Mahatma Gandhi once said something worth sitting with properly today. "Non-violence is the greatest force at the disposal of mankind. It is mightier than the mightiest weapon of destruction devised by the ingenuity of man." The United Nations anchored this observance to his birthday in 2007, co-sponsored by 140 countries, built on a belief that still sounds almost naive against the evening news: that restraint, properly organised, outperforms force. South Africa in 2026 is not an easy place to test that belief. On 21 November 2025, government formally declared gender-based violence and femicide a national disaster, a designation reserved for events that overwhelm ordinary institutional capacity. Between April and June this year, 9,201 rapes were reported to police, roughly one hundred every single day, alongside 11,532 recorded sexual offences overall. The national murder count for 2025/26 came in at 23,096, a real decline from 25,421 the year before, and still a number that means someone's child, partner, or parent was killed on average every twenty three minutes. A national disaster declaration exists to force sustained, coordinated, accountable action. Ten months on, a September opinion piece in this country's own press, written after bodies were discovered in Ekurhuleni, asked the only honest question left: what has the declaration actually achieved, and what difference has it made to the safety of women. Nobody in government has yet offered a satisfying answer. Civil society groups have separately raised alarm that authorities suspended the detailed, disaggregated crime data that would let anyone outside government actually check. This is where Gandhi's idea turns from a commemorative quote into something closer to a test. Non-violence was never, in his own practice, a passive absence of conflict. It was a discipline, requiring institutions willing to be held accountable for whether it actually worked. South Africa has those institutions. Community Safety Forums, established under the Civilian Secretariat for Police Service Act, bring municipalities, police, and residents together locally. Community Policing Forums exist at station level. Shelters depend heavily on municipal support, functioning streetlights, safe taxi ranks, and whether a ward councillor treats a police station's responsiveness as their own problem. Which brings this piece directly to 4 November 2026. A councillor cannot rewrite the Domestic Violence Act or personally close the gap in national crime data. A councillor can, however, answer a very specific set of questions that reveal exactly how seriously they take this crisis, long before any ballot gave them a reason to perform concern. Has this candidate ever attended, supported, or pushed to establish a Community Safety Forum in this ward. Has this candidate advocated for functioning streetlighting in the areas residents describe as unsafe after dark. Has this candidate engaged, specifically and by name, with a local shelter or support service, rather than only referencing GBV in a campaign speech during Women's Month. What exact, fundable plan are they proposing, with a number attached, rather than a sentiment. A candidate with no answer to any of these has not yet earned the word non-violent as a description of their leadership, whatever language appears on their poster. Gandhi's claim was that non-violence is a force, not a feeling, measured by what it actually prevents. Before you vote, ask each candidate in your ward to meet that same standard, with evidence rather than adjectives. Character first. Then the Ballot.

The Only Poll That Counts Night before election. Newspaper has obeyed the law and printed nothing. Somewhere in a family WhatsApp group, an aunt forwards a screenshot of a poll anyway, sourced from nowhere, paid for by nobody anyone can name. Which of those two things did the regulation actually stop? That image sat at the heart of Ebrahim Fakir's presentation at the Electoral Commission Gauteng Thought Leadership Seminar, held at the Birchwood Hotel and Conference Centre on 30 September 2026. The seminar asked whether pre-election and exit polls can influence voter choices ahead of the 4 November local government elections. Fakir, a researcher and analyst in politics and governance, answered with a warning of his own: be careful what you ban. He opened with a confession. He has problems with some polls, he said, and has written about it twice, once in 2019 with Ivor Sarakinsky and again in 2024. His objection, though, is political, not legal. The remedy he prefers is public education, literacy and debate, not what he called the blunt instrument of law. Instead of asking whether we are for or against regulation, Fakir asked three quieter questions. What democratic value is at stake? What harm is actually proven? What is the least intrusive tool that would fix it? The evidence, it turns out, is messier than the headlines. A French study found that exposure to exit-poll results cut turnout by about 11 percentage points. A large Dutch experiment, by contrast, found that simple poll results on their own barely moved anyone. His conclusion was careful, almost stubbornly so. Polls do not determine votes. Under certain conditions, they can produce measurable effects. Those are very different claims. Pre-election polls and exit polls, he insisted, should not be treated as twins. A poll taken weeks out tells us something about the public mood. An exit poll released while stations are still open can tell a voter in the queue that the race is already over. South Africa's Electoral Act, under section 109, already bars publishing exit-poll results during voting hours. Fakir would keep that rule, even tighten it. His concern lies with Bill B33 of 2025, the proposal to create a Polling Ombud, register pollsters and publishers, and ban opinion polls in the final seven days before an election. He credited its strengths, especially the push for disclosure of sponsors and methods. The worries were sharper, though. A seven-day blackout, possible prior approval of publication, and criminal penalties of up to two years all sit uneasily with section 16 of the Constitution, which protects the right to impart and receive information. Other democracies have walked this road. France cut its seven-day ban to the eve and day of the vote after its highest court found the longer ban incompatible with European human rights law. Canada's Supreme Court struck down a three-day ban in 1998. A long blackout, it seems, rarely survives contact with a courtroom, or with the internet. What Fakir proposed instead was a ladder, starting at the lowest rung that works. Strict exit-poll embargoes. Voluntary disclosure of sample sizes, weighting and who paid. Journalists trained to interrogate a poll before they echo it. An IEC that keeps its energy for its core job, running free and fair elections. Set beside Prof. Bheki Mngomezulu's call for regulation earlier in the day, the debate felt genuinely alive, which is exactly what a seminar like this should offer. Fakir's argument landed close to home for us. Democracy literacy has always been our work, and his case rests on a simple belief we share, that citizens can be trusted with information if they are equipped to question it. Perhaps that is the real test. A law can silence a newspaper for a week. Only an informed voter can look at a forwarded screenshot and ask, who made this, and why?

