In the spring of 1936, three women met in Madrid and founded an organization that would never quite be permitted to exist by the movement it belonged to. Lucía Sánchez Saornil was a poet and CNT organizer. Mercedes Comaposada was an educator and writer trained in law. Amparo Poch y Gascón was a doctor who ran a maternity clinic and wrote on sex education at a time when both subjects were dangerous. The organization they founded was called Mujeres Libres — Free Women. Within three years, Franco's victory would scatter its surviving members into exile across France and Mexico. In its brief life, it grew to perhaps twenty thousand members, published a journal, ran literacy and vocational schools across Republican Spain, and insisted — against the wishes of the CNT-FAI, the anarcho-syndicalist confederation it grew within — on its own autonomous existence.
In the years since, scholars have variously called Mujeres Libres an early anarcha-feminist organization, a precursor to second-wave feminism, a doomed experiment of the Spanish Revolution. All these descriptions are accurate, and all of them miss the point. The point is structural. Mujeres Libres was the first time in the modern era that women inside a revolutionary movement organized autonomously within it — not as a faction, not as a women's auxiliary committee, but as a parallel architecture, insisting that the new society could not be built unless it was being built simultaneously on three fronts: against capital, against fascism, and against the patriarchal habits of their own comrades.
In July 2012, in the chaos of the first months of the Syrian Civil War, the Assad regime quietly withdrew its forces from a strip of northeastern Syria populated largely by Kurds. What happened in that vacuum, over the years that followed, became known as the Rojava Revolution. Most of the international attention has gone to the all-women's military units — the YPJ — that fought ISIS to a standstill in places like Kobanî. Less visible has been the parallel civilian architecture: Kongra Star, the women's umbrella organization; the Mala Jin, women's houses for dispute resolution; the mandatory co-presidencies in which every administrative position is held jointly by a woman and a man; the women's communes inside the communes; the development of Jineolojî, the "science of women," as a body of theory shaped by Kurdish women over decades.
Between Barcelona in 1936 and Qamishli in 2012 there is no direct line of transmission. The Kurdish women's movement did not derive its insights from Mujeres Libres. Its own genealogy runs through different soil — through the long decades of the Kurdish freedom struggle, through Sakine Cansız and her co-founders, through the development of Jineolojî in the 2000s, through Öcalan's prison writings and his absorption of Murray Bookchin (who in turn had written The Spanish Anarchists in 1977, but whose engagement with Mujeres Libres specifically was limited). The political theory traveled. The women's-organizational form, in its specific architectural detail, did not.
And yet the two movements, separated by seventy-six years and three thousand miles, arrived independently at the same conclusion. They built, in the same shape and for the same reasons, parallel structures inside their respective revolutions. The question is why.
Barcelona, 1936
When the Spanish Civil War began in July 1936, the CNT — the National Confederation of Labor, the anarcho-syndicalist trade union — was already the largest workers' organization in Spain. In Catalonia and Aragón, where the Republican government's hold was weakest in the early months, the CNT and its affiliated militias effectively ran social and economic life. Factories were collectivized. Land was held in common. The streets of Barcelona, as George Orwell observed in Homage to Catalonia, were briefly transformed: tipping had disappeared, Señor had been replaced by Camarada, class distinctions seemed to have dissolved.
What Orwell did not write about, because it was not visible to him, was the position of women inside the CNT itself. The organization formally endorsed gender equality. In practice, women were expected to support the revolution by carrying it out as men understood it: in the factory, on the front lines, at the barricade. The household remained the household. The men of the CNT who marched in the streets of Barcelona returned home to wives who had cooked the dinner that fueled the march.
Mujeres Libres named this gap and refused to bridge it without altering its structure. The organization had two stated programs. Capacitación — preparation — was the work of building women's individual capacity through literacy classes, vocational training, political education, and healthcare. By 1938 Mujeres Libres ran schools across Republican Spain. Comaposada wrote on pedagogy in language strikingly close to what Paulo Freire would develop three decades later — education as the practice of freedom, beginning from the learner's own experience. Captación — incorporation — was the work of building the organization itself, of pulling women out of domestic isolation and into movement life.
