ZA-K-01: The 2007 Polokwane Decision and the Mbeki Recall (2005–2008)

Status: DRAFTWords: 10,326

1. Key Takeaways

  • Polokwane and the recall are best read as a single two-stage decision, not two events. At the African National Congress's 52nd National Conference at the University of Limpopo's Turfloop campus outside Polokwane (16–20 December 2007), approximately 4,000 branch-mandated delegates elected Jacob Gedleyihlekisa Zuma over the incumbent State President, Thabo Mvuyelwa Mbeki, as ANC President — 2,329 votes to 1,505, a 60.7 to 39.3 per cent margin — the first defeat of a sitting leader in the party's then-95-year history. Nine months later, on 19–20 September 2008, the Zuma-aligned National Executive Committee elected at Polokwane "recalled" Mbeki from the state presidency, eight months before the end of his constitutional term. Stage one transferred the party; stage two demonstrated that whoever holds the party holds the state. The pair of decisions is the hinge of post-Mandela South African politics, and every subsequent presidency — Zuma's, Ramaphosa's — has operated inside the doctrine the pair established.

  • The decision arc began with Mbeki's own calculated risk: the 14 June 2005 dismissal of Deputy President Zuma. On 2 June 2005 the Durban High Court (Squires J) convicted Zuma's financial adviser Schabir Shaik on corruption and fraud counts arising from payments to Zuma and the solicitation of a bribe from the French arms company Thomson-CSF/Thint, sentencing him to fifteen years [TBD-VERIFY: the phrase "generally corrupt relationship," universally attributed to the Squires judgment, does not appear verbatim in it; Squires stated publicly in 2006 that he had not used the words — the formulation derives from press summaries, and its substance was later endorsed on appeal]. Twelve days later Mbeki told a joint sitting of Parliament that he had decided to "release" Zuma from his duties as Deputy President of the Republic. The dismissal was constitutionally impeccable and politically catastrophic for its author: it converted Zuma from a compromised deputy into a martyr around whom every constituency alienated by the Mbeki presidency could assemble. The ANC National General Council of 29 June–3 July 2005, which restored Zuma's full functions as ANC Deputy President against the leadership's intentions, was the first measurable signal that the branches would not follow Mbeki against Zuma.

  • The Zuma coalition was a negative coalition — assembled less by what it wanted than by what it opposed. Its components were: COSATU and the SACP, alienated by GEAR-era macroeconomic orthodoxy and by their marginalisation under what SACP General Secretary Blade Nzimande's party had since 1998 called the "1996 class project"; the ANC Youth League under Fikile Mbalula, which declared for Zuma early and supplied the campaign's mobilising theatre; a cohort of provincial and ex-prisoner politicians sidelined by Mbeki's centralised deployment machinery; and Zuma's own KwaZulu-Natal base. The 2005–06 prosecutions became the coalition's mobilising engine rather than its liability: the rape trial (acquittal, 8 May 2006) and the corruption case (struck off the roll by Msimang J, 20 September 2006) were reframed by the coalition as instalments of a political conspiracy run from the Union Buildings — the victim-of-conspiracy narrative that COSATU General Secretary Zwelinzima Vavi captured in his "unstoppable tsunami" formulation [TBD-VERIFY: Vavi's June 2005 wording is variously rendered; the common paraphrase is that stopping Zuma's rise to the presidency would be "like trying to fight against the big wave of a tsunami"]. The press coinage "Zunami" registered the phenomenon: legal jeopardy converted into political momentum.

  • Mbeki's decision to seek a third term as ANC President — the move Polokwane punished — rested on a systematic misreading of the branches. Barred by the Constitution from a third state-presidential term, Mbeki stood for a third party term on the logic that the ANC presidency would let him manage his own succession and protect his legacy (and, his critics held, remain the power behind a chosen state president). The June 2007 ANC Policy Conference signalled branch hostility to the "two centres of power" arrangement; the November 2007 provincial nomination conferences gave Zuma a clear majority of branch nominations [TBD-VERIFY: precise nomination tallies; contemporary reporting placed Zuma's share at roughly 60 per cent, consistent with the eventual conference result]; and Mbeki's own October 2007 KwaZulu-Natal campaign tour played to thin halls while Zuma filled stadia. Mbeki's camp ran on incumbency, governmental record, and control of the deployment machinery — precisely the assets the branches had assembled to vote against. ZA-B-02 carries the full procedural record; the analytical point here is that the third-term gambit transformed a succession contest Mbeki might have managed into a referendum on his presidency that he could only lose.

  • What the branches voted against at Polokwane is better documented than what they voted for. The conference's iconography — the booing of Mbeki and his ministers from the floor, the "substitution" hand-rolling gesture borrowed from football crowds, the singing of Umshini wami — expressed a revolt against an accumulated indictment: the imperial-presidency style and the centralisation of appointment power; the HIV/AIDS denialism whose mortality cost is treated at ZA-G-01 (the Harvard School of Public Health estimate attributes roughly 330,000–365,000 preventable deaths to the 2000–2005 ARV delay [TBD-VERIFY: the Chigwedere et al. 2008 figure and period]); the "quiet diplomacy" on Zimbabwe; and the redeployment machinery's long tail of resentment among those it had passed over or purged. What the delegates voted for divides into three readings that this document holds at strength in Section 5: a leftward economic correction (the COSATU/SACP reading), a redistribution of patronage from one network to another (the structuralist reading), and a personalist transfer to a more accessible, less remote leader (the political-culture reading). The Polokwane resolutions were in fact largely continuous with existing policy; the rupture was about power, not programme.

  • The 12 September 2008 Nicholson judgment supplied the trigger the Polokwane NEC had been waiting for — and its later reversal demonstrated how thin the recall's stated justification was. Nicholson J's procedural holding (that the NPA's December 2007 recharging of Zuma was invalid for failure to take representations) was narrow and curable; the judgment's political force lay in obiter passages inferring that the prosecution had been "politically directed," naming Mbeki and former NDPP Bulelani Ngcuka without their having been parties or evidence having been led. Within forty-eight hours the SACP, COSATU, and the KwaZulu-Natal ANC were demanding Mbeki's head; the NEC convened at Esselen Park on 19–20 September 2008 and resolved to recall him. On 12 January 2009 the Supreme Court of Appeal (Harms DP) overturned Nicholson comprehensively, rebuking the political-interference findings as unsustainable — but Mbeki was four months gone, and no restoration mechanism existed or was contemplated. ZA-B-03 carries the full judgment-to-inauguration sequence; the decision-analysis point is that the recall's formal predicate was destroyed in court within four months and the recall stood anyway, confirming that the judgment had been occasion rather than cause.

  • Mbeki's compliance was itself a decision — the constitutionalist exit that stabilised the precedent it submitted to. The NEC had no constitutional power to remove a state president; the recall was a party instruction that acquired legal effect only because Mbeki chose to obey it. His televised resignation address of Sunday 21 September 2008 was deliberate in register: loyalty to the ANC's decision, a point-by-point defence of his record, an explicit denial of interference in the Zuma prosecution, and not one sentence of constitutional resistance. Parliament elected Kgalema Motlanthe (ZA-H-PRES-03) as caretaker president on 25 September 2008, and the Motlanthe interregnum (to 9 May 2009) was conducted with deliberate institutional sobriety — including the reversal of the Mbeki-era HIV/AIDS posture. The immediate organisational cost was the Congress of the People split (founded 16 December 2008; 7.42 per cent and 30 seats in April 2009; effectively defunct within a decade), whose failure validated, ex post, the Polokwane coalition's bet that an Mbeki-loyalist breakaway could not break the ANC's majority.