The Number Before the Vote Picture a queue outside a school hall in Tshwane on 4 November. The morning is cold. You have your ID in one hand and, somewhere at the back of your mind, a headline you read three days ago telling you who is going to win. Does that headline walk into the voting booth with you? That was the question on the table at the Electoral Commission Gauteng Thought Leadership Seminar, held at the Birchwood Hotel and Conference Centre on 30 September 2026. The theme was blunt: can pre-election and exit opinion polls in South Africa influence voter choices? With the local government elections just over a month away, and the IEC's "Get Up, Show Up, Vote X" campaign in full swing, it felt less like an academic exercise and more like a warning light on the dashboard. Prof. Bheki R. Mngomezulu of Nelson Mandela University delivered the presentation, and he did not pretend there was an easy answer. Polls, he reminded the room, are everywhere now. Survey companies run them. Media houses run them. Political parties run them too, which is where it gets interesting. One comparison stayed with me, polls are a bit like university rankings. Who did the counting, and why, matters almost as much as the count itself. He took us around the world first. America, heading into its midterms. Israel, where surveys suggest no bloc will reach a governing majority on its own. Then Zambia, where the Zambia Election Research Network found that 41% of respondents felt the economy had worsened, yet 55% still backed the governing party. President Hakainde Hichilema went on to win with 60.49%. Sometimes, it seems, the poll gets it right. Here at home, the projections are already circulating. The Social Research Foundation expects the ANC's support in Johannesburg to fall and the DA's to grow. Some surveys suggest the DA and the FF+ could jointly govern Tshwane. In eThekwini, the MKP is projected to lead, though possibly below its 2024 showing. Do those numbers change how people vote? Honestly, nobody fully agrees. Prof. Mngomezulu was candid about that. Some voters never look at a poll, they already know where their X is going. Others, especially the undecided, may not want to "waste" a vote on a party that is supposedly losing anyway. He cited Joel Netshitenzhe's view that polls can be both cause and effect of voter sentiment, which, I think, was the most unsettling idea of the day. A poll that predicts a mood can also manufacture one. Parties feel it too. A good number can make a campaign lazy. A bad one can make it frantic. What, then, should be done? South Africa currently has no law governing pre-election polls. Section 109 of the Electoral Act offers some guidance on exit polls, and not much more. ActionSA's proposed Election Integrity Bill would change that, requiring poll sponsors and methodologies to be declared, establishing an Office of the Polling Ombud, and banning the publication of polls in the week before an election. Prof. Mngomezulu acknowledged the obvious risk, those who benefit from polling, whether the firms or their sponsors, may well head to court. His conclusion was still firm. Polls can neither be ignored nor stopped. They can, however, be regulated, monitored and verified. The Thuli Madonsela Foundation was in the room as thought leaders, sharing our views on what honest, transparent information means for a democracy, and, just as importantly, listening. Our work in democracy literacy has taught us that a vote is only as free as the information shaping it. We left with more questions than answers, which is probably how it should be. What is at stake is not a percentage point here or there. It is whether ordinary South Africans can trust that the numbers in front of them are honest, and that trust is what we must protect as we pave a better way forward. Thirty-four days remain. Read the polls if you like. Ask who paid for them. Then go and vote anyway.