What made Mujeres Libres specifically anarchist in its analysis was the refusal of the dominant feminist solutions of its era. They did not want women's suffrage on equal terms with men's, in part because they did not want a parliamentary system at all. They did not want women's entry into the male workforce on male terms. They wanted to remake the structure of work, of family, of education itself, simultaneously and as parts of one project. Sánchez Saornil published a series of essays in Solidaridad Obrera — the CNT's daily newspaper — arguing that women's oppression was not a separable issue that could be addressed after the revolution. It was constitutive. A revolution that left it intact would not be a revolution.
The CNT-FAI did not agree. At successive plenums, the confederation declined to recognize Mujeres Libres as an autonomous federation within the broader anarchist movement, on the grounds that creating a separate women's organization would divide the working class. The argument was familiar — it was the same argument made against autonomous Black organization within the American labor movement in the same decade, and against women's caucuses within the American New Left thirty years later. Mujeres Libres did not accept it. They continued to organize on their own terms, with or without recognition, and by 1938 had grown to a size the confederation could no longer pretend not to see.
The organization's analysis of women's oppression became known as the triple esclavitud — the triple slavery. Women, Mujeres Libres argued, suffered three distinct but interlocking forms of subjugation: the slavery of ignorance (denied education), the slavery of being women (subject to patriarchal control in family and society), and the slavery of being workers (exploited under capitalism). Any project that addressed only one of these — that liberated workers but left patriarchy intact, or extended formal rights to women without restructuring economic life — would leave the architecture of oppression standing.
This is the source of the phrase that has become the organization's lasting summary: women had to fight on three fronts. Against fascism, which was at the door. Against capitalism, which had taught its violence to fascism. And against the patriarchal habits of their own comrades, which would reproduce both in the new society if left unaddressed.
Franco's victory in April 1939 ended the Spanish experiment. Most of Mujeres Libres' surviving members crossed the Pyrenees and were interned in French camps; some later reached Mexico, where Sánchez Saornil lived out the rest of her life, working in a typewriter shop and writing poetry she did not publish. The journal — fourteen issues in total — was preserved by a handful of exiles. The schools were destroyed or absorbed into the Franquist state. The organization itself, as a functioning thing, ended.
What did not end was the argument. It remained available, as it would later turn out, to anyone willing to read for it.
Qamishli, 2012 onward
The Kurdish women's movement did not begin in 2012. It is not even principally a movement of Rojava — its center of gravity, for most of its history, has been on the other side of the border, in the Kurdish regions of Turkey, and inside the long internal life of the Kurdistan Workers' Party (PKK).
Sakine Cansız was twenty years old in November 1978 when she helped found the PKK. She was one of two women among its founding cadres. She spent twelve years in Turkish prison after a 1979 arrest, during which she was tortured and refused to break; her account of those years remains one of the central texts of the Kurdish movement. After her release in 1991, she helped build the women's wing of the PKK — what would become, after several reorganizations, YJA Star — and traveled in exile through Europe, organizing among the Kurdish diaspora. In January 2013, she was assassinated in Paris, along with two other Kurdish women activists, Fidan Doğan and Leyla Şaylemez, by an operative linked to Turkish intelligence. She was fifty-four.
This compressed biography matters for a particular reason. The autonomous women's structure inside the Kurdish movement was not a concession won at the moment of the Rojava opening. It was four decades in the making, fought for inside an organization that — like the CNT-FAI — was nominally committed to gender equality and in practice committed to the deferred-to-later version of it. The Kurdish women's movement organized within the PKK on the same principle Mujeres Libres had organized within the CNT-FAI: not as faction, but as parallel architecture, with its own institutions and its own theory.
