs' and headmen's formal assumption of policing duties in the interwar years did not restore to the NAD the precedence it had lost to the SAP in the struggle over control of African policemen.Footnote ⁷⁹ The SAP's opinions were notably decisive, for example, in adjudicating chiefs' and headmen's requests for permission to possess firearms, “notwithstanding,” as one Chief Native Commissioner put it sourly, “that the possession of fire-arms is of definite significance in Native ways of thought.”Footnote ⁸⁰ Here again, claims to possess specialized knowledge of Africans no longer conferred authority in what was now more clearly circumscribed as the domain of policing policy.

African commentators confirmed the growing identification of chiefs and headmen with the SAP. “The correct definition of a headman,” argued Transkeian Territories General Council member Siroqo in 1936, “is a constable…. The only reason why he is not recognised as a policeman is that he lives and sleeps at his kraal.”Footnote ⁸¹ A year later, SAP Commissioner I. P. de Villiers spelled out the theory of policing through headmen: “Even with the ordinary police station in the Transkei the man who does the real police work as such is the native chief. If a [white] sergeant is an intelligent man, apart from routine duties he has only to inform the native chief who will present him with the criminal and with all the evidence too.”Footnote ⁸² We would expect, however, that headmen, being subject to less direct invigilation than “slack” African constables, also enjoyed greater latitude to obstruct white authority.

The Commissioner's comment is remarkable for its absence of ethnographic curiosity about the internal workings of the machinery of indirect policing. Local white detectives were less confident than De Villiers that headmen were sufficiently tied by external controls and incentives to the administration to deter them from trying cases themselves and receiving bribes rather than bringing the information to the nearest police post (which could require a considerable journey). The result, thought one, was that crime was “being compounded” and police investigations hampered.Footnote ⁸³ Yet most station commanders trusted headmen, and instructed African constables to go to them for guidance in cultivating informers, which headmen welcomed as opportunities to dispense patronage. One former black policeman told me that chiefs and headmen “were quite reliable people, because they are the law-abiding people too…. They are well paid by the government.”Footnote ⁸⁴

In the decades after 1900 the tensions between a decentralized patronage politics of everyday local policing and the concentration of police control within differentiated, hierarchical organizations were further evident in the policing of urban locations and mine compounds, which also depended heavily on African intermediaries. Municipality-appointed superintendents throughout the country were legally responsible for regulating specific locations, and their budgets allowed a number of African “headmen” or constables for this purpose. In the late 1930s, older SAP and municipal officials nostalgically recalled an earlier era of cooperation when there had been “no differentiation between [their] two departments” in policing urban Africans.Footnote ⁸⁵ Although this claim was exaggerated, it did fairly indicate that the SAP assumed more and more of this work as the legislature instituted more, discriminatory pass, tax, and especially liquor laws, whose enforcement location superintendents shunned as “a difficult, unpopular and dangerous job.”Footnote ⁸⁶

The post-Union jostle for departmental aggrandizement only encouraged this tendency. Hence Truter urged in 1912 that the “best locations” were those where the municipality relied “entirely on Government police for the prevention of disorder and the suppression of crime generally.” Local governments were seldom reluctant to cede control of their locations to the SAP since doing so was in their fiscal interest, but it cost them significant oversight powers.Footnote ⁸⁷ Increasingly in the interwar years black municipal police focused on guarding town property, collecting rents, and other duties analogous to those performed at the national level by NAD constables.Footnote ⁸⁸

The mining industry was the most significant private employer of African police, who before the 1920s exercised a brutal and venal regime over the workforce housed in compounds on mine property.Footnote ⁸⁹ In 1903, Lagden, one of whose chief responsibilities as Transvaal Commissioner for Native Affairs was to facilitate the mine labor supply, was alarmed to learn of “a story running riot” among Cape Africans that they would not be safe in Johannesburg because “the Zulu Police on the mines” were ill-disposed towards them.Footnote ⁹⁰ This reputation led the industry, with NAD prompting, to make their compound police forces more ethnically diverse.Footnote ⁹¹ Racketeering continued, however, and the CID accused mine compound managers and their police of collusion with illicit liquor dealers, a charge Truter also leveled against municipal black policemen.Footnote ⁹² Since the compounds were private property, there was little likelihood of the SAP taking over their policing directly, but soon after Union, Truter advocated “much stricter supervision of the mine police,” and he sympathized with calls for them to “be under the control of some white Government official.”Footnote ⁹³ The matter was urgent because, under the Native Labour Regulations, mine police in executing compound managers' dictates exercised state police powers.Footnote ⁹⁴ In the meantime, Truter pressed the outward differentiation of the two forces to a logical conclusion by insisting that the mining groups outfit their compound police with uniforms dissimilar to those worn by black SAP members, although this likely did little to change black mineworkers' perception of their functional similarity.Footnote ⁹⁵

What is striking about the subsequent development of policing within the differentiated sphere of the mines is how closely it resembled the drama being played out on the broader South African stage between the NAD and SAP, but with compound managers and mine police departments as the protagonists. The larger point is that the same forces promoting and limiting rationalization were organizing institutional arrangements similarly throughout the social order.

Before World War I, some of the mining corporations began hiring private agencies, such as G. Bush and Company, staffed by former CID members, to monitor gold theft and illicit liquor dealing on their properties.Footnote ⁹⁶ By 1914, a few trend-setting companies like the Crown Mines had created their own detective departments.Footnote ⁹⁷ In the interwar years the differentiation of mine policing beyond simple compound control gathered pace as the industry evolved large and highly structured detective organizations, managed by retired senior CID officers, with responsibility for all the mines within a corporation.Footnote ⁹⁸ In the early 1940s there were four such departments with combined staffs of 142 whites and 1,887 Africans (a force the size of the SAP's entire Natal provincial division), controlling fifty-three mines.Footnote ⁹⁹ By 1945, the manager of one mine police force called “unsatisfactory” the fact that on some “smaller Mines” the compound manager was “also the Police official.” Clearly this was no longer the situation on most mines.Footnote ¹⁰⁰

The mine police departments' brief to prevent gold theft, guard mine property, and keep liquor out of the compounds grew steadily into what the radical senator H. M. Basner termed, in a statement that provoked a celebrated libel suit, “an extensive and efficient espionage system” with “the worst features of American labour racketeering.”Footnote ¹⁰¹ Basner's antagonist was A. E. Trigger, a former Witwatersrand CID chief in charge of the largest of the mine forces, who issued instructions prescribing surveillance of “labour unrest” and “communism.” These orders licensed considerable intervention in compounds and workplaces.Footnote ¹⁰² Much as NAD officials had yielded to the SAP, then, compound managers in the interwar decades ceded to new mine police organizations some of their control over the recruitment and supervision of African mine police. Trigger, for example, whose SAP experience left him with the impression that unsupervised compound managers and compound police abused their authority, strove to lay down rules detailing the linguistic and cultural knowledge compound managers should possess, and the conduct they should observe, so as to secure respect as a “chief” and avoid being overly dependent on their “Induna and the police boys as being [their] eyes and ears.” There were constraints, however, on the mine police departments' superintendence and rationalizing colonization of compound managerial patriarchalism, most notably the limits entrenched by the greater legal powers that compound managers enjoyed to “maintain law and order” in the compounds. Thus mine police departments screened compound police applicants, and cooperated with, supervised, and reported on, but never ultimately displaced, compound managers' policing function.Footnote ¹⁰³