A concrete tool worth knowing exists, because it is one of the few places this test moves from character judgement to an actual paper trail. Every councillor in South Africa is legally required, within sixty days of election, to declare their financial interests in writing to the municipal manager, including shares, directorships, property, other employment, and any gift above one thousand rand. This is not optional. It is written directly into the Code of Conduct for Councillors under the Municipal Structures Act. Here is the honest complication. The law leaves it to each municipal council to decide which parts of that declaration become public, and multiple independent reviews have found that many municipalities do not publish these registers proactively at all. Residents are often required to submit a formal request under the Promotion of Access to Information Act just to see what a sitting councillor has already legally declared. This is worth doing anyway. A councillor whose declared interests suddenly shift, or whose declaration is difficult to access at all, is telling you something either way. Ask your municipality directly whether the register exists and how to view it. The right to ask is already yours. Character first. Then the ballot.

Here are the specific patterns worth training yourself to notice, since self interest rarely announces itself directly. It shows up in decisions that, individually, each have a plausible excuse. Watch who gets appointed to well paid positions once a candidate takes office. Family members, business associates, or people who supported their campaign filling roles they are not obviously qualified for is a pattern, not a coincidence, once it repeats more than once. Watch whether tenders and contracts start flowing toward companies connected to the candidate or their inner circle, even loosely. Watch their visible lifestyle against their declared income, not out of suspicion for its own sake, but because a rapid change with no obvious explanation deserves one. Compare that against the opposite pattern. A service motivated leader tends to make decisions that cost them something personally or politically, protecting a process even when it slows them down, appointing people more qualified than convenient, staying reachable to residents long after the campaign photos are taken. Self interest optimises for the leader's position. Service optimises for the outcome, even when the leader personally loses something by choosing it. Character first. Then the ballot.

Why does this pillar matter more than it first appears to? Because service and self interest produce nearly identical campaign speeches, and only diverge once someone actually holds power. A leader motivated by service treats the office as a responsibility they happened to be given. A leader motivated by position, power or financial gain treats the office as a resource they finally have access to. From the outside, in a manifesto or a doorstep conversation, these two people can sound indistinguishable. The difference only becomes visible in what they actually do once elected, when nobody is forcing them to choose between the two. Watch Ekurhuleni again, from this specific angle. A mayor reshuffled his own Mayoral Committee not once but twice in a matter of months, once removing opposition portfolios entirely, once removing a member of his own party who happened to be a rival for his own position. Neither move was about competence. Both were about consolidating power in one person's hands. That is what it looks like when position, not service, is the actual motive. The test is never what a candidate says they will do. It is what they protect first, once the office is theirs to use. Character first. Then the ballot.

