Source: Revitalising Labour

Instead it committed to building a broad anti-fascist front and backing a unity candidate of the gauche de rupture (the “left of rupture,” the bloc distinguishing itself from social-liberalism, the term used across much of European politics to describe social democratic parties that have adopted neoliberal politics), naming Jean-Luc Mélenchon as the candidate currently best placed to do that job. This move was motivated by the risk of the far right winning the presidency in 2027.

While a small far-left party making such an announcement might appear irrelevant, or self-important, in the current context of French politics it’s important. Mélenchon is currently running close to level with Édouard Philippe for the second qualifying spot behind Jordan Bardella, within the margin of error in recent polling. Fringe left candidacies typically poll under 1% each, but in a contest this tight, even a few thousand votes freed up by NPA-A’s withdrawal could be the difference between Mélenchon reaching the run-off and not.Subscribe

For readers who’ve followed this Substack since March, a discussion of the intricacies of French left electoral politics might seem an odd tangent. But I’ve had a deep interest in French politics going back decades, and the polarisation we’re seeing in Australian politics has, in important respects, been more advanced in France. So while the two countries operate in different political systems with very different balances of forces, there are lessons Australian progressives can draw from how France has responded to the challenges now on our own horizon as we deal with One Nation’s current polling surge, particularly how to most effectively build a left bloc capable of blocking a government of the right, especially one dominated by or reliant on the far right. The mechanics of how France’s electoral system handles exactly this kind of choice are different to Australia’s, but the approaches operating in France, and how badly its cordon sanitaire against the far right has held up in practice, are worth understanding. I’ve written in more detail on what One Nation’s rise might mean for Australia, including this comparison.

In the 2024 French legislative elections, the cordon sanitaire against the Rassemblement National (RN, National Rally) held only selectively: of 306 initial three-way contests involving RN, 89 remained after the round of tactical withdrawals, because a number of Macronist candidates refused to stand aside for La France Insoumise (LFI, France Unbowed) candidates specifically, treating LFI as an acceptable casualty of the same logic meant to block RN.

Understanding the NPA-A announcement, its broader significance and its political continuity, requires understanding not just French politics today, but how it got there.

The origins

The process that produced the NPA, and the sequence of splits and near-fusions that followed it through to this year’s decision, didn’t begin with the 2009 founding congress. It began with a result the Ligue Communiste Révolutionnaire (Revolutionary Communist League, LCR) didn’t expect and didn’t quite know what to do with: 2002.

Olivier Besancenot took 4.25% in that year’s presidential first round, comfortably above what the LCR’s own organisational size would have predicted, and alongside Arlette Laguiller’s 5.72% for Lutte Ouvrière (LO, Workers’ Struggle), both individually outpolled the Parti Communiste Français (PCF, French Communist Party), which had been the dominant force to the left of the PS for decades. That was a real advance. But it landed inside the same election that eliminated Lionel Jospin from the run-off. Each candidate to the left of the PS individually fell short of the gap between Jospin and Le Pen, but their combined vote would have carried a hypothetical unitary left candidate through to the second round instead. What actually happened was a political earthquake: Jean-Marie Le Pen, leader of the Front National (National Front, FN), went through against Chirac in a result the whole country was forced to treat as an emergency.

Whether the charge that followed, that the far left had let Le Pen through, was fair is a separate question, one with an added irony: Jospin, like Mélenchon, had been part of a Trotskyist entryist current inside the PS in the 1970s, the Lambertist Organisation Communiste Internationaliste. Jospin’s own government had pursued the most extensive privatisation programme in French history to that point, after promising the opposite, and had bled a large part of its own coalition’s support well before polling day. But the accusation didn’t need to be accurate to be consequential, and the LCR couldn’t out-argue it in real time. It set the terms for a decade of internal argument about the real meaning of the LCR’s own electoral success.

2007: The argument before the NPA existed

The 2006 collectifs unitaires process aimed to produce a single left-of-PS (Parti Socialiste, Socialist Party) unity candidate. It fractured into three separate campaigns instead. The PCF engaged with the process and then broke from it unilaterally, choosing Marie-George Buffet. The LCR didn’t sign the founding appeal at all, judging its wording left open an eventual “gauche plurielle” arrangement, support for a PS-led government, which it treated as non-negotiable to exclude. The collectifs then split from the PCF once Buffet was chosen, turning to José Bové instead.

Inside the LCR, a current around Christian Picquet had opposed running Besancenot again and campaigned for Bové once that argument lost. Besancenot went on to outperform both Bové and Buffet, a result that vindicated the LCR majority’s refusal to subordinate itself to a process that hadn’t addressed its concerns about independence from the PS. It didn’t settle the underlying argument, though, and that argument carried directly into the NPA’s founding.

