Exit Under Duress
Maria
Maria runs a small e-commerce business selling handmade ceramics. She learned the craft from her grandmother in Oaxaca, refined it through two years of art school in Mexico City, and started selling online seven years ago when she moved to Portland. Her studio is the back room of a rented storefront. Two part-time assistants handle shipping and customer communications. Annual revenue is three hundred and forty thousand dollars. Her margin is twelve percent: about forty-one thousand in take-home income. The platform takes a fifteen-percent commission and provides the infrastructure that makes the business possible: the marketplace where customers find her, the payment processing, the review system that builds her reputation, the advertising tools that bring new customers to her listings.
Seven years of commercial life have deposited layers of value inside the platform's systems. Her product listings (photographs she took, descriptions she wrote, pricing she set) are unambiguously hers. But the assets that matter most are the ones she cannot simply download. Her customer list includes 4,200 buyers accumulated over seven years: regular customers who order holiday gifts every December, a restaurant owner in Chicago who buys serving platters for each new season's menu, a gift shop in Sedona that stocks her work on consignment. Every relationship exists because she made something worth buying and someone chose to buy it; yet the data also contains her customers' names, addresses, and purchasing patterns, implicating their privacy in any export.
Her seller rating (4.9 out of 5.0, based on 3,847 reviews) is the single most valuable asset she holds on the platform, more valuable than any individual listing. It determines her visibility to new customers, her access to premium placement, her eligibility for favorable commission rates, and the trust a first-time buyer places in her before the first purchase. The platform treats it as proprietary. Maria cannot take her 4.9 rating with her. She cannot even prove to another platform that the rating exists.
Beyond the rating sits her network position: visibility in search results, ranking relative to competitors in handmade ceramics, integration with the recommendation engine, the algorithmic advantages that accrue to a seller who has been active and successful for seven years. Network position is not data in any exportable sense. It is an emergent property of her interaction with the platform's systems over time: a property the platform controls and that vanishes the moment she leaves. Unlike listings, which can be copied, or a reputation score, which can be attested, network position exists only inside the system that produced it; and it is the asset whose loss makes exit prohibitively expensive for most sellers most of the time.
What Breaks
In March, the platform changes its terms of service. Commission rises from fifteen to eighteen percent, and a mandatory advertising spend of four percent is deducted automatically from payouts. All customer communications must go through the platform's messaging system, severing the direct email relationships Maria has built with repeat buyers. She runs the numbers and watches her margin drop from twelve percent to approximately five. Five percent of three hundred and forty thousand is seventeen thousand dollars: less than she pays in rent for the studio alone.
Maria creates an account on a competitor platform. She spends two weekends reformatting product listings for the new interface, uploading photographs, rewriting descriptions to match the new platform's conventions.
On day one she has zero reviews, zero rating history, zero algorithmic visibility, indistinguishable from a seller who opened an account five minutes ago. Seven years of commercial history, 4,200 customer relationships, 3,847 positive reviews, a verified track record of reliable shipping and honest descriptions, all of it left behind, attached to an account on a platform she can no longer afford to use, invisible to the new platform's customers. Not deleted. Simply inaccessible.
She emails the customers whose direct addresses she kept in a personal spreadsheet, about half the total. Some reply. Most do not. Of those who reply, some visit her new storefront and place orders. Others say they'll come back later. A few say they did not know she had left the old platform and already bought replacement items from another seller.
After six weeks, Maria has recovered eighteen percent of her customer base. Annualized revenue on the new platform is sixty-one thousand dollars, down from three hundred and forty thousand. She cannot meet her studio lease payments and lets one assistant go. She begins exploring whether she can take custom orders through Instagram while she rebuilds.
Her exit was not prohibited. It was not obstructed in any formal sense. The old platform did not prevent her from leaving. It did not threaten her. It did not lock her data behind a paywall. It simply did not help her leave, and the cost of leaving without help (the destruction of seven years of accumulated commercial value) made the exit feel less like a choice and more like a catastrophe. A political philosopher examining the situation would observe that Maria's freedom to exit was formally intact. A political economist would observe that her freedom was substantively null. The distinction between formal and substantive freedom is the one the republican tradition has drawn for centuries: the distinction between the slave who is told he may leave and the slave who has nowhere to go.
What makes Maria's situation specifically computational is the totality of the lock-in. A merchant in the Champagne fairs who was dissatisfied with the fair wardens' fees could trade through Bruges or Geneva; his reputation traveled with him in letters from correspondents who knew his character. A shopkeeper in a nineteenth-century town who was dissatisfied with his supplier could switch to a competitor; his customers lived on his street and shopped in his store because it was convenient, not because an algorithm directed them there. Maria's commercial life is constituted by the platform in a way the medieval merchant's was not constituted by the Champagne fairs. Customer relationships, reputation, visibility, transaction history, shipping and tax integrations: all exist within the platform's systems and vanish when she leaves. The platform is not a marketplace she visits. It is the substrate on which her commercial identity is constructed, and departing the substrate means dismantling the identity.