  • What the recall settled was the party-state question: the state president is the ANC's deployee, holding office on the party's sufferance. The 1996 Constitution makes the president accountable to the National Assembly; the recall demonstrated that, so long as the ANC commanded the Assembly, accountability to the party's NEC was the operative mechanism and the Assembly a notarising formality. The "deployee-president" doctrine has disciplined every president since: Zuma was himself recalled by the NEC on 13 February 2018 and resigned on 14 February (a symmetry treated in Section 6); Ramaphosa's December 2022 survival of the Phala-Phala Section 89 report was decided in the ANC NEC and caucus before it was decided in Parliament [TBD-VERIFY: the 13 December 2022 Assembly vote tally against adopting the panel report]. Whether the doctrine is an accountability mechanism (parties disciplining presidents between elections) or an instability mechanism (presidents governing for internal-party survival rather than for the electorate) is the recall's central contested legacy, treated in Sections 5 and 7.

  • In the state-capture genealogy, Polokwane is the enabling decision — but the corpus holds the enabling reading and the overdetermination reading in tension rather than declaring a winner. The enabling reading, developed against ZA-J-01's structural account, runs: Polokwane installed a leadership whose path to power ran through legal jeopardy and whose coalition was held together by patronage obligations; the 6 April 2009 withdrawal of the Zuma charges and the post-2009 capture architecture (ZA-C-01, ZA-C-02) followed from the 2007 victory as project follows charter. The overdetermination reading replies that the capture vulnerabilities — cadre deployment, conference-delegate purchasability, SOE board patronage, the criminal-justice system's exposure to executive pressure — predated Polokwane, were built partly under Mbeki, and would have been available to any winner; Polokwane selected the occupant, not the architecture. The honest synthesis is that Polokwane was necessary but not sufficient for what followed, and that the conference system's demonstrated capture vulnerability — branches as purchasable currency — is itself among the decision's most consequential revelations.

  • Comparatively, Polokwane is the rare dominant-party succession rupture that the party machinery absorbed rather than externalised. Kenya's KANU imposed Moi's chosen successor from above in 2002, split, and lost power within months (KE-K-01) — the contrast case of a dominant party that could not metabolise an internal revolt and collapsed. UMNO expelled its 1998 insurgent (Anwar Ibrahim) rather than let him contest, externalising the rupture into two decades of Reformasi opposition. The ANC, by contrast, let the insurgency win inside the party: the machinery survived, the coalition stayed under one roof, and the electoral majority held for another seventeen years. But the absorption had a price structure of its own — the conference system became the site of permanent internal warfare, the recall precedent made every presidency provisional, and the slow leak of splinters (COPE 2008, EFF 2013, MK 2023) externalised the rupture in instalments until the majority finally broke at 40.18 per cent in May 2024 (ZA-D-04, ZA-D-05). Whether Polokwane postponed the ANC's reckoning or merely restructured its repayment schedule is the question Section 7 leaves open.

2. The Road to Polokwane (2005–2007)

2.1 The Dismissal as Calculated Risk

The decision arc opens with a decision of Mbeki's own. The 2 June 2005 conviction of Schabir Shaik — Zuma's financial adviser, convicted in the Durban High Court on two counts of corruption and one of fraud arising from a sustained pattern of payments to Zuma totalling approximately R1.2 million over the late 1990s, and from the solicitation of an annual R500,000 bribe from Thomson-CSF (later Thint) in exchange for protection from the Arms Deal investigations [TBD-VERIFY: the precise payment aggregates as found by Squires J versus the higher figures advanced at trial] — implicated the sitting Deputy President in a corrupt relationship without trying him. The judgment created an unstable situation: Zuma stood publicly implicated by a court's findings yet legally unconvicted, and the NPA under Vusi Pikoli announced on 20 June 2005 that Zuma himself would be charged.

Mbeki's response, delivered to a joint sitting of Parliament on 14 June 2005, was the most consequential personnel decision of his presidency: "I have come to the conclusion that the circumstances dictate that, in the interest of the Honourable Deputy President, the Government, our young democratic system, and our country, it would be best to release the Honourable Jacob Zuma from his responsibilities as Deputy President of the Republic." The decision was widely praised at the time — domestically and internationally — as a demonstration that no one stood above the law, and it was constitutionally unimpeachable: section 91(2) of the Constitution gives the President an untrammelled power to dismiss the Deputy President.

It was also a calculated risk whose calculation failed. The Mbeki camp's premise was that a criminally implicated Zuma was politically finished — that the ANC's traditions of discipline and deference would hold, that the legal process would grind Zuma down before December 2007, and that the dismissal would be read as principle rather than as the elimination of a rival. Every element of the premise proved wrong. Critically, Mbeki dismissed Zuma from the state deputy presidency but could not dismiss him from the ANC deputy presidency, which was the branches' gift, not his. The gap between the two offices became the channel through which the insurgency organised: Zuma retained the party platform, the party's organisational access, and the party's protective symbolism while shedding the constraints of government office. Mark Gevisser's biography (Thabo Mbeki: The Dream Deferred, 2007) treats the dismissal as the moment the succession question escaped Mbeki's control; William Gumede's earlier Thabo Mbeki and the Battle for the Soul of the ANC (2005, revised 2007) had already documented the coalition of resentment waiting for an organising occasion.

The first measurable signal came within a fortnight. The ANC National General Council of 29 June–3 July 2005 — a mid-term policy gathering, not an elective conference — revolted against the leadership's attempt to have Zuma's voluntary withdrawal from party functions noted and confirmed, and instead restored his full functions as ANC Deputy President from the floor. Susan Booysen (The African National Congress and the Regeneration of Political Power, 2011) identifies the 2005 NGC as the first occasion on which the post-1994 ANC's branch delegates openly overrode the National Working Committee on a leadership question. The Polokwane majority was visible, for those prepared to see it, thirty months before Polokwane.

2.2 The Assembly of the Negative Coalition

The coalition that delivered Polokwane assembled around Zuma but was not primarily about Zuma. Its components shared a single property: each had been marginalised, defeated, or insulted by the Mbeki presidency, and each calculated that a Zuma ANC would restore what the Mbeki ANC had withheld.

COSATU and the SACP. The Tripartite Alliance's left flank had been in structured opposition to Mbeki's economic management since the 1996 adoption of GEAR (ZA-E-01) — the macroeconomic stabilisation programme adopted without alliance consultation, which the SACP from the late 1990s designated the "1996 class project": the capture of the National Democratic Revolution by a technocratic-capitalist class agenda. The substance of the grievance (fiscal consolidation, privatisation signalling, inflation targeting, jobless growth) was compounded by the manner: Mbeki treated the alliance partners as constituencies to be managed rather than partners to be consulted, and his public polemics — including the ANC Today columns deriding the "ultra-left" — personalised the breach. Zwelinzima Vavi (COSATU General Secretary) and Blade Nzimande (SACP General Secretary) committed their organisations' mobilising machinery to Zuma from mid-2005, supplying the insurgency with what it otherwise lacked: disciplined national organisation, a programmatic vocabulary, and the workerist legitimacy that Zuma's own record did not provide. That Zuma had no documented left-economic convictions was noted at the time and proved decisive later; the left lent its machinery against Mbeki and discovered after 2009 that it had purchased very little (the point at which ZA-J-01's structural account picks up the thread).