2009: The founding split

The dispute at the NPA’s 2009 founding congress was whether to enter discussions with the emerging Front de Gauche (FdG, Left Front) before any resolution of its relationship to the PS, or to negotiate first and decide after. Around 15% of delegates, roughly 3.5% of the membership, took the first position and walked when they lost, founding Gauche Unitaire (GU, Unitary Left). GU joined the FdG, which won seats in that year’s European elections while the NPA won none.

GU’s own subsequent history complicates any easy reading of which path served an organisation’s survival. GU split twice more, over whether to join the “Ensemble!” formation in 2013-14, with only part of its membership going in, before what remained, reduced to a few hundred members, dissolved into the PCF in 2015. Fourteen years after leaving the NPA to avoid what it saw as sectarian isolation, it no longer existed as an independent organisation at all.

2011–2012: The argument at scale

The same dispute resurfaced at the NPA’s 2011 congress, split three ways: entry into the FdG (a quarter of delegates), continuation of an independent anticapitalist pole (the largest minority, over 40%), and greater orientation to the social movements (another quarter). By the July conference, the field had narrowed to a 50.4% majority for an independent presidential campaign against a 40% minority arguing the majority was retreating into sectarianism. That minority constituted itself as Gauche Anticapitaliste (GA, Anticapitalist Left).

This is also where the argument first got tested against outside scrutiny, in a form still recognisable today. Dick Nichols’s January 2012 talk to the Australian Socialist Alliance’s national conference judged the NPA’s refusal almost entirely on the FdG’s electoral prospects. My own response at the time argued this missed the actual question: not what the FdG would do in office, but which orientation better positioned militants to reach the working class as it actually existed, a standard I’ve since set out more fully, and one I believe was at the core of the historic Trotskyist movement’s debates over the “French Turn.”

2017: The same test, a different government

The 2017 election put the same question in a sharper form. Mélenchon, having broken from the PS in 2008 to found the Parti de Gauche (PdG, Left Party) and then built LFI as an independent, personally-led vehicle, took 19.6% in the first round: the strongest result to the left of social democracy since 1969. FI itself split internally over whether to call a Macron vote in the run-off against Le Pen; the NPA declined to call one at all, judging that a government that would go on to attack the people organising against Le Pen wasn’t a bulwark against her in any sense worth the name.

The 2017 legislatives that followed showed what the alternative to unity costs: FI and the PCF, unable to agree a common banner, ran against each other in seats the PCF already held. The combined left vote collapsed to its lowest since the war; abstention hit 51.29%; Macron’s LREM (La République En Marche) secured a majority resting on the active support of barely 16% of registered voters.

2022: Engagement up to the line, and no further

NUPES (Nouvelle Union Populaire, Écologique et Sociale, New Ecological and Social People’s Union) brought FI, the PCF, the PS, the Greens and Génération.s together in 2022, on the strength of Mélenchon’s 21.95% first-round score. Génération.s was itself a PS breakaway, founded by Benoît Hamon in 2017 after his own collapsed PS presidential run convinced him the party had had its day.

NPA-A was not simply aloof, the charge LO levelled from a genuinely different independence than NPA-A has ever claimed. NPA-A engaged FI’s leadership on joining, and said it would have considered a joint slate on the proposed platform, ruling out participation only once the PS was admitted, the same line running unbroken from 2007. It called for active votes for NUPES candidates in every seat but the ten it was running in, while declining to stand against them, drawing criticism from its own internal “Spark” tendency for going too far toward NUPES even in that qualified form.

Why the PS, still

The classical fear, the PS-PCF-Green government of 1997-2002, was about a dominant PS using junior partners as ballast. By 2022, and the parliament that followed 2024’s snap election, that scenario had inverted: a NUPES or Nouveau Front Populaire (NFP, New Popular Front)-led government was never going to be PS-dominated. The risk instead is the leverage a small partner holds through the threat of defection. Over the 2025 budget, the PS negotiated separately with the finance ministry while LFI was excluded, then declined to join the rest of NFP’s censure motion against Bayrou’s government after it forced the budget through by decree, a non-censure Mélenchon called treachery.

There’s a deeper reason the barrier has held. The 1983 tournant de la rigueur (austerity turn), usually read as an external shock forcing Mitterrand’s government off the 1981 Common Programme, is more precisely the resolution of an argument already running inside the PS’s own finance apparatus, won by a technocratic current around Jacques Delors. That current didn’t disappear once it had won; Delors carried it to the European Commission and helped build the Maastricht-era constraints every subsequent French government has operated inside. When Macron built En Marche in 2016, he built it substantially from that same current’s personnel, his own background as a PS-aligned economy minister, Christophe Castaner and Richard Ferrand both drawn from the PS apparatus. Renaissance’s own lineage still lists a split from the PS as one of its two roots.