The theory has a name. Jineolojî — from jin, the Kurdish word for woman, plus the Greek -logia — is the body of analysis developed primarily in the 2000s by Kurdish women working in study circles, prisons, and mountain camps. It draws on Marxist feminism, on indigenous knowledge traditions, on the long history of the women's question inside Kurdish society, and on Öcalan's prison writings, which themselves absorbed and transformed Bookchin's libertarian municipalism. Jineolojî argues that the subjugation of women is the original act of hierarchy, prior to and underlying the formation of the state itself; that the recovery of women's authority is therefore not a reform within civilization but a return to a more ancient form of human life; and that no liberation of any kind can be achieved without it. The framework is contested by some Western feminists — its historical claims about prehistoric matriarchy are debatable, and its essentialism about gender is at odds with much contemporary feminist theory. But its function inside the movement is unambiguous. It is the intellectual scaffolding for parallel women's organization.
When the Assad regime withdrew from northeastern Syria in 2012, the Kurdish women's movement was already there, with decades of practice in autonomous organization. What happened over the next several years was the rapid construction, under siege conditions, of an entire civilian women's architecture inside the broader Autonomous Administration of North and East Syria.
Kongra Star — originally Yekîtiya Star — is the umbrella confederation of women's organizations. Beneath it are women's communes, women's councils at every administrative level, women's economic cooperatives (bakeries, agricultural cooperatives, textile shops, dairies), women's academies for political education, and the Mala Jin, the "women's houses" that handle disputes involving women — domestic violence, divorce, inheritance, marriage — outside the male-coded structures of customary law. Every administrative body in the Autonomous Administration is led by a co-presidency: one woman, one man, equal authority. The principle holds from the smallest neighborhood commune to the highest tier of regional government. Women hold a minimum of forty percent of positions in all decision-making bodies; in practice the figure is often higher.
The YPJ, founded in 2013, is the military wing — the Women's Protection Units, organized parallel to the mixed-gender YPG. It was the YPJ that drew international attention during the battle for Kobanî in 2014–15, when small units of Kurdish women fought ISIS in street-to-street combat and helped turn the tide of the war. The image of the female Kurdish fighter that has circulated globally since then is in some ways a misleading frame — it focuses Western attention on the most legible part of the project, the part that can be assimilated to existing categories like "women in combat" — and tends to obscure the civilian architecture that is the deeper achievement. The YPJ exists because Kongra Star exists, because the Mala Jin exist, because four decades of organizing exist. Take away the civilian structure and the military formation is unintelligible.
The siege has not lifted. Turkey has launched repeated cross-border operations, occupying Afrîn in 2018, and threatens further intervention. The Assad regime, even in its current weakened form, claims the territory. International sanctions on Syria fall on the Autonomous Administration as well, despite its political independence from Damascus. The infrastructure is fragile. ISIS detainees and their families remain in camps the Administration cannot afford to maintain indefinitely. The system carries contradictions: tensions between the official commitment to direct democracy and the de facto authority of the PYD's security apparatus, the limits of pluralism in practice, the difficulty of sustaining co-presidencies in remote villages where the practice has shallower roots.
These difficulties are real, and the Kurdish women's movement names them more honestly than most outside observers do. What is also real is that the architecture has held, and grown, and continued to produce its effects, for fourteen years and counting, under conditions of permanent siege. The civilian women's structure has outlived multiple Turkish offensives, the rise and fall of ISIS, the collapse of the Syrian state, and the abandonment by international allies. It is not a finished thing. It is also not a fragile experiment that requires protective gloves. It is, by this point, a fact.
Why It Recurs
Between Barcelona in 1936 and Qamishli in 2012, the political theory traveled but the women's-organizational form did not. Bookchin reached Öcalan. The CNT did not reach the PKK. Sánchez Saornil and Cansız never read each other. And yet what the two movements built, in the shape of their internal architecture, is almost the same thing.
The question is what to call this. It is not lineage in any direct sense. It is also not coincidence. The right term might be structural recurrence — the appearance, in widely separated revolutionary moments, of the same architectural conclusion, arrived at independently by people working from their own circumstances and their own analysis.