This is exit without fork rights. The right to leave exists. The capacity to arrive does not.
Maria's Alternative Exit
Under the fork-rights architecture, Maria's departure would proceed differently.
She would click "export" and receive her owned data in a standardized, machine-readable format: product listings complete with images, descriptions, pricing, and inventory status, formatted according to an interoperable standard rather than the old platform's proprietary schema. No weekends reformatting. She would upload the file to the new platform, where her listings would appear in the same form they had on the old one.
With her listings would come a portable reputation credential: a cryptographically signed attestation of her seven-year track record. Seller rating: 4.9. Return rate: 1.2 percent. Dispute resolution rate: 0.3 percent. Customer satisfaction: 97th percentile in her category. The credential would not disclose which platform issued it; it would attest to facts and allow the new platform to incorporate the attestation into its own ranking system. Maria would arrive not as a stranger but as a verified seller with a documented track record, starting with visibility proportional to demonstrated reliability, not at zero.
Because the platforms would communicate through a shared protocol, her new storefront would appear in search results alongside sellers on the old platform. A customer searching for "handmade Oaxacan ceramics" would find Maria regardless of which platform she occupies. Federated search, not siloed. Her network position would be diminished (she would have lost the algorithmic advantages specific to the old platform), but she would not have lost visibility entirely.
Her customers would receive automated notification of the migration, with a link to her new storefront and a mechanism to transfer saved preferences, wish lists, and order histories. Each customer would decide whether to follow: a simple choice, not a research project.
Under this architecture, Maria would lose some customers, those who prefer the old platform for its interface, its other sellers, its loyalty program. She would lose some network position. But she would preserve the core of her commercial value: her reputation, her product catalog, and the relationship with the majority of customers who follow her because they care about her ceramics, not about which platform displays them.
Hardening
Portable reputation credentials raise a genuine concern: if they can travel, they can be gamed, and a gamed reputation that travels is more dangerous than one confined to a single platform. Under the receipt regime, credentials are cryptographically signed by the issuing platform. Forging one would require forging the platform's cryptographic signature: computationally infeasible with current or foreseeable technology. A credential is as trustworthy as the platform that issued it, which means receiving platforms can track the correlation between attested reputation and actual performance, discounting attestations from platforms whose sellers consistently underperform their credentials. Markets discipline attesters through the same mechanism they have always used to discipline providers of unreliable information: recipients stop trusting it.
Interoperability creates a free-rider problem: a new platform displaying sellers from the old one benefits from the old platform's investment in verification without paying for it. Symmetry resolves this. Interoperability is a mutual obligation: the old platform displays sellers from the new just as the new displays sellers from the old. Both benefit from an expanded market, and neither can degrade the connection without losing access to the other's users. The equilibrium is cooperative, sustained by mutual advantage rather than adversarial extraction.
The deepest vulnerability is the cost of alternatives. Fork rights make exit possible, but someone must build the destination platform. If building a viable alternative is prohibitive, the fork right is theoretical: a right that cannot be exercised because nowhere exists to exercise it. What fork rights do is reduce the demand-side barrier: a new platform that can accept portable credentials, import seller data, and interoperate with the existing ecosystem faces a dramatically lower entry barrier than one that must build its user base from scratch. Supply-side barriers (server infrastructure, regulatory compliance, the engineering talent to build and maintain a platform) remain. Fork rights are necessary but not sufficient for competitive governance.
What the Worked Example Proves
Maria's story tests the fork-rights architecture against the platform whose market power depends on lock-in, with specific assets at stake: data, reputation, relationships, network position.
On three of four dimensions, the architecture passes. Portable data and interoperable protocols mean Maria would not spend two weekends reformatting and would not arrive as a stranger. Credentialed migration means her seven-year track record would travel with her, visible to new customers on the first day. Federated search means a customer looking for handmade Oaxacan ceramics would find her regardless of which platform she occupies. Network position is the residual loss (the algorithmic advantages specific to a single platform are emergent properties that no export mechanism can fully reconstruct), but the loss is partial, not total.
Fork rights do not make exit costless. They make exit survivable, and the difference between costless and survivable is the difference between a mathematical abstraction and a constitutional right. Maria under the current architecture loses everything; her business dies. Maria under fork-rights architecture loses some things and rebuilds; her business contracts but endures. The first is captivity. The second is freedom with friction: the only kind of freedom that has ever existed in practice, and sufficient to make voice meaningful and governance accountable.
The claim survives if Maria can exit and arrive. It fails if she cannot.