The ANC Youth League. The League under Fikile Mbalula declared for Zuma early and loudly, reprising its kingmaker tradition (it had played the equivalent role for Mbeki against the 1990s left). The League supplied the campaign's theatre — the T-shirts, the songs, the courtroom crowds — and its talent for transgressive escalation kept the Zuma cause permanently in the news cycle. Julius Malema, who succeeded Mbalula as League president in April 2008, belongs chronologically to the post-Polokwane, pre-recall phase — his "we are prepared to kill for Zuma" statement of June 2008 marked the escalation of recall-period rhetoric [TBD-VERIFY: the verbatim formulation and date; widely reported from a June 2008 Youth Day rally] — but the Malema energy was the continuation of a League posture set under Mbalula from 2005.

The provincial and security-adjacent periphery. A third stratum comprised politicians sidelined by Mbeki's deployment machinery: provincial figures passed over for premierships (which Mbeki had centralised into presidential gifts), former intelligence and security figures alienated by the Mbeki-era purges, and the KwaZulu-Natal ANC, for which Zuma's rise carried an explicit provincial-turn logic after a decade of Xhosa-speaking presidents — an ethnic reading that the campaign exploited at the margins and officially disavowed.

2.3 The Trials as Mobilisation: The Zunami

The conventional expectation in 2005 was that Zuma's legal calendar would destroy him. The inversion of that expectation — the conversion of each legal episode into a mobilisation event — is the signature political phenomenon of the period, and the reason contemporary reporting reached for the "Zunami" coinage.

The rape trial ran from February to May 2006 in the Johannesburg High Court before van der Merwe J: Zuma was charged with raping a 31-year-old family friend and HIV-positive AIDS activist known publicly as Khwezi (Fezekile Ntsukela Kuzwayo, named publicly after her death in 2016). He was acquitted on 8 May 2006, the court finding the sex consensual on the evidence before it. The trial's political content lay outside the verdict: the daily crowds outside the court burning photographs of the complainant and singing Umshini wami ("Bring me my machine gun," the exile-era song that became the campaign's anthem); Zuma's testimony that he had showered after unprotected sex with an HIV-positive woman to reduce infection risk — from the former chair of the South African National AIDS Council — which made him an international caricature even as it did him no measurable harm with the mobilised base; and the construction, around the trial, of the conspiracy narrative: that the charge itself was a fabrication of the Mbeki state. The trial's gender politics produced a durable scar (the persecution of the complainant, driven into exile, is a standing reference point of South African feminist scholarship) and demonstrated the coalition's central discovery: that the base's loyalty was not contingent on respectability.

The corruption track ran in parallel. Pikoli's NPA charged Zuma in 2005; on 20 September 2006 Msimang J struck the case off the roll in the Pietermaritzburg High Court, refusing the NPA's request for a postponement and castigating its state of unreadiness. The strike-off was procedural — the NPA could and did recharge — but the Zuma camp received it as vindication, and the crowds received it as proof that the conspiracy was unravelling. Each legal event thus ratcheted the narrative: prosecution as persecution, the courts as terrain of struggle, and Zuma as the embodiment of the movement against its usurpers. By the time the NPA recharged Zuma on 28 December 2007 — ten days after Polokwane — the recharge itself could be presented as the conspiracy's confirmation, with consequences for September 2008 treated in Section 4.

2.4 The Third-Term Gambit and the Misreading of the Branches

Mbeki's decision to contest the ANC presidency a third time converted the contest from a succession question into a referendum. The Constitution barred a third state-presidential term; the ANC constitution imposed no term limit on the party presidency. Mbeki's stated logic was continuity and the protection of the National Democratic Revolution from a criminally compromised candidacy; the logic attributed to him by allies and critics alike was succession management — holding the party presidency through 2009 would let him determine the next state president and protect his legacy (and his appointees) from a Zuma administration. The "two centres of power" debate that ran through 2007 was, formally, about whether party and state presidencies should be split; substantively, everyone understood it as the Mbeki-stays question.

The gambit rested on a misreading with two layers. The first was informational: the Presidency's political intelligence ran through the very deployment machinery — premiers, provincial executives, the headquarters apparatus — whose occupants owed their positions to Mbeki and reported what he preferred to hear, while the branches' temper was legible to anyone counting nomination returns. The June 2007 Policy Conference's hostility, the November 2007 provincial nomination conferences (Zuma majorities in the provinces' branch returns [TBD-VERIFY: precise aggregate nomination tallies; the reconstructed range in ZA-B-02 is 55–65 per cent]), and the contrast between Mbeki's empty halls and Zuma's stadium overflows in the same October 2007 week in KwaZulu-Natal all pointed one way. The second layer was conceptual: Mbeki's camp campaigned on the assets of incumbency — the governmental record, the macroeconomic numbers, the continental statesmanship — to an electorate of branch delegates for whom those assets were either abstractions or grievances. The branches were not the country, and the conference was not an election; Mbeki ran for the wrong office before the wrong electorate, and the December result was arithmetically visible weeks before the delegates flew to Limpopo.

3. The Conference (December 2007)

3.1 The Scenes

The 52nd National Conference convened at the University of Limpopo's Turfloop campus on 16 December 2007 with approximately 4,000 voting delegates, some 90 per cent of them mandated by branches — the constituency the Zuma coalition had spent thirty months organising. The procedural record is carried at ZA-B-02 §3; what belongs here is the conference's decision-texture, because Polokwane's scenes were themselves part of the decision's meaning.

The hall's hostility to the incumbent leadership was open, sustained, and unprecedented in the ANC's conference tradition. Mbeki's allies were booed from the floor when they rose to speak or chair; delegates deployed the rolled-hands "substitution" gesture from football crowds at the platform; Umshini wami was sung over proceedings the leadership attempted to run. Mbeki's own organisational report on the opening day — a long defence of the presidency's record, received in near-silence by a hall, as the Mail & Guardian's correspondent wrote, "that was already counting votes" — was the imperial presidency's last address to its party. The discourtesy was the message: the deference culture on which the ANC's consensus-succession tradition had rested since exile was being publicly retired, and the delegates intended the platform to watch them retire it.

The voting confirmed what the scenes announced. Zuma defeated Mbeki for the ANC presidency by 2,329 votes to 1,505 — 60.7 per cent to 39.3 per cent of valid ballots — with the result announced on the evening of 18 December 2007. The slate logic ran the table: all six National Office Bearer positions went to the Zuma slate — Kgalema Motlanthe as Deputy President, Gwede Mantashe as Secretary General, Baleka Mbete as National Chairperson — and the eighty-member NEC elected alongside them was comprehensively Zuma-aligned, with Mbeki's senior ministers tumbling down or off the list. The "slate" itself — a pre-packaged full ticket, voted as a bloc, organised through provincial brokers — was Polokwane's procedural innovation as much as its instrument, and it has organised every ANC elective conference since: Nasrec 2017 and the December 2022 conference were both fought slate-against-slate under rules of engagement Polokwane wrote.