Silvia Federici's Caliban and the Witch (2004) offers a theoretical account of why this keeps happening. Federici's argument is that the violent suppression of women's reproductive autonomy — most spectacularly in the European witch hunts of the sixteenth and seventeenth centuries — was a foundational act of capitalist accumulation, on equal footing with the enclosure of common lands and the colonization of the Americas. Capital requires the unpaid labor of social reproduction; it cannot produce its workers, feed its workers, raise its workers' children, or maintain its workers' bodies without this labor; and it has therefore always required the violent subordination of women to ensure that the labor is performed unpaid.
If Federici is right — and the historical evidence she marshals is substantial — then the gendered architecture of capitalist accumulation is not contingent. It is structural. And what follows is that any revolutionary project that does not directly address this architecture, that treats it as a downstream issue to be resolved after the main work, will reproduce the foundational violence of capital in its own institutions. Patriarchal habit inside a workers' organization is not a personal failing of its members. It is the unreckoned-with material substrate on which the organization's collective life is being conducted.
This is what Mujeres Libres and the Kurdish women's movement, working in different idioms, both saw. The CNT could not be the vehicle of women's liberation, not because its men were uniquely benighted, but because no movement can deliver what its internal architecture forecloses. The PKK in its early years could not be the vehicle of women's liberation for the same reason. The solution in both cases was the same: build parallel structure inside the movement, equipped with its own institutions, its own theory, and its own decision-making power, so that the revolution's claim to gender equality could be tested against the actual texture of organizational life rather than asserted as a principle to be honored later.
This is what autonomy within autonomy means. It is not separatism. Both Mujeres Libres and Kongra Star worked alongside the mixed-gender organizations of which they were part. The point was not to leave; the point was to ensure that the new society being built was being built in fact and not only in name.
That serious revolutionary thinkers, working independently across continents and decades, keep arriving at this same conclusion is the most important evidence we have that they are not mistaken.
For The People's Share
The relevance of all this to a Field Guide to the Commons — and to the question of who should own the AI economy — is not immediately obvious. It becomes obvious if you ask the right question.
The right question is this: what is the gendered architecture of the AI commons we are being asked to imagine, and who is doing the unwaged or underwaged reproductive labor that holds it up?
The answer is not difficult to find. AI systems are trained on an enormous body of human-produced text, image, and code, scraped without permission or compensation, much of it produced by women and other historically uncompensated creators. They are made usable by armies of content moderators in the Global South — disproportionately women — who work for low wages under traumatic conditions to filter out the violent and sexual material that would otherwise saturate the systems' outputs. They are made accurate by data labelers performing repetitive cognitive work that has, like all forms of routinized cognitive labor historically, become a feminized sector. They are deployed in care economies — eldercare, childcare, customer service, healthcare administration — where they are positioned as labor-saving devices in industries already structured around the extraction of value from women's underpaid work.
A commons project that addresses ownership of the capital of AI — the models, the compute, the patents — and leaves this gendered architecture of reproductive labor intact will reproduce the structure of accumulation it claims to oppose. It will be Federici's witch hunt with better marketing.
Mujeres Libres and Rojava are, as far as I know, the two most serious historical answers to the question of what not doing that looks like in practice. They suggest a structural principle: that any serious commons project must build, as an integral part of its architecture, autonomous women's structures — not as moral concession but as functional necessity, because the alternative is the silent reproduction of an enclosure that the project will not have the internal capacity to see, name, or undo.
What this would mean concretely for a worker-owned AI infrastructure cooperative, or for a public option in cloud computing, or for a federated network of municipal AI commons — these are questions the Field Guide can begin to ask. They are not questions any existing AI policy framework, abundance-movement document, or open-source manifesto known to me has yet asked seriously.
The Barcelona militiawoman on the rooftop, looking out over her city in 1936. The young Kurdish woman with a tablet in her hands and smoke on the horizon in the second decade of the twenty-first century. They are not the same person. They are not even part of the same movement. But they belong, by the structural logic of what they were doing, to the same argument. The argument is not finished. It needs to be made again, in the form of the present.