A continuing test for the centre

When Macron’s 2024 dissolution forced legislative run-offs, 217 candidates withdrew to block RN in triangular contests. The withdrawal wasn’t uniform: Édouard Philippe and Bruno Le Maire took the position that neither RN nor LFI should be treated as acceptable, and at least one Renaissance candidate, Loïc Signor, refused to stand down for a sitting LFI deputy even with RN leading, treating an LFI win as functionally equivalent to the danger the withdrawal convention exists to prevent.

That matters for 2027 because Philippe is currently level with Mélenchon for the second qualifying spot. Recent polling has Le Pen beating Philippe narrowly in a hypothetical run-off, and beating Mélenchon by a landslide. Whatever the reasons for that gap, it’s the empirical shape of the question NPA-A’s decision poses to the wider French political system. The left has reflexively backed centrist candidates against the RN and its predecessors since 2002, most starkly in Chirac’s 2002 landslide. The centre’s record of reciprocating that support for the left has been far less consistent, and 2024 is the clearest recent evidence of it: parts of the centre chose to treat LFI as no more acceptable than RN rather than close ranks against the far right. Whether that same asymmetry holds if Mélenchon reaches a presidential run-off against Le Pen is still an open question.

The problem of Mélenchon as candidate

Mélenchon’s own record is one of repeated, largely unilateral rupture: leaving the PS in 2008, then in February 2016 announcing LFI’s formation without prior coordination with his Front de Gauche partners, later stating the FdG had been dead for him since 2015. François Ruffin, a documentary filmmaker turned LFI deputy and one of the movement’s best-known figures outside Mélenchon himself, left LFI in 2023, accusing Mélenchon of running the movement like a royal court; his departure is the most prominent of several internal ruptures since.

The PCF’s own 2016 internal consultation is instructive on how this reads even to natural allies. Delegates initially voted, against their national secretary’s explicit preference, to run an independent PCF candidate rather than back Mélenchon, a rebuke rooted substantially in his having launched his campaign without consulting the party. It took a wider membership ballot to reverse that into what’s accurately described as critical, rather than enthusiastic, support.

The Aude allocation

NPA-A has its own, more recent experience of what alignment with Mélenchon’s project can cost a small partner. In the 2024 NFP scramble, NPA-A’s own most recognisable figure, Philippe Poutou, its three-time presidential candidate, a Ford factory worker and CGT (Confédération Générale du Travail, General Confederation of Labour) organiser from the Blanquefort plant’s closure fight, and a sitting municipal and metropolitan councillor in Bordeaux, was allocated a seat in the Aude, a different department entirely, disconnected from the base he’d spent two decades building. The placement served no one’s evident interest: the local PS federation objected too, running its own dissident candidate rather than accept it.

There’s no evidence this was deliberate rather than simple carelessness in a rushed, top-down carve-up, and it would be dishonest to claim otherwise. What is demonstrable is that the process failed by its own stated logic: a seat-maximising alliance should have run its most locally embedded candidate where his standing could do the coalition some good. Doing so would also have rewarded NPA-A for its own move to support and maximise the united left vote, rather than treating that support as something to be taken for granted.

The choice, made with eyes open

NPA-A is not backing Mélenchon because it trusts him, or because it has forgotten what happened to Poutou two years ago. It’s backing him because it judges that two conditions have, for the first time, aligned: a far-right threat serious enough to demand consolidation, and a pole worth consolidating behind that, unlike the FdG in 2007, 2009 and 2012, doesn’t require an orientation of directly supporting and following the PS.

NPA-A’s own declaration is careful about what it is and isn’t offering: it retains its political and organisational independence, proposes building “cadres of action” with LFI and the wider social-movement and union forces rather than dissolving into anyone else’s structure, and insists the fight is happening in the streets as much as at the ballot box, language that seeks to pre-commit against the specific failure that undid the 1936 Popular Front, where mass action was eventually treated by its own governing coalition as a threat to manage. NPA-A can’t guarantee that commitment holds any better than the PCF’s mass base could in 1936. What it can do, being the smaller formation, is try to create a pole pointing in the direction it believes meets the demands of the moment.

What this actually demonstrates

Core to the LCR majority and NPA-A’s orientation, from the 2002 election onward, is a single, consistent method rather than competing loyalties: orient toward the most radical, organisable layer of the class as it actually exists; set political boundaries by what that orientation requires at this moment; and hold the one line, no accommodation of the PS specifically, based on what experience has shown, repeatedly and in different guises: accommodation to the PS demobilises and discredits the radical left in the eyes of the class.

NPA-A’s approach isn’t premised on the outcome being certain. Whether the centre extends the front républicain to Mélenchon; whether a Mélenchon government would honour “the streets as much as the ballot box” any better than 1936’s coalition honoured its own version of that commitment; whether NPA-A’s read of where the working class currently stands is correct: none of these are things this piece, or NPA-A’s own declaration, can know in advance. What can be said is narrower: NPA-A has applied the same test it has applied since 2007, at real cost to itself, and concluded that the answer to how best to advance the interests of the class heading into 2027 is unity.