3.2 What the Branches Were Voting Against

The negative content of the Polokwane vote is well documented because the coalition's components each published their indictment. Four counts recur across the COSATU/SACP documentation, the conference-floor culture, and the contemporaneous reportage.

The imperial presidency. Mbeki had built the most centralised executive in democratic South Africa's history: a Policy Co-ordination and Advisory Services unit that ran government from the Union Buildings; premiers and mayors effectively appointed from the centre; the party headquarters subordinated to the Presidency; and a personal style — remote, textual, allergic to consultation — that personalised the structure. The centralisation had administrative logic (treated at ZA-B-01) but distributed humiliation widely, and Polokwane was where the humiliated were counted.

The AIDS catastrophe. Mbeki's questioning of HIV science and his government's delayed antiretroviral rollout (the full record at ZA-G-01) had by 2007 a quantified mortality cost — the Harvard estimate of roughly 330,000–365,000 preventable deaths over 2000–2005 [TBD-VERIFY: Chigwedere et al. 2008 figures] — and a political cost less often counted: the denialism years had taught ANC branches, in the most intimate register possible, that the centre's intellectual self-confidence could be lethally wrong, and had supplied the Treatment Action Campaign with victories against the ANC government that normalised internal dissent.

Zimbabwe. The "quiet diplomacy" toward the Mugabe government — defensible as mediation strategy, indefensible to a labour movement watching Zimbabwean workers' organisations crushed and millions of displaced Zimbabweans arriving in South African townships — was COSATU's signature foreign-policy grievance, sharpened by episodes such as the 2005 COSATU delegation's deportation from Harare while Pretoria stayed silent.

The machinery itself. Beneath the policy counts ran the redeployment economy's ledger of resentments: every cadre redeployed downward, every provincial chair dissolved from the centre, every premier imposed over a province's objection had a constituency at Polokwane. Booysen's analysis is the standard treatment: the Mbeki machine's discipline was extracted, not earned, and extraction's price was payable on the first occasion the branches could vote by secret ballot against the extractor.

3.3 What the Branches Were Voting For — Three Readings Held at Strength

The positive content of the vote is genuinely contested, and the corpus declines to adjudicate it; the three readings below are each carried by serious literature and each explains part of the post-2007 record.

The leftward-correction reading. On the COSATU/SACP account, Polokwane was a programmatic decision: the defeat of the 1996 class project and the recapture of the ANC for the working class. The conference resolutions give the reading some support — the developmental-state language sharpened, and the resolutions on the National Health Insurance concept, industrial policy, and (fatefully) the disbanding of the Scorpions were left-coalition deliverables. The reading's difficulty is the sequel: the Zuma government's macroeconomic management remained substantially orthodox (Treasury's continuity through 2015 is the standard evidence), and the left's principals were progressively marginalised — suggesting the programme had been the coalition's vocabulary rather than its purpose.

The patronage-redistribution reading. On the structural account (developed fully in ZA-J-01 §5), Polokwane was a transfer of the deployment economy from one network to another: the branches that voted were themselves embedded in patronage chains, conference delegates were already by 2007 a partially purchasable currency, and the coalition's provincial brokers were buying future access, not future policy. The reading's strength is predictive — it anticipates the post-2009 record better than the others — and its weakness is reductive: it cannot explain the genuine moral energy of the anti-Mbeki revolt, the AIDS grief, or the documented sincerity of the left's self-deception.

The personalist reading. On the political-culture account, the branches voted for a recognisable human being against an abstraction: Zuma the singing, dancing, funeral-attending, vernacular politician against Mbeki the pipe-smoking quoter of Yeats. The reading is condescended to in the academic literature but should not be: the delegates' own explanations, recorded in the contemporaneous reportage, were saturated with the language of accessibility and restoration — the ANC being returned to its members — and the Zuma presidency's later resilience among the base is unintelligible without it.

The honest synthesis is that the vote was overdetermined against Mbeki and underdetermined for anything; Polokwane was a decision about power that left the question of purpose to be settled afterwards, by whoever moved fastest. Section 5 returns to what that settlement looked like.

4. The Recall (September 2008)

4.1 The Nine-Month Cohabitation and the Nicholson Trigger

Between December 2007 and September 2008 South Africa ran an experiment without precedent in its democratic history: a state president whose party had repudiated him, governing alongside a party president awaiting trial. The cohabitation was structurally unstable — the Polokwane NEC controlled the party machinery and the parliamentary caucus's career incentives, while Mbeki controlled the state — and the Zuma camp's operating question from January 2008 was not whether to end it but when and on what pretext. The 28 December 2007 recharging of Zuma by the NPA (eighteen counts spanning racketeering, corruption, money laundering, and fraud [TBD-VERIFY: the count structure of the December 2007 indictment]) had meanwhile hardened the coalition's conviction that the prosecution was Mbeki's instrument, and made the neutralisation of that prosecution the cohabitation's real agenda.

The trigger arrived on 12 September 2008, when Nicholson J delivered judgment in Zuma v National Director of Public Prosecutions in the Pietermaritzburg High Court. The judgment's holding was narrow and procedural: the NPA's December 2007 decision to recharge had been a "review" of an earlier decision within the meaning of section 179(5)(d) of the Constitution, requiring representations from Zuma that had not been invited; the recharge was accordingly invalid. The holding was curable — the NPA needed only to invite representations and recharge — and on its own would have moved nothing. The judgment's political payload was its obiter: Nicholson inferred from the timing of successive NPA decisions that the prosecution had been "politically directed," characterised the executive's conduct as "political meddling," and named Mbeki and former NDPP Bulelani Ngcuka adversely — although neither was a party, no evidence had been led on the interference question, and the relief sought was procedural. The legal academy's contemporaneous criticism was severe and proved prophetic: on 12 January 2009 the Supreme Court of Appeal (Harms DP), in NDPP v Zuma, reversed Nicholson comprehensively and rebuked the political findings as findings made against non-parties, without evidence, beyond the relief sought. ZA-B-03 §§3 and 9 carry both judgments in full.

The sequence matters to the decision analysis because of what it reveals about the recall's structure. The NEC acted on Nicholson within a week; the SCA destroyed Nicholson within four months; the recall was not revisited for a moment. The judgment was the occasion the coalition had been waiting for — a judicial imprimatur, however legally fragile, for the conspiracy narrative constructed since 2005 — and its function was permissive, not causal. Frank Chikane (Eight Days in September, 2012), Director-General in the Presidency through the crisis and the recall's closest inside chronicler, records that the Mbeki side understood immediately that the judgment's appealability was politically irrelevant: the NEC majority wanted the outcome, and the reasoning was fungible.

4.2 The NEC Decision

The NEC convened at Esselen Park, Kempton Park, on Friday 19 September 2008, with one agenda item — "the political situation in the country" — and resolved, in a debate running into the early hours of Saturday 20 September, to recall Mbeki from the state presidency [TBD-VERIFY: the vote was not officially enumerated; press reconstruction indicated a substantial majority of the roughly 80-member NEC, with reported figures around 60 in favour and a minority opposed or abstaining — ZA-B-03 carries the same caveat]. Secretary General Gwede Mantashe announced the decision on the afternoon of 20 September. The stated grounds tracked Nicholson: the judgment's interference findings had made Mbeki's continuation untenable. The unstated grounds were the cohabitation's: the Polokwane majority saw no reason to let a repudiated president control the state — including the security and prosecutorial apparatus bearing on Zuma's trial — for another eight months until the April 2009 election.

"Recall" itself was a terminological invention worth pausing on. The word belongs to the ANC's internal deployment vocabulary — cadres deployed by the party may be recalled by it — and its application to the head of state was a constitutional category error committed deliberately: the Constitution provides exactly two mechanisms for removing a president mid-term (impeachment under section 89; a motion of no confidence under section 102), and the NEC invoked neither. The recall was an instruction from a private association to its member. It worked because the member complied and because the party's parliamentary caucus stood ready to convert non-compliance into a section 102 motion within days — a fact both sides understood and neither needed to articulate.

4.3 The Constitutionalist Exit

Mbeki's compliance was delivered in a televised address on the evening of Sunday 21 September 2008, and the address's manner is part of the decision's historical content. He announced his resignation in the language of movement discipline ("I have been a loyal member of the African National Congress for fifty-two years... I remain a member of the ANC and therefore respect its decisions"); he defended his government's record at length; he denied, on oath-adjacent terms, ever interfering in the prosecution of Zuma ("I would like to state this categorically..."); and he offered not one syllable of constitutional resistance — no challenge to the NEC's authority, no appeal to his parliamentary or popular mandate, no caretaker manoeuvre. The resignation took effect upon Parliament's processes, and on 25 September 2008 the National Assembly elected Kgalema Motlanthe — Polokwane's Deputy President, but a figure the Mbeki side regarded as the Zuma slate's most institutionally serious member — as President of the Republic.

The exit's constitutionalism cut two ways, and both readings have a literature. On the first reading, Mbeki's submission was his finest hour: confronted with a demand of dubious legitimacy, he subordinated personal vindication to institutional continuity, prevented a confrontation between party and Constitution whose outcome no one could underwrite, and handed over power peacefully — twice over, since the handover validated the 1994 settlement's deepest promise that South African power transfers by procedure, not force. On the second reading, the submission constituted the precedent: by obeying, Mbeki ratified the NEC's claim to dispose of the state presidency, taught every successor that the party's instruction is final, and converted a constitutional anomaly into a working doctrine. Chikane's account inclines to the first reading; the constitutional-law commentary of the period inclines to the second; the corpus records that both are true at once — the exit was simultaneously a vindication of constitutionalism's forms and a subordination of its substance.

The Motlanthe interregnum (25 September 2008 – 9 May 2009; full treatment at ZA-B-03 §§6–7 and ZA-H-PRES-03) was conducted with deliberate sobriety: the cabinet substantially retained (after the market-shaking Trevor Manuel resignation episode of 23 September was walked back within hours), the HIV/AIDS posture reversed with the appointment of Barbara Hogan to Health, and the state visibly run as a caretaking trust. The interregnum's competence was itself an argument in the period's politics — evidence, deployed by all sides for different purposes, of what the ANC's institutional bench could deliver when the deployment logic was suspended.

4.4 The COPE Split and the Question the Recall Settled

The recall's immediate organisational cost was the first major ANC split of the democratic era. Mosiuoa Lekota's "divorce papers" announcement of 8 October 2008 led through the 1 November Sandton convention to the founding of the Congress of the People in Bloemfontein on 16 December 2008 — the date and city chosen for their 1912 resonance. COPE's April 2009 result (7.42 per cent, 30 seats) was simultaneously the strongest first outing of any ANC splinter and a decisive failure against its strategists' 15–20 per cent projections; its subsequent implosion through the Lekota–Shilowa leadership litigation reduced it to statistical residue by 2014 (the full arc at ZA-B-02 §6). For the decision analysis, COPE's failure mattered prospectively: it validated the Polokwane coalition's bet that the ANC brand, not its incumbent personnel, commanded the loyalty of the base, and thereby lowered the perceived cost of every subsequent act of internal ruthlessness.

What the recall settled was the question Polokwane had posed: where does the presidency live? The answer — the state president is the ANC's deployee, holding office on the NEC's sufferance, with the National Assembly's removal powers held in reserve as the party's enforcement mechanism — was never legislated, never tested in court, and never thereafter seriously contested inside the ANC. The deployee-president doctrine is the recall's principal institutional product, and its career since 2008 is the subject of Section 6.

5. The Decision Accounts

The corpus's contested-record convention applies: Polokwane and the recall support three principal accounts, each carried by serious literature, each explaining part of the record, none declared the winner.

5.1 The Democratic-Renewal Account

On the first account, Polokwane was the most democratic event in the ANC's organisational history: the branches reclaiming their party from a remote, centralising, unaccountable leadership. The evidence is genuine. The Mbeki machine had hollowed the party's internal democracy — premiers imposed, provincial executives dissolved, dissent pathologised as conspiracy — and the 2005 NGC and 2007 conference were the only available instruments of correction. The delegates used them lawfully, by secret ballot, under the party's own constitution; the incumbent accepted the result; and the subsequent recall, on this account, merely enforced the verdict — a party that had voted out its leader was entitled not to have him run the state against it for eighteen months. The account's proponents (much of the COSATU/SACP literature of 2008–2010; parts of Booysen's analysis) note that the alternative on offer was a third-term party presidency expressly designed to perpetuate the centralisation the branches opposed. The account's difficulty is the sequel: a democratic renewal that installed the Zuma presidency, disbanded the Scorpions within months, and led by documented stages to the capture era (ZA-C-01, ZA-C-02) must either disown its consequences or explain them, and the renewal literature has spent fifteen years doing both.

5.2 The Patronage-Coalition Account

On the second account — the structural reading developed at ZA-J-01 §5 — Polokwane was the assembly of a capture coalition: a network whose principals needed the state (for legal protection, for rents, for the deployment economy's redistribution) and who used the party's democratic forms as the acquisition mechanism. The evidence is the sequel read backwards: the Scorpions' disbandment (a Polokwane conference resolution, executed by legislation in 2008–09); the 6 April 2009 withdrawal of the Zuma charges on the eve of the election; the post-2009 colonisation of the SOEs, the intelligence services, and the criminal-justice senior ranks documented across the Zondo record; and the conference system's own subsequent career, in which branch credentials, delegate transport, and provincial brokerages became line items in a documented marketplace. On this account the branches' grievances were real but instrumentalised — the moral energy of the anti-Mbeki revolt was the coalition's fuel, not its purpose — and the conference system's capture vulnerability (a few thousand purchasable delegates standing proxy for a million members standing proxy for seventeen million voters) is the account's enduring analytical contribution. Its difficulty is teleology: it reads 2007 through 2016, attributes to the coalition's median member an intentionality the evidence locates only in its core, and cannot explain why serious institutionalists — Motlanthe, Mantashe, much of the SACP — rode the wave.

5.3 The Institutional-Failure Account

On the third account, the decisive fact is not who won but that the ANC could manage succession only by rupture. The party's succession technology — consensus among elders, refined in exile and adequate to the managed handovers of 1991 and 1997 — failed its first contested test: Mbeki would not retire, the machinery would not referee, and the only mechanism left standing was open warfare for branch mandates, fought with money, martyrdom narratives, and slates. On this account Polokwane and the recall are symptoms of a missing institution — the ANC has no honourable exit, no neutral umpire, and no way to separate the party presidency from the state presidency's shadow — and the same absence has reproduced the same rupture at every subsequent conference: Nasrec 2017 was Polokwane with the slates reversed, the 2022 conference was a Polokwane that fizzled, and the recall mechanism is the rupture's enforcement arm rather than an accountability instrument. The account's strength is comparative (Section 7); its difficulty is that it under-weights agency — the institutional vacuum was real, but specific men chose the third-term gambit, the conspiracy narrative, and the recall, and the vacuum did not choose for them.

No winner is declared. The renewal account best explains 2005–2007 (the revolt's moral texture), the patronage account best explains 2009–2016 (what the victors did), and the institutional account best explains 2017–2026 (why the pattern repeats with the personnel changed). A reader needs all three; the corpus's obligation is to keep all three at strength.

6. The Long Consequences (2008–2026)

6.1 The Recall Precedent's Career

The deployee-president doctrine established in September 2008 has been exercised once and brandished continuously.

The Zuma recall, February 2018 — the symmetry. The precedent's first re-use was against its principal beneficiary. After Cyril Ramaphosa's narrow Nasrec victory in December 2017, the new NEC followed the 2008 playbook with near-ceremonial fidelity: a marathon NEC meeting (12–13 February 2018), a recall resolution, a Mantashe-era communications protocol now executed by his successors, an initial refusal by the incumbent, and capitulation under the explicit threat of a parliamentary no-confidence motion — Zuma resigning in a televised address on the evening of 14 February 2018. The symmetry was total and was noted by every participant: Zuma was removed by the instrument his own coalition had forged, on the authority of the doctrine his own elevation had established, by a party majority assembled exactly as his had been — through the branches, conference by conference. The principal asymmetry was the resistance: where Mbeki complied within a day, Zuma extracted negotiations, tested the NEC's resolve for a week, and his post-2018 career (the Zondo defiance, the July 2021 unrest that followed his imprisonment, the 2023 founding of the MK Party that took 14.58 per cent of the 2024 vote) constitutes a long refusal of the doctrine's finality — the recalled president who, unlike Mbeki, never accepted the recall's legitimacy. ZA-C-01 and ZA-J-01 §6 carry the record.

The Phala-Phala survival, December 2022 — the precedent's limit. The doctrine's boundary was mapped when it met an incumbent who controlled the NEC. The Section 89 independent panel chaired by former Chief Justice Sandile Ngcobo reported on 30 November 2022 that the Phala-Phala matter disclosed prima facie grounds on which Ramaphosa might face impeachment; for some days his resignation was widely expected and reportedly drafted [TBD-VERIFY: the contemporaneous reporting that a resignation statement was prepared on 1 December 2022]; the ANC NEC then resolved to oppose the panel report's adoption, and on 13 December 2022 the party's parliamentary caucus voted it down [TBD-VERIFY: the Assembly division — reported at approximately 214 against adoption to 148 in favour]. The episode completed the doctrine's specification: the recall mechanism is not an accountability instrument that binds presidents, but a power instrument that binds weak presidents. Mbeki fell because the NEC was his enemy's; Zuma fell because the NEC had changed hands at Nasrec; Ramaphosa survived a constitutionally weightier finding than Nicholson's because the NEC was his. The party disciplines the president exactly when, and only when, the president has lost the party — which is to say the doctrine reduces, on inspection, to the Polokwane rule itself: hold the party or lose the state.

6.2 The Party-Over-State Trajectory to 2024

The recall settled party supremacy as practice; the seventeen years that followed priced it. The doctrine's operating costs ran through every layer documented elsewhere in the corpus: presidents governing for NEC arithmetic rather than national constituency; cabinet appointments as coalition maintenance; the criminal-justice and SOE senior ranks as conference currency; and the permanent campaign between conferences crowding out government (the Eskom record at ZA-D-03 is the standard exhibit of decisions deferred across electoral and conference cycles). The electoral ledger is unambiguous in direction: 65.90 per cent in 2009, 62.15 in 2014, 57.50 in 2019, 40.18 on 29 May 2024 — the loss of majority that produced the Government of National Unity (ZA-D-04), whose coalition mechanics (ZA-D-05, ZA-E-03) have forced the ANC, for the first time, to negotiate the deployee doctrine against partners who do not recognise it. The GNU-era question — whether a president heading a multiparty cabinet can still be "recalled" by one party's NEC without collapsing the coalition — is open as of 2026 and flagged forward to ZA-D-08; the doctrine has never operated outside the condition of ANC parliamentary majority that gave it force in 2008.

The causal weight Polokwane carries for the 2024 result is itself contested. The decline trajectory's proximate drivers — capture-era governance failure, load-shedding, unemployment, the MK split — each have their own genealogy, and the counter-trajectory (an ANC that managed its 2007 succession by consensus) is unobservable. What can be said within the corpus's discipline: every one of the proximate drivers passes through decisions made by the Polokwane coalition or by the conference system it created, and the splinter parties that broke the majority — COPE, the EFF (founded 2013 from the expelled Malema's Youth League network), MK — are each fragments of a Polokwane-era formation. The 2024 result is not Polokwane's verdict, but it is Polokwane's arithmetic.

6.3 The Mbeki Rehabilitation Arc

Mbeki's posthumous-in-office reputation has run a full cycle. The recall-era characterisation — aloof, denialist, paranoid, author of his own fall — dominated the 2008–2012 literature. The rehabilitation began as the Zuma record accumulated: by the mid-2010s the Mbeki-era growth and fiscal numbers, the functioning SOEs, and the continental statesmanship read as a lost golden age in precisely the commentariat that had cheered his fall, and "we did not know how good we had it" became a recognised genre of South African political writing (the genre's own excesses are flagged at ZA-N-01 — the rehabilitation airbrushes the AIDS record with the same energy the demonisation airbrushed the institutional record). Mbeki himself managed the arc with discipline: the Thabo Mbeki Foundation's elder-statesman production line, the African Union mediation portfolios, the periodic open letters dissecting ANC degeneration, the conspicuous 2024 return to the ANC campaign trail at 81, and the post-2024 positioning as the GNU era's constitutional conscience. The rehabilitation's significance for this document is doctrinal: Mbeki's standing as the recall's dignified loser — the man who obeyed — is itself part of what stabilised the precedent, and his rehabilitation has proceeded within the party whose decision he never publicly disputed. The AIDS record (ZA-G-01) remains the rehabilitation's hard limit; no serious account writes it down.

6.4 The Counterfactual Debates

Two counterfactuals organise the retrospective literature, and the corpus records them as debates rather than findings.

The Mbeki-retirement path. Had Mbeki declined a third party term in 2007 and sponsored a successor — Motlanthe being the consensus candidate of this counterfactual, with Cyril Ramaphosa the more speculative variant (Ramaphosa, then in business and untainted by the 2005–07 wars, was repeatedly floated as a compromise and repeatedly declined to organise; his actual 2007 path was election to the NEC's top ranks without a slate) — the Zuma coalition would have faced a candidate who shared none of Mbeki's negatives. The debate divides on whether the coalition's momentum was stoppable by 2007 at all (the Zunami reading: no — the branches had been captured and the conspiracy narrative was self-sustaining) or whether the coalition was held together only by Mbeki's continuation (the negative-coalition reading: yes — subtract the common enemy and COSATU, the SACP, and the provincial brokers had no shared candidate). The evidence mildly favours the second reading — the coalition fractured within four years of victory along exactly the predicted seams — but 2007 nominations arithmetic favours the first.

The early-Ramaphosa path. The deeper counterfactual asks what a 2007 or 2009 Ramaphosa presidency — the 2017 outcome arriving a decade early, without the capture interregnum — would have yielded. Its proponents read the post-2018 record as evidence of arrival-too-late: by Nasrec the institutions Ramaphosa inherited were the capture era's shells. Its sceptics read the same record as evidence that the Ramaphosa method (consensus, caution, NEC arithmetic) would have been no more decisive in 2009 than in 2019, and that the method is itself a product of the Polokwane system — every post-2007 ANC president governs first against the next conference. The counterfactual is unresolvable but analytically productive: it isolates how much of the 2009–2026 record is attributable to personnel (the patronage account's emphasis) versus the succession system (the institutional account's), which is the live question of Section 5 restated in the optative mood.

6.5 Polokwane in the State-Capture Genealogy

The corpus's formal position, harmonised with ZA-J-01: the enabling reading holds that Polokwane is the capture era's charter — it installed a principal whose freedom depended on neutralising the criminal-justice system, resolved at conference to disband the Scorpions, and assembled the brokerage networks through which the Gupta operation later moved; on this reading 6 April 2009 (the charge withdrawal) and the Zondo record are Polokwane's execution. The overdetermination reading holds that the vulnerabilities — cadre deployment (formalised under Mbeki), SOE board patronage, conference purchasability, prosecutorial exposure to the executive (the Nicholson and SCA records both document pre-2007 executive entanglement with the NPA) — constituted a standing capture surface that any sufficiently motivated winner could have exploited, and that several Mbeki-era episodes (the Arms Deal's protection, treated from ZA-A-01 forward) show the surface in prior use. The genealogical synthesis the corpus carries: Polokwane did not build the weapon, but it demonstrated, before a national audience and to every subsequent political entrepreneur, that the weapon could be picked up through the branches — and the demonstration may be the decision's most durable consequence.

7. Comparative Dimension

7.1 Dominant-Party Succession Ruptures

Polokwane belongs to a small comparative class — succession crises in long-dominant parties — within which its distinguishing feature is absorption.

The Kenyan contrast (KE-K-01). KANU's 2002 succession crisis is Polokwane's inverse. Daniel arap Moi, term-limited like Mbeki, attempted succession management from above — imposing Uhuru Kenyatta on a party whose barons had not been consulted — and the party split outward: the Rainbow Rebellion's principals left, merged with the opposition, and defeated KANU within months, ending its four-decade rule. The ANC's branches, facing an analogous attempt at succession management from above, defeated it inside the party: the insurgency captured the machinery rather than abandoning it, the losers (COPE) rather than the winners departed, and the ANC retained power for another seventeen years. The comparative variables the literature emphasises: the ANC's elective conference gave the insurgency a lawful internal route to total victory that KANU's delegate system (controlled from State House) did not, and the ANC brand's liberation equity was vastly less transferable than KANU's patronage equity — KANU's barons could take their networks to NARC; ANC dissidents discovered with COPE that the voters stayed with the name. The contrast cuts both ways: KANU's rupture cost it power but released Kenyan politics into genuine (if turbulent) electoral competition; the ANC's absorption preserved dominance while internalising permanent warfare.

The UMNO parallel. Malaysia's UMNO faced its insurgency in 1998 and chose the third available strategy: expulsion. Mahathir Mohamad's destruction of Anwar Ibrahim — dismissal, party expulsion, prosecution on charges his supporters read as fabricated — rhymes with the Mbeki–Zuma sequence (the deputy dismissed and prosecuted; the prosecution read as conspiracy; the martyrdom narrative) with the decisive difference that UMNO's machinery held against the insurgent: Anwar could not contest inside the party and was forced outward into Reformasi, two decades of opposition coalition-building, and eventual victory over UMNO itself (2018, then the premiership in 2022). The pairing is analytically clean: the same opening move (incumbent dismisses and prosecutes the deputy) produced internal capture where the conference system was open (ANC) and external insurgency where it was closed (UMNO) — and both dominant parties ultimately lost their majorities anyway, on a generation's delay.

The liberation-movement frame. The southern African literature (Roger Southall's Liberation Movements in Power, 2013, is the standard treatment) reads Polokwane within the regional pattern: liberation movements struggle to separate party from state, treat the state presidency as a deployment, and manage succession badly because the movement's legitimacy narrative has no retirement clause. ZANU-PF's November 2017 sequence — a central committee "recall" of Mugabe, backed by the military, days before the ANC's Nasrec conference — is the pattern's hard-power variant: the same party-over-state doctrine, enforced by armour rather than caucus arithmetic. Tanzania's CCM is the pattern's managed variant: regularised two-term rotation, succession brokered within the elite, no incumbent ever defeated at conference — the counterexample proving that dominant-party succession can be institutionalised, at the price of a presidential bench chosen for acceptability over distinction. On the regional spectrum from CCM's management to ZANU-PF's force, Polokwane sits in the contested middle: succession by genuine internal ballot, which is the democratic reading's evidence, fought by slate warfare and settled by rupture, which is the institutional-failure reading's.

7.2 The Recall as Institutional Innovation

The recall mechanism itself deserves comparative placement, because party removal of an incumbent president between elections is rarer than its South African normalisation suggests. Westminster systems remove prime ministers by party-room vote routinely (Australia's 2010–2018 sequence; the UK's 2022), but a prime minister is constitutionally the legislature's creature; the ANC innovation was applying caucus logic to a presidential office that the Constitution renders separately elected and impeachable only on defined grounds. The accountability reading holds that the recall supplies what pure presidentialism lacks — a between-elections sanction short of impeachment, exercised twice (2008, 2018) against presidents who had each lost their party's confidence, both times without a shot fired, a court defied, or an election missed; on this reading South Africa has quietly parliamentarised its presidency and is the better for it. The instability reading holds that the recall transfers the sanction from constitutional organs to a private association's internal majority — unreviewable, standardless, available against good presidents and bad on identical terms — and that its true function (demonstrated by the 2022 Phala-Phala limit case) is to enforce factional victory, not accountability; on this reading the recall is the party-state's signature instrument and its normalisation a constitutional erosion that the 2024 loss of majority has suspended rather than repealed. The corpus holds both readings; the GNU era is the natural experiment that will adjudicate them, since a recall attempted against a coalition president would for the first time require partners outside the doctrine to honour it.

8. Conclusion

The Polokwane decision and the Mbeki recall form a single hinge on which post-Mandela South African politics turns. In December 2007 the ANC's branches did something no liberation-movement-in-power's membership had done by ballot: they dismissed their own incumbent, 2,329 votes to 1,505, ending the consensus-succession tradition and installing the slate system as the party's permanent operating mode. In September 2008 the NEC those branches elected did something the Constitution does not contemplate: it instructed the head of state to leave office, and he left — establishing by a party resolution and a televised resignation the deployee-president doctrine that has governed the Union Buildings ever since.

The decision pair's ledger is written through the rest of this corpus. It opened the Zuma era and stands at the head of the state-capture genealogy (ZA-J-01), whether as charter or as occasion; it produced the COPE split, the EFF expulsion, and ultimately the MK breakaway — the instalment plan by which the rupture Polokwane internalised was externalised after all, down to the 40.18 per cent of May 2024; it gave South Africa a removal mechanism that worked twice without violence and a precedent that makes every presidency provisional upon conference arithmetic; and it staged, in Mbeki's constitutionalist exit, the period's most genuinely ambiguous act — a submission that was at once constitutionalism's vindication and the party-state's ratification. The corpus's three accounts of the decision — democratic renewal, patronage coalition, institutional failure — divide the record between them, and the document's refusal to declare a winner is not agnosticism but accuracy: the branches' revolt was real, the capture coalition was real, and the succession vacuum that forced the contest into rupture was real, simultaneously.

What remains open in 2026 is whether the doctrine survives its enabling condition. The recall presupposed an ANC parliamentary majority that no longer exists; the GNU has placed the deployee-president inside a coalition that does not recognise the deployment; and the party that learned at Polokwane to discipline its presidents must now learn whether it can discipline itself without one. The forward thread runs to ZA-D-05 and ZA-D-08; the backward thread, to every account of how dominant parties end — not at the elections they lose, but at the conferences where they teach themselves what power is for.

Sources

  1. Mark Gevisser, Thabo Mbeki: The Dream Deferred (Johannesburg: Jonathan Ball, 2007; abridged international edition A Legacy of Liberation, 2009).
  2. William Mervin Gumede, Thabo Mbeki and the Battle for the Soul of the ANC (Cape Town: Zebra Press, 2005; revised edition 2007).
  3. Frank Chikane, Eight Days in September: The Removal of Thabo Mbeki (Johannesburg: Picador Africa, 2012).
  4. Susan Booysen, The African National Congress and the Regeneration of Political Power (Johannesburg: Wits University Press, 2011).
  5. Jeremy Gordin, Zuma: A Biography (Johannesburg: Jonathan Ball, 2008).
  6. Alec Russell, Bring Me My Machine Gun: The Battle for the Soul of South Africa from Mandela to Zuma (New York: PublicAffairs, 2009).
  7. Anthony Butler, The Idea of the ANC (Johannesburg: Jacana, 2012).
  8. Adam Habib, South Africa's Suspended Revolution: Hopes and Prospects (Johannesburg: Wits University Press, 2013).
  9. Richard Calland, The Zuma Years: South Africa's Changing Face of Power (Cape Town: Zebra Press, 2013).
  10. Andrew Feinstein, After the Party: A Personal and Political Journey Inside the ANC (Johannesburg: Jonathan Ball, 2007).
  11. Roger Southall, Liberation Movements in Power: Party and State in Southern Africa (Woodbridge: James Currey; Pietermaritzburg: UKZN Press, 2013).
  12. S v Shaik and Others, Durban and Coast Local Division (Squires J), judgment 31 May – 2 June 2005; and S v Shaik and Others [2006] SCA 134 (appeal judgment, 6 November 2006) [TBD-VERIFY: precise citation and date of the SCA judgment].
  13. Zuma v National Director of Public Prosecutions and Others, KwaZulu-Natal High Court, Pietermaritzburg (Nicholson J), Case No. 8652/08, 12 September 2008.
  14. National Director of Public Prosecutions v Zuma [2009] ZASCA 1 (Harms DP), 12 January 2009.
  15. African National Congress, 52nd National Conference: Resolutions (Polokwane, 16–20 December 2007), and the conference's official results announcements as carried in the ANC's post-conference documentation.
  16. Thabo Mbeki, Address to the Nation on the Occasion of the Resignation of the President of the Republic of South Africa, 21 September 2008 (televised; transcript via the Presidency and the Thabo Mbeki Foundation archive).
  17. Pride Chigwedere, George R. Seage III, Sofia Gruskin, Tun-Hou Lee, and M. Essex, "Estimating the Lost Benefits of Antiretroviral Drug Use in South Africa," Journal of Acquired Immune Deficiency Syndromes 49, no. 4 (2008): 410–415.
  18. Mail & Guardian and Daily Maverick contemporaneous coverage: the June 2005 dismissal week, the 2006 trials, the December 2007 Polokwane conference reportage (including the Drew Forrest conference reports), and the 12–25 September 2008 recall sequence.
  19. Mosiuoa Lekota and Mbhazima Shilowa, founding documentation of the Congress of the People (Founding Declaration, Bloemfontein, 16 December 2008), with the 2009 Independent Electoral Commission national results.
  20. Sandile Ngcobo (chair), Report of the Section 89 Independent Panel: Phala Phala (Parliament of South Africa, 30 November 2022), and the National Assembly division record of 13 December 2022.
  • ZA-J-01: State Capture in the Zuma–Gupta Era — Three Accounts — the structural account whose §5 carries the patronage-coalition reading of Polokwane, and whose genealogy this document's §6.5 harmonises with.
  • ZA-K-03: The 16 August 2012 Marikana Massacre — Decision and Aftermath — the K-block companion; the Zuma-era state's most lethal decision-chain, made by the government Polokwane installed.
  • ZA-D-05: The Government of National Unity (2024–2025) — the post-majority coalition order in which the deployee-president doctrine faces its first test outside ANC parliamentary dominance.
  • ZA-E-03: The GNU First Year — Clearing House and Gauteng Coalition Stress (2024–2025) — the coalition-mechanics record against which the party-over-state doctrine's GNU-era limits are read.
  • ZA-H-PRES-01: Nelson Mandela Biography — the consensus-succession tradition (the 1997 Mafikeng handover) that Polokwane retired.
  • ZA-H-PRES-02: Thabo Mbeki Biography — the biographical companion to the dismissed-and-recalled principal, including the rehabilitation arc.
  • ZA-H-PRES-03: Kgalema Motlanthe Biography — the interregnum president and the consensus counterfactual candidate.
  • ZA-H-PRES-04: Jacob Zuma Biography — the biographical companion to the Polokwane victor, the 2018 recall, and the MK afterlife.
  • ZA-I-01: Chapter Nine Institutions and the Accountability Architecture (1994–2026) — the constitutional accountability machinery that the recall mechanism bypasses, and that the Zuma era then tested.
  • ZA-N-01: South Africa in International Perceptions — the external reception of Polokwane and the recall, and the Mbeki-nostalgia genre's place in the perception cycle.
  • ZA-O-01: South Africa Megatrends — The 2030s Questions — the long-horizon document for which the succession system's reformability is a standing variable.
  • KE-K-01: The 2002 Moi Succession Decision and the KANU Collapse — the comparative contrast case: the dominant party that externalised its succession rupture and lost power within months.
  • ZA-G-04: Land Reform and the Expropriation Question — Restitution, Redistribution, Tenure, and the Section 25 Politics
  • ZA-M-01: Nonracialism, Rainbow Nation, and the Contested Ideas of South African Nationhood
  • ZA-K-02: The 1996 GEAR Decision and the Macroeconomic Settlement
ArchiveSourcesChat