Can You Hold Someone Accountable for a Document They Cannot Read In 2023, the Department of Cooperative Governance and Traditional Affairs ran a skills audit across KwaZulu-Natal's councillors. The finding, first reported by News24 that July and repeated since in the Mail and Guardian, TimesLive and the Sunday Times, was specific and unsettling. Nearly three hundred councillors in that province alone could not read or write. No equivalent audit exists for the other eight provinces, so there is no verified national figure, and any claim of a national majority remains unproven. What is verified, and what deserves your full attention regardless, is this. Nearly three hundred people, in one province, were voted into offices that require them to approve multi million rand budgets, scrutinise integrated development plans, and hold municipal officials accountable, without being able to read a single page of any of it. Here is the question this campaign has asked all month, applied now to its sharpest possible edge. What are the actual minimum requirements to become a ward councillor in South Africa? The honest answer is startling in its brevity. You must be a South African citizen. You must be a registered voter. You must not fall under a narrow list of disqualifications, insolvency, certain criminal convictions, unsoundness of mind. Nowhere in that list does literacy appear. Nowhere does any minimum education requirement appear at all. Sit with what that actually means in practice. A ward councillor's job, as this campaign has described in detail this month, includes representing residents in council, voting on the municipal budget, and chairing a ward committee meant to feed real community concerns into real municipal decisions. Every one of those duties assumes the person doing them can read the documents involved. A councillor who cannot read a budget cannot meaningfully vote on one. They can only be told what it says, by whoever benefits from them voting a particular way. This is not a question of dignity or intelligence, and it would be dishonest to pretend it were. Illiteracy is not a moral failing, and it reflects, overwhelmingly, decades of unequal access to education that this country has still not fully repaired. The question here is narrower and more practical. Whatever the cause, can a person who cannot read a document actually exercise independent oversight over that document? Can they hold a Municipal Manager accountable for a budget they cannot verify themselves? Can they represent a ward's written complaints, correspondence, and formal grievances, when reading those complaints is itself part of the job? Dr Nonhlanhla Ngcobo, the North West University law lecturer whose research first drew wide attention to this issue, has been careful on exactly this point. She has said plainly that whether the true national number is three hundred or three thousand, the underlying problem is the same. A councillor who cannot read cannot independently verify anything placed in front of them, and becomes, in her words, a seat warmer, someone who votes as instructed rather than as informed. This campaign has spent three weeks arguing that character and competence both matter, and that voters have to test for both before they vote, not after. Literacy sits underneath every other test this month has offered. A candidate can be perfectly ethical and still be unable to catch a fraudulent budget line they cannot read. Ask directly, before you vote, whether your candidate can read the documents the job actually requires. It is not an insulting question. It may be the most basic one you can ask. Character first. Then the Ballot. Source: Mail & Guardian, "Majority of South African councillors cannot read or write," 14 October 2025, mg.co.za/2025-10-14-majority…, drawing on the Department of Cooperative Governance and Traditional Affairs' 2023 KwaZulu-Natal skills audit and Dr Nonhlanhla Ngcobo's research at NWU

One line from this year's manifesto season is worth lifting out and handing back to every party equally, not because one party said it best, but because it names a standard any of them could be held to. At the launch of Build One South Africa's 2026 manifesto, leader Mmusi Maimane told supporters that South Africans should judge political parties not by the promises they make during elections, but by whether they can deliver on those promises once elected. Whatever you make of that party's own politics, the standard itself belongs to nobody. It is simply what accountability sounds like when someone is willing to say it about themselves in advance, before the votes are counted, rather than only demanding it of opponents afterward. Take that exact sentence and put it to any candidate from any party at your door this year. Would you be comfortable being judged, five years from now, purely on whether this specific promise came true? A purpose driven candidate answers that question quickly, because they already expect to be measured against something real. A transactional one hesitates, because the promise was never meant to survive that long in the first place. Character first. Then the ballot

A practical checklist for the individual candidate in front of you, since a manifesto is written by a party and a doorstep conversation is not. Consistency across time. Ask what this person was saying about your community three years ago, before any campaign existed. A purpose driven candidate's concerns barely shift. A transactional one's concerns track whatever issue is currently popular. Consistency across audience. Notice whether their pitch changes depending on who is listening. A purpose driven candidate tells a business forum and a street meeting a version of the same underlying conviction. A transactional one tells each room exactly what that room wants to hear, and the two versions rarely survive being placed side by side. The presence of an unpopular position. Purpose occasionally costs a candidate support, because a genuine conviction does not reliably align with whatever the crowd wants this week. If every single position a candidate holds is also the popular one, that is not evidence of good judgement. It is evidence you have not yet found the one place their purpose and the crowd disagree. None of these questions require inside information. They only require listening across more than one conversation, and comparing what you hear. Character first. Then the ballot.

A distinction worth knowing before you read a single manifesto this year, because it explains why two documents can list the same promises and still mean something completely different. Leadership researchers have long separated transactional leadership, an exchange, your vote for my delivery, from purpose driven leadership, which asks something larger of both sides. Every major party contesting this election has, correctly, promised water, electricity, jobs and safety in its 2026 local government manifesto. That is not the test. Every serious party should promise the basics. The test is what sits underneath the list. Read any manifesto and ask one question. Does it explain why this party exists at all, in a sentence that would still make sense if you deleted every specific service promise from the page? A transactional document survives that deletion as nothing. A purpose driven one survives it as a conviction, a theory of what the community is actually for, that the service list is only one expression of. This is not about which party passes that test. It is about training yourself to notice the difference, because the same test applies just as precisely to the individual candidate standing at your door. Character first. Then